0:00:05 The content here is for informational purposes only, should not be taken as legal business
0:00:10 tax or investment advice or be used to evaluate any investment or security and is not directed
0:00:14 at any investors or potential investors in any A16Z fund.
0:00:18 For more details, please see a16z.com/disclosures.
0:00:20 Hi and welcome to the A16Z podcast.
0:00:26 I’m Lauren Murrow and in this episode, Patreon co-founder Sam Yam, an A16Z consumer tech
0:00:32 partner, Lee Jin, join me to talk about the passion economy, the rise of online platforms
0:00:35 that enable people to make a living off their unique interests and skills.
0:00:40 Though this episode was recorded before the current shelter-in-place orders, it’s a trend
0:00:45 that’s become increasingly relevant as we see a massive growth in digital work.
0:00:49 Our conversation covers the new forms of work made possible by these online platforms, why
0:00:55 creators today are effectively making more money off fewer fans, and what all this means
0:00:57 for the future of entrepreneurship.
0:01:04 You can also check out our posts on this topic as well as related content at a16z.com/passioneconomy.
0:01:08 So first off, how would you define the passion economy?
0:01:14 When I talk about the passion economy, I really mean any potential path of monetization where
0:01:18 people are leveraging their individuality.
0:01:25 So it’s beyond just podcasters and YouTubers and visual content creators and it encompasses
0:01:32 things like video course creators or virtual teachers or tutors or virtual coaches for
0:01:34 professional coaching.
0:01:40 So some of the really interesting platforms allow creators to create their own online
0:01:45 courses and these online courses encompass everything from how to organize your home
0:01:51 or how to not kill all of your plants to more traditional educational subjects like how
0:01:55 to start your own online business or how to start a podcast.
0:02:00 And then in terms of virtual teachers and tutors, there’s platforms that are giving
0:02:05 people the ability to come up with their own online courses and offer them to kids.
0:02:11 The way that we think broadly about the creator economy, which ties into the passion economy,
0:02:18 is any individual who has an interest, has a passion and is trying to grow community
0:02:19 around that.
0:02:24 So there was a study last year by the Recreate Coalition on the new creator economy.
0:02:29 They found that there are 17 million creators who are earning an income from posting creations
0:02:31 online.
0:02:35 They were covering just nine platforms, I think it included platforms like Instagram,
0:02:37 Twitch, YouTube, Tumblr.
0:02:42 The creator economy has been around for a long time, so that isn’t necessarily new.
0:02:47 In the 2010s, we saw the rise of the influencer industry, YouTube has been around for something
0:02:53 like 15 years now, Sam, you co-founded Patreon in 2013.
0:02:57 What has changed since the early days of creator platforms?
0:03:02 In 2013, people were making money online and even online for content, but the premise was
0:03:05 that the economics for these creators were broken.
0:03:09 And so you had my co-founder Jack, who was making music on YouTube, and his story is
0:03:13 that he would get hundreds of thousands, even millions of views on his videos, as he likes
0:03:14 to put it.
0:03:20 Several football-sized stadiums filled with fans, but only to log in on his ad revenue
0:03:24 dashboard and see like a few hundred dollars transferred to his back account, right?
0:03:27 And if you look at my own life, I studied almost 15 years of classical piano.
0:03:32 I performed at Carnegie Hall in New York before college, so I took the craft very seriously,
0:03:37 because I couldn’t see any viable path into making my own passion for music a career.
0:03:39 You just had to be the best.
0:03:45 And so the crux of Patreon was that there was an opportunity to provide an ongoing sustainable
0:03:48 salary to creators and that the internet enabled this.
0:03:54 Ten years ago, it would have been really difficult for many creators to make a living because
0:03:58 what has happened since then is that social networks have reached global scale.
0:04:04 The entire internet has been connected on these large platforms like Facebook and YouTube
0:04:05 and Instagram.
0:04:10 We needed that to happen first, so that creators could find their tribe, so that fans could
0:04:15 find the creators that they resonated with in order for this next wave of new creator
0:04:17 platforms to arise.
0:04:20 What types of work would fall under the passion economy umbrella?
0:04:25 In addition to the ones that people often talk about as new forms of creative careers
0:04:33 like podcasters, online writers, online course creators, there’s also a lot of emerging professions
0:04:37 that I think a lot of people aren’t even aware of yet.
0:04:41 After the passion economy blog post came out last fall, I was pleasantly surprised by how
0:04:46 broadly some founders read it and thought that it applied to their businesses.
0:04:52 There were people who were starting marketplaces for after school kids activities or summer
0:04:57 enrichment activities, and thought of themselves as a passion economy platform.
0:05:01 There’s this new potential profession in the form of a conference organizer.
0:05:07 There’s also community leaders that are going private or paid and making a living through
0:05:09 basically moderating a community.
0:05:14 And then in China, there’s platforms that enable people to find video gaming buddies
0:05:18 for either training or companionship, and I think a lot of people in the West aren’t
0:05:21 aware that people are actually making a living from doing that.
0:05:27 In our own network, we see this on mainstream areas like podcasters, YouTubers, even web
0:05:28 comics.
0:05:34 But also, we see fringe communities and categories around things like animal rescue or knitting
0:05:40 or dating advice or even a group that just sails around the world and records their adventures.
0:05:47 So for us, the creator economy is any individual that is able to post something to the rest
0:05:48 of their community online.
0:05:54 The commonalities between all of those examples is people are providing a service that really
0:06:00 only they themselves can create versus previous generations of turnkey work platforms may
0:06:07 have prioritized consistency and tried to reduce variation between workers.
0:06:09 So individuality is a feature, not a bug.
0:06:16 A lot of previous marketplaces were really about providing the most consistent experience
0:06:22 as possible, such that, for instance, whenever you hit the get a ride button, you knew exactly
0:06:24 what to expect.
0:06:29 The difference between that versus passion economy work is that as a consumer, you’re
0:06:34 specifically looking for a very, very individual type of service and the person who’s rendering
0:06:37 that service is of utmost importance to you.
0:06:42 And so it’s not about taking any online course or subscribing to any newsletter.
0:06:44 They’re not interchangeable.
0:06:47 Have you noticed a shift in the types of work that you’re seeing on Patreon?
0:06:52 Lee has made the analogy that it’s like the difference between Amazon and Shopify.
0:06:58 Like that standardized mass produced monolith and then the more indie focused goods.
0:07:02 Have you seen creators become more individualized or more niche?
0:07:03 Yeah.
0:07:06 Previously in order to survive, you have to appeal broadly to the masses, otherwise the
0:07:10 cost for marketing or distribution were never worth it, right?
0:07:14 But with the internet, you can have individuals find other members of a community.
0:07:19 In smaller gatherings, but it still becomes a critical mass.
0:07:20 This is something I want to hone in on.
0:07:23 Is this type of creative work growing?
0:07:28 Is the passion economy or what we would formerly call the creator economy, is it growing?
0:07:34 The number of one million plus subscriber YouTubers has been increasing 65% each year.
0:07:39 But on top of that, existing YouTubers are having their subscriptions monotonically increasing
0:07:42 and then all our cohorts are growing over time.
0:07:46 The way I like to actually look at this when you’re talking about overall growth is it’s
0:07:51 through this study that seems slightly unrelated where last year during the 50th anniversary
0:07:55 of the moon landing, there was a Harris study shared in USA Today.
0:07:58 They asked children, what do they want to be when they grow up, right?
0:08:05 And they found that in the US, the number one answer was that children want to be YouTubers.
0:08:10 And that there was an answer of children want to be astronauts too, but it was like three
0:08:14 to one, more children wanted to just be online and be a YouTuber.
0:08:20 And so I think this speaks to what’s going on in the whole industry.
0:08:25 Everyone’s attention and interest is in being able to build communities and follow what they’re
0:08:26 passionate about.
0:08:30 And that’s just going to dictate continual growth over time as these children grow up
0:08:35 and decide what they want to do alongside this whole passion economy.
0:08:39 Another way to think about the total potential market for the passion economy, which by the
0:08:44 way was the biggest question that I got after publication of the passion economy blog post.
0:08:48 Lots of people said, you know, this all sounds great, but how many people in the world can
0:08:51 actually make a living from their passions?
0:08:54 How big can the passion economy actually get?
0:08:55 There’s two ways to think about it.
0:09:00 One is by looking on all of the different creative platforms like YouTube and Instagram
0:09:05 and TikTok and seeing how many people have amassed a decently sized audience where they
0:09:08 could probably be able to monetize that audience.
0:09:12 But the other way that I like to think about TAM is by thinking through the demand side
0:09:19 on the consumer side, what are all the different pockets of spend that Americans are putting
0:09:24 their money into that could potentially be up for grabs by creators?
0:09:31 So for instance, Americans spend more than $32 billion a year on health and fitness, $10
0:09:34 billion on self-improvement.
0:09:39 The entire online education market is $130 billion.
0:09:46 And a lot of these categories, I think, could shift to creators and more niche content that
0:09:48 individuals are producing.
0:09:52 You mentioned that you got a lot of feedback that was like, this sounds great, but how
0:09:54 many people can actually make a living this way?
0:09:55 Yeah.
0:10:00 And you have argued that thanks to the rise of some of these modern digital platforms,
0:10:06 in fact, creators need less fans to make a living than they did in the past.
0:10:10 More than a decade ago, Kevin Kelly, who is an editor-wired, predicted the rise of the
0:10:11 passion economy.
0:10:17 So he basically argued that rather than pursuing, say, widespread fame, that creators only needed
0:10:25 to engage a base of 1,000 true fans, so 1,000 fans that would pay $100 per year.
0:10:31 So Lee, you argue that with these new tools, creators actually only need to amass 100 true
0:10:35 fans, but paying them $1,000 a year.
0:10:36 Exactly.
0:10:41 It really builds on the same idea as Kevin Kelly’s original post, which was that it doesn’t
0:10:46 take building a huge, massive audience in order to be a successful creator who can make
0:10:47 a living.
0:10:52 But I took it one step further, and this blog post called 100 True Fans, which outlined
0:10:58 the differences in the 1,000 true fans model versus the 100 true fans model.
0:11:02 It’s a difference in the kind of content they’re creating, right?
0:11:07 Content and I think the mindset that they’re tapping into among their fans.
0:11:14 So the 1,000 true fans model where people are paying you slightly less than $10 a month
0:11:21 is really about tapping into the idea of fan altruism, where users are willing to pay you
0:11:26 for something that benefits the creator because they want to support you and they want to
0:11:27 donate to you.
0:11:34 In order to get users to pay $1,000 a year, obviously there’s a tremendous amount of
0:11:39 value that needs to be offered to them in exchange for that increase in price point.
0:11:46 And so the model of the 100 true fans is really about the fan benefit rather than the creator
0:11:47 benefiting.
0:11:52 And so in order to get people to part with $1,000 a year, users need to be willing to
0:11:57 pay for premium content in community that isn’t really available elsewhere.
0:12:02 A lot of it is oriented around delivering tangible value and results.
0:12:08 So Sam, do you think there’s a shift from that influencer model of people subscribing
0:12:13 because they’re really into a creator versus creators making something for the benefit
0:12:14 of the fans?
0:12:20 Well, to the least point, we’ve certainly seen the average initial pledge amount increase
0:12:26 over time and also a shift in patrons who are paying over $100 per month, which is over
0:12:28 $1,000 a year.
0:12:32 We’ve seen that grown over 20% over the past few years.
0:12:34 I certainly think price point is relevant.
0:12:41 I think to your earlier point, what’s being offered is highly relevant to, like there’s
0:12:44 this comedy podcast called Righteous and Ratchet.
0:12:48 There’s two podcasters who offer commentary on entertainment news.
0:12:52 When they originally launched, they were generating $2,000 a month from their community
0:12:53 of supporters.
0:12:58 But when they focused on offering exclusive content, extra commentary, even early access
0:13:04 to their content, they saw that actually the business jumped up to $25,000 a month.
0:13:06 And I think there are a few reasons for this.
0:13:12 When we look at our creators are able to retain subscriptions at 90 plus percent several months
0:13:13 out.
0:13:18 And because they have a consistent stream of content that I’d argue is very intimate
0:13:23 with users, I think what happens with podcasting is, A, you’re almost speaking directly into
0:13:27 the ear of your audience and you’re doing it on a very regular basis.
0:13:32 And so when we talk about being able to move communities and having influence over them,
0:13:37 I think building that connection and that relationship through engagement is what we’ve
0:13:38 seen be very effective.
0:13:44 I think broadly there’s a rotting trust in ad driven social platforms and creators have
0:13:47 the trust and distribution of their fans.
0:13:52 For a lot of people, creators have become their main social experience and fans are
0:13:55 craving a way to engage more with them.
0:13:56 Yeah.
0:14:02 For us, we’ve seen all our creator cohorts grow over time, but the more recent cohorts
0:14:06 have actually been growing faster on overall revenue than past years of cohorts.
0:14:11 People have been able to send creators over a billion dollars through the first six years
0:14:15 of our existence, but then over half a billion of that happened in the most recent year.
0:14:20 And so we’re seeing this acceleration of growth across these communities.
0:14:27 Now with the creator economy, there is also not only different ways to monetize, but the
0:14:30 potential to more directly monetize.
0:14:33 So what are the lovers in which you do that?
0:14:38 Well, just to start off, the gig economy is obviously huge and isn’t going to go anywhere,
0:14:44 but one of the key differences between passion economy work versus gig economy work is that
0:14:48 the lovers that a creator has for growing their business are different.
0:14:53 So in the gig economy, when it was really about one time transactions through doing
0:14:59 a particular gig, the lovers for growing your revenue was just through doing more, spending
0:15:05 more time doing gigs, driving more miles, doing more jobs, et cetera, et cetera.
0:15:10 Whereas in the passion economy, the lover for growing one’s business is really about
0:15:15 expanding your audience and offering something that’s more differentiated than anything else
0:15:16 that exists.
0:15:22 So it’s about the fan base growth and the quality of your service or product.
0:15:26 So I think expansion of your audience is absolutely one dimension, which is important.
0:15:33 I think what we’re trying to promote is the depth of that relationship, too.
0:15:34 How so?
0:15:38 So Facebook and Google broke out their revenues this year for Instagram and YouTube, respectively,
0:15:44 and they basically said that Instagram was making $20 billion in 2019 and YouTube made
0:15:49 $15 billion, but both almost entirely through advertising.
0:15:53 But there’s also been a narrative that these sort of social media companies are now getting
0:16:00 criticized for focusing entirely on this notion of growing engagement and developing addiction
0:16:02 among individuals.
0:16:07 And we see it, too, where some creators that have developed a large following or likes on
0:16:12 their Facebook page or whatever, that when it comes to asking their fans to help support
0:16:15 them, you still get wildly different results.
0:16:20 And that speaks to the point of depth of connection with your audience, where you could have a
0:16:25 much smaller creator in the niche community where those fans really appreciate the content
0:16:28 in that creator, and they’re willing to pay a lot more.
0:16:35 And so you can focus more instead of mass audience penetration to what is the craft
0:16:39 that I care most deeply about that I think others will care deeply about, too.
0:16:42 What are some ways that you see creators fostering that depth?
0:16:45 It’s an easy thing to say, but how is that done?
0:16:49 When people talk about community, they typically mean a sense of belonging, which I think is
0:16:50 fostered.
0:16:54 But more than that, I think what communities mean nowadays online is actually just more
0:16:55 content.
0:17:01 And so when you go to a Reddit thread and you look at the commentary there, a lot of people
0:17:03 don’t even read the main Reddit article anymore.
0:17:07 They’re just there to see what the community has to offer in terms of additional information
0:17:09 or their opinion on a thing.
0:17:13 When you go into a Discord channel, that’s the thing that people care about.
0:17:16 They want to be very knowledgeable in a particular area.
0:17:21 Someone actually wants to learn more and they want to watch and observe other creators online
0:17:24 to understand how they’re making things.
0:17:29 There was a stat that YouTube shared that 70% of millennial YouTubers use the platform
0:17:30 to learn a new skill.
0:17:35 I think people even through merchandise are just looking for a way to represent that they
0:17:40 have this belonging in this interest for themselves as a representation of what they care about.
0:17:46 So as an example, if you look at the DJ Cascade and you’re going to one of his exclusive content
0:17:51 concerts and then you get that merch, that t-shirt or that hat or whatever that’s just
0:17:57 tied to that particular event, then when you’re out in the real world wearing this, you’re
0:18:02 effectively signaling that you’re part of this community and you’re such a zealot about
0:18:06 being a Cascade fan that you went to this exclusive concert.
0:18:09 And so I guess that’s what I mean when I’m talking about engagement, whether creators
0:18:15 are able to foster that sort of excitement around wanting to get a depth of knowledge
0:18:20 in either their interest area or the community’s passion and interests.
0:18:27 So I think of depth as a very critical element of a lot of these new platforms as Sam mentioned.
0:18:32 When I’ve observed a lot of the startups in this space, a lot of them have this initial
0:18:37 go-to-market wedge of a single product that they give to creators that they can monetize
0:18:38 through.
0:18:43 So for instance, Substack started with newsletters, Run the World is starting with online events,
0:18:47 but they oftentimes have a vision of offering greater depth over time.
0:18:55 In other words, they are trying to offer additional content types or products that give creators
0:18:59 the ability to earn more even if they keep the audience size the same.
0:19:04 So for instance, for Substack, beyond just newsletters, now they also support the ability
0:19:08 for creators to publish paid podcasts.
0:19:14 And for Run the World, a lot of the long-term vision is about supporting ongoing community
0:19:17 building and community interaction and engagement.
0:19:22 And you can imagine that creators are going to be able to earn more if they not only charge
0:19:28 for these one-time ad hoc events, but they’re able to charge for belonging to an ongoing
0:19:29 existing community.
0:19:36 And that’s part of your 100 True Fans theory as well, that you can start with a broad fan
0:19:41 base and then only a small percentage of those actually make it up the ranks in which they’re
0:19:46 paying $1,000 a year or more for really premium content.
0:19:47 Exactly.
0:19:52 But at the same time, what we’ve seen is that the network of all these creators is getting
0:19:54 denser and closer over time.
0:20:00 And so with new creators on Patreon, at least 48% of their initial revenue comes from the
0:20:03 existing network of patrons who are already on Patreon.
0:20:08 And we’ve seen this increase over time, and actually over half of revenue overall now
0:20:11 comes from patrons who are subscribing to multiple creators.
0:20:15 And that only flipped for us a few years ago, where instead of most of our revenue coming
0:20:19 from people just pledging to one creator, now people are paying multiple creators at
0:20:20 a time.
0:20:23 And so yeah, what does that all mean?
0:20:28 From my perspective, this notion of a niche community is actually becoming less isolated.
0:20:30 They’re not orphan groups.
0:20:34 And so you’ll see these collaborations happen within creators on areas or communities that
0:20:36 look niche to begin with.
0:20:37 I think that’s super interesting.
0:20:42 And when you first told me that, I sort of puzzled over it because I had been thinking
0:20:47 that the whole creator economy space and everyone monetizing online was really about tapping
0:20:51 into the specific niche that they couldn’t access anywhere else and that each of those
0:20:54 fan bases were probably dependent of each other.
0:21:02 My takeaway is that I think overall consumers broadly are shifting their information diets
0:21:05 to things that are creator led.
0:21:07 As opposed to what?
0:21:09 As opposed to traditional media sources.
0:21:14 So, for instance, before Substack, any aspiring writer had to go get a job at a traditional
0:21:16 media company.
0:21:20 Someone at that company had to give you a platform to write and distribute to an audience.
0:21:24 With passion economy platforms, it’s about directly connecting the creator with their
0:21:25 audience.
0:21:29 Except the passion economy is basically cutting out the middleman.
0:21:30 Yeah, exactly.
0:21:35 It’s those layers of intermediaries that sit between a creator and his or her audience
0:21:41 that is taking a cut that really makes it more difficult for a creator to make a living
0:21:43 off of a smaller audience size.
0:21:47 Consumers are deciding that they’re not going to go with the content subscriptions that someone
0:21:50 else has bundled and determined came together.
0:21:55 But they’re creating their own information bundles from creator led content.
0:21:56 Yeah.
0:21:59 Your point that people are unbundling their consumption, we’ve seen across the board.
0:22:05 You have folks like Try Guys, who used to be part of Buzzfeed, spin out where their fans
0:22:06 are supporting them directly.
0:22:10 You have this group kind of funny, which used to be part of IGN.
0:22:14 And then again, also managed to build a sustainable career on their own.
0:22:19 We’ve talked about how previously it used to be that only established businesses could
0:22:24 create websites and edit software and build apps, and now that’s no longer the case.
0:22:30 So how do you think about not only attracting creators and patrons, but then providing support
0:22:33 for continued growth and engagement?
0:22:35 How do you think about that from the platform perspective?
0:22:40 I think it’s happening through the network, meaning whereas previously you had to go through
0:22:42 large media providers.
0:22:45 Those media providers can now be other creators.
0:22:49 And so this notion of collaborating, where the barrier of entry of collaboration is much
0:22:50 lower too.
0:22:51 Creators are willing to work with each other.
0:22:56 I think it’s sort of this shared plight of figuring out how to see sustainability in
0:22:57 this business.
0:23:02 If you look at what happened with the YouTubers Jeffree Star and Shane Dawson.
0:23:06 So they on YouTube collectively have 40 million subscribers, but they collaborated together
0:23:11 on a makeup line and they sold out tens of millions of dollars on this launch in just
0:23:12 half an hour.
0:23:18 And then immediately the makeup stock was then being resold on eBay and they were outselling
0:23:24 billion dollar plus company makeup lines because of their power as an influencer.
0:23:29 So direct connection with an audience is immensely powerful and you only develop this credibility
0:23:33 through meaningful connection with the audience, which is what the internet has enabled.
0:23:38 So we’ve talked about how one of the key features of the passion economy is building this direct
0:23:41 channel between the creator and the consumer.
0:23:44 And we’ve talked about how you need to foster engagement.
0:23:50 But once you create this direct engagement, there is the risk that then that creator will
0:23:53 take their fans and go off your platform.
0:23:55 How do you guard against that?
0:24:00 The short answer is we started from day one thinking about your audience very differently
0:24:03 and your community very differently than other platforms.
0:24:08 And our thinking around this is this creative first mentality.
0:24:12 For us, it’s the belief that if we can provide you the right tools to interact with your
0:24:17 audience, that’s why you’re going to stay here, not because you’re effectively stuck
0:24:18 on the platform.
0:24:19 That’s interesting.
0:24:24 So in the same way that these creators have to provide specific value to their users,
0:24:29 the platforms also have to provide appreciable value to the creators in order to keep them
0:24:30 on platform.
0:24:34 Yeah, it’s a refreshing take, I think, in the industry.
0:24:39 The whole thesis about the passion economy stemmed from doing a lot of in-depth research
0:24:45 in consumer marketplaces and from studying the evolution of different marketplace models
0:24:47 over time.
0:24:51 And my colleague Andrew Chen and I had traced all of these different marketplace paradigms
0:24:56 from the Craigslist model to the Uber for X model to managed marketplaces.
0:25:02 And a trend that I saw evolving there was that newer generations of marketplaces were
0:25:08 taking a much more supply focused approach and thinking about how can we become a really
0:25:12 attractive place for these service providers to work.
0:25:19 So one catalyst of the passion economy thesis was observing the evolution of marketplaces.
0:25:25 Another driver is my personal experience of having studied art and painting intensively
0:25:26 growing up.
0:25:33 Between the age of five and 18, I took private art classes every single weekend.
0:25:37 And when I was a senior in high school, my path really branched out between going to
0:25:41 art school versus going to traditional college.
0:25:48 And like Sam, I just didn’t see a viable path towards actually becoming a professional
0:25:53 artist and being able to make a living and sustain myself that way.
0:25:56 And you took the much more risk averse path, right?
0:26:00 Like me too of startups, which is what I believe is before this.
0:26:01 Yeah.
0:26:06 So the passion economy is not just this intellectual exercise about how people make a living and
0:26:07 the future of work.
0:26:10 Well, thank you so much for joining us on the A16Z podcast.
0:26:14 If you start painting again and Sam, if you start playing the piano, I’ll subscribe to
0:26:16 both of your channels.
0:26:18 Thanks, Lauren.
0:26:23 Thank you.

0:00:05 Hi, and welcome to the A16Z podcast. I’m Hannah. In some ways, the coronavirus feels
0:00:08 like something that came out of nowhere, a sort of black swan event. But at the same
0:00:13 time, it’s also been exposing a lot of the fundamental cracks and flaws in our healthcare
0:00:17 system that feel like a perfect storm when they all come together and have hurt our ability
0:00:22 to address a problem that many say we should actually have seen coming. So we’re here today
0:00:27 to talk with A16Z General Partners Jorge Conde and Julie Yu about some of those big forces
0:00:32 and dynamics in the healthcare system that led us to this moment at the intersection of
0:00:37 business, policy, and public health, and how in healthcare in particular, broken policy
0:00:42 can lead to broken business models that in the wrong circumstances lead to major failures
0:00:46 in public health like the one we’re seeing today. We talk about where we’ve seen this
0:00:51 before in the markets of vaccines, antibiotics, and diagnostics, and what should be different
0:00:56 next time so that when a new pandemic hits, we aren’t facing another perfect storm.
0:01:00 Since we have talked quite a bit about some of the cracks on the healthcare system side,
0:01:04 let’s start with the business or the market side. What was the underlying problem on the
0:01:06 market side that got us here?
0:01:12 Much like a virus attacks the weaknesses in the human body, this pandemic spreads by effectively
0:01:16 attacking or exploiting the weaknesses across the healthcare system writ large in the United
0:01:22 States, healthcare is a fairly unique case in that this is one industry in one area
0:01:30 where policy sets business models and where you have bad policy, you have bad business
0:01:34 models which can lead to market failures, which can lead to public health failures. We’ve
0:01:40 seen it happen in the vaccines industry. We’ve seen it happen in diagnostics. We’ve seen
0:01:45 it happen in anti-infectives and antibiotics more broadly. And so these aren’t isolated
0:01:47 examples where the system fails.
0:01:52 As a system, historically, we’ve made very little investment into prevention in general.
0:01:57 Really where the money has been made is in the treatment of patients who get sick. And
0:02:01 so there’s been really just an orientation around the incentives being aligned with waiting
0:02:08 until those patients do get sick to then provide treatments and therapies and procedures that
0:02:13 generate more revenue, unfortunately, and also higher margins for physicians and hospitals.
0:02:16 And you’re seeing a version of that here where, again, there’s been very little investment
0:02:20 in preparedness for these kinds of pandemic disasters.
0:02:24 So let’s pull apart those threads in those three different areas. So because vaccines
0:02:29 are so top of mind, let’s dive into vaccines. Why does that not a successful market?
0:02:33 Well, I think vaccines in general has been a difficult industry for a couple of reasons.
0:02:39 Number one, if you’re developing a vaccine for something that already affects humankind
0:02:45 broadly, those are considered pretty commodity products today, right? Your mom’s vaccines,
0:02:49 your, you know, any of the number of vaccines that children get in their regular stables
0:02:53 are relatively commodity product. They’re not differentiated. You can’t charge a lot
0:02:58 for them because these are almost basic staples of public health. And therefore, you know,
0:03:01 relatively speaking, we all need them. Yeah, we all need them. They’re well covered and
0:03:05 they’re widely available. The trick comes when you have something that emerges quickly
0:03:09 and has the potential to spread rapidly. COVID-19 is not the first example of this that we’ve
0:03:15 seen. We’ve seen this happen numerous times, whether it’s with SARS or H1N1 or West Nile
0:03:20 or Zika. Obviously, Ebola was one that was a wake up call and made a lot of people nervous.
0:03:25 And I think if you look at a lot of those historical examples, generally speaking, what
0:03:33 we saw happen was the companies that did have active vaccine programs were asked essentially
0:03:38 to stop doing everything they were doing to develop programs against whatever a specific
0:03:43 emerging threat was. And they responded to that call for action. And what happens with
0:03:48 vaccines is when the threat goes away, you know, the urgency tends to go as way as well.
0:03:52 And once the urgency has gone, the market goes with that, along with that, these companies
0:03:57 that were developing vaccines for these specific newly emerging threats. In many cases, they
0:04:01 were left holding the bag where, you know, in some cases, orders that governments had
0:04:06 put in for them kind of went away when the threat went away. And so the business model
0:04:09 of saying, well, ramp up production for an emerging threat, where if you are successful
0:04:13 the market for what you’re developing goes away. And so the public health efforts are
0:04:16 successful in containing the disease and you don’t need the vaccine. And therefore there’s
0:04:21 no market for it. And therefore you essentially wasted the effort. The other thing that’s true,
0:04:25 that has been true in previous pandemics is there’s generally a call on the industry by
0:04:31 governments to make sure that the vaccines become available at cost or at low cost. And
0:04:35 it’s very hard to build a business around that, where if you’re building for, you know,
0:04:40 an event like you called it a black swan event. But at the moment of that black swan event,
0:04:43 you’re not able to recoup profits for some valid reasons that you obviously have to make
0:04:47 this widely available. It’s a very difficult business model unless you’re in a situation
0:04:51 where it’s the entire globe, and then it’s all of a sudden. Well, yeah, but then you
0:04:55 need the ability to actually ramp up and be able to produce vaccines at that scale and
0:05:00 at that speed. Something that’s a silver lining in all of this is how rapidly some of these
0:05:05 novel platforms have been at looking to develop vaccine candidates. But for that to work as
0:05:10 a business model, you need to either be able to produce a vaccine for every oncoming pandemic,
0:05:15 and you need to be able to have a business model that will enable you to essentially recoup
0:05:19 and make a profit from the investments you’ve made in developing that platform. And that
0:05:20 goes back to policy.
0:05:22 Yeah. I mean, one interesting thing that comes to mind, Horace, you’re talking about this
0:05:28 is there’s a lot of characteristics of vaccines that are somewhat similar to some of the more
0:05:32 novel sort of gene therapies and cell therapies that we’ve talked about, where you essentially
0:05:36 need the vaccine once or maybe a handful of times when it comes to things like booster
0:05:41 shots for your entire life. And therefore, the opportunity to monetize that particular
0:05:46 intervention is very rare in the context of any one patient. You just have to wonder
0:05:51 whether some of the dialogue that’s happening around value-based payments for different
0:05:55 types of treatments, how that would be applied here, because to take the traditional fee-for-service
0:06:01 way of thinking about getting paid a commodity price for a one-time intervention just doesn’t
0:06:05 seem to match well with the paradigm of how vaccines are actually administered.
0:06:08 So I think the analogy is valid. And by the way, I think a lot of the proof is in the
0:06:15 pudding because that kind of incentive structure that you just described hasn’t existed historically.
0:06:20 There have been several calls for better policy put in place. Some of the key areas that have
0:06:27 pioneered this have been sort of the one and done therapies for very rare diseases. And
0:06:32 I think the reason why there was room to have that discussion was, number one, because the
0:06:36 prevalence of those diseases is fairly well understood, even though they’re very rare.
0:06:41 And so you can essentially run the actuarial calculation to say, “Okay, well, if we charge
0:06:47 $2 million to treat somatic muscular atrophy, we’re still benefiting the system a significant
0:06:53 amount by extending life and reducing the need for supportive care and all of those things.”
0:06:58 And so you can actually model those out. And it’s a rare event that with a high value,
0:07:03 you can put a high price on. Here, it’s a little bit trickier because it’s hard to model
0:07:08 out the actuarials on this because a pandemic is, generally speaking, an unknown event.
0:07:13 So unless you have a policy for pandemics broadly, where you essentially provide some
0:07:18 sort of incentive, whether they’re block grants or success fees, et cetera, that would have
0:07:23 to be really large for any institution that comes up with a effective vaccine against
0:07:29 a newly emerging threat, you’re sort of trying to solve for an unknown number. But I think
0:07:34 we could learn a lot from what we’re seeing with these new modalities like gene therapies
0:07:39 and cell therapies that have in many ways trailblazed novel business models to make them viable.
0:07:44 I understand how for vaccines that model falls apart when sort of a massive event like this
0:07:48 happens really quickly. But how about with antibiotics? That feels like something that
0:07:53 everybody needs, that we know there’s an increasing demand for the right kind of antibiotic. Where
0:07:58 does the market policy, health care, public health failure come together there?
0:08:03 Well, so there it’s, you know, similar situation with different circumstances, right? So in
0:08:08 the case of antibiotics, public health agencies and the medical practice broadly have been
0:08:15 very focused on how you prescribe antibiotics and in what order, in order to prevent the
0:08:21 emergence of resistance. And so we have broad spectrum antibiotics, and then we have narrow
0:08:27 spectrum antibiotics, and we have ones that are more potent than others. And once bacteria
0:08:32 become resistance to all things, and we have the threat of superbugs, where you basically
0:08:38 have no last line of treatment against a bacteria that has become resistant. And so the conventional
0:08:43 thinking around antibiotics has been to think of it as first line and second line and third
0:08:49 line. And so if you develop a novel antibiotic, you might be addressing a very important unmet
0:08:57 need. But by definition, physicians are going to use these new antibiotics as sort of last
0:09:03 line of therapy. I develop a new antibiotic, and it’s good, it’s only going to be used
0:09:09 sparingly. And if I’m charging per treatment or per use, that’s obviously not going to
0:09:14 be a very effective model for me either. Right. You don’t want to be having to use these unless
0:09:18 everything else has failed. That’s right. And by the way, we’ve seen the real world
0:09:23 examples of this as well. Large pharmaceutical companies have also exited, like with vaccines,
0:09:29 they’ve also exited the antibiotic space. Novartis has exited, Sanofi has exited. It’s
0:09:33 very hard for them to find a way to make a profitable business for all the reasons we’ve
0:09:37 been talking about. There was a startup called the Kaogen that was actually successful in
0:09:43 developing a novel antibiotic against a nasty bug. So they had a public health success.
0:09:48 But what they found was that they didn’t have a business model. And so they went bankrupt.
0:09:51 This is a really strange industry where you have something that’s so successful that you
0:09:55 cannot build a business on it. Scott Gottlieb, the former head of the FDA, who’s of course
0:10:01 been so vocal and helpful throughout this COVID-19 pandemic. He had floated the idea
0:10:06 a year ago of similarly creating some sort of incentive structure so people could develop
0:10:12 new business models around novel antibiotics. And these included things like, similar to
0:10:18 vaccines, success fees for developing a novel therapy. He even floated an idea on something
0:10:23 like subscriptions where, you know, institutions would subscribe to get access to antibiotics
0:10:28 and they wouldn’t pay for use. They would pay for access. So it’s almost like an all-you-can-treat
0:10:33 model versus, you know, a paper pill model. Why would that need to come from policy? Why
0:10:37 wouldn’t that just make good market sense for the market to respond in that way? Well,
0:10:40 I think you’re going to see a combination of those things. So some of it needs to come
0:10:46 from policy in the sense that you can require institutions to have a broad toolkit of antibiotics
0:10:50 and require them to carry all of them. And then if you require that, the companies are
0:10:53 developing novel antibiotics, come up with new business models to make it viable for
0:10:58 them and to make it viable for the hospital systems. But I mean, I think you do need policy
0:11:04 at some level to define what is required for standard of care. And if you can do that,
0:11:09 then you at least create some room for companies to develop novel business models. A great
0:11:15 example is in self-therapy and gene therapy. Like unless you change the way CMS pays for
0:11:20 therapies and how they calculate cost of therapy, you can’t have installment models in medicine.
0:11:22 So there had to be a policy change there.
0:11:26 People do look to CMS to be kind of the tip of the spear with regards to these things.
0:11:33 CMS reimbursement is the ultimate policy change that generally is what triggers this type of
0:11:38 business model innovation and adoption of these types of things in the market. And commercial
0:11:44 payers tend to follow suit when CMS makes those kinds of policies. CMS did put into place
0:11:49 a what’s called a qualified infectious disease products program in partnership with the FDA
0:11:56 around antibiotics to essentially up the reimbursement rate for the use of antibiotics in hospital
0:12:01 based settings under certain circumstances. So there is some precedent that shows that
0:12:05 there is a willingness to do this sort of from the top down. There was other legislation
0:12:11 proposed that has not yet been passed around the concept of essentially mandate that the
0:12:17 hospital report on the utilization of antibiotics and sort of demonstrate that they are abiding
0:12:21 by standard of care. And it’s sort of a quid pro quo where in exchange for doing that and
0:12:26 sort of demonstrating compliance with that kind of program, the proposal is that there
0:12:28 would be higher reimbursement rates for that.
0:12:32 The other interesting thing about antibiotics that I find is I’ve had a number of experiences
0:12:36 myself where let’s say you go to an urgent care clinic, you think you have strep, they’ll
0:12:40 do a test, but the test results aren’t going to come back for a couple of days and therefore
0:12:46 you are put sort of prophylactically on an antibiotic without yet knowing what the pathogen
0:12:51 is or whether or not you even have that pathogen. And there’s studies that show that it’s upwards
0:12:57 of a third of antibiotic prescription could be unnecessary, but given the delay in the
0:13:00 information needed to determine whether or not it’s correct, that’s just sort of the
0:13:04 general practice of how it plays out. So it’s interesting in that this sort of intersects
0:13:10 with how having better real time diagnostic capabilities could actually impact and just
0:13:13 reduce the inefficiency amount antibiotic use as well.
0:13:18 So it’s interesting not just the sort of market policy public health failure starts to come
0:13:22 together, those actual three buckets start coming together as well. But let’s talk about
0:13:29 diagnostics now and pull again apart these sort of different failures of market policy
0:13:35 public health. Why is the diagnostic market not as successful? Why was that broken?
0:13:39 Diagnostics historically have been a challenging area for a couple of reasons. The first one
0:13:46 is it takes a while to develop a diagnostic. It’s time consuming and expensive and it wasn’t
0:13:51 easy to really recoup that investment because reimbursement rates historically haven’t been
0:13:52 very high.
0:13:56 But why would it not always be better to know if you have some, I’m sort of like not clear
0:14:00 on the very question of why is a diagnostic not always valuable?
0:14:03 You know, look, I think diagnostics are a form of prevention. And I think if one thing
0:14:08 we’ve seen generally across the healthcare system is prevention is not always the most
0:14:12 effective business model. Whether or not it’s valuable isn’t the debate. It’s how it’s
0:14:16 reimbursed. There’s plenty of examples where reimbursement rates don’t always track to
0:14:21 value, especially in the system where a lot of things have been around fee for service
0:14:24 versus value based care.
0:14:30 Diagnostics as an industry is a fairly tricky one because reimbursement levels for diagnostics
0:14:36 have over time have trended down. If you look at things like diagnostics that could be used
0:14:42 to determine which patients should be a candidate for a therapy, for any therapy, for any number
0:14:47 of conditions, those have historically been underutilized. One of the reasons that people
0:14:51 speculate is because physicians are going to prescribe the treatment that they think
0:14:55 their patients are going to benefit from. And unless the diagnostic is meaningfully going
0:14:58 to change that decision, they probably won’t use it. They’ll use other factors like risk
0:15:04 factors or family history. In many ways, the diagnostics industry has been viewed, I think
0:15:12 unfairly, as a commodity industry. There is an element of competing on price and trying
0:15:15 to essentially make it up with volume. And that’s why you have very large players in
0:15:20 the industry that have these very, very large menus of tests that they provide and they
0:15:24 do it at scale because the amount you’re going to reimburse on a per test basis is relatively
0:15:30 small. And so this whole area of diagnostics as being viewed as a commodity has led to
0:15:31 generally underinvestment.
0:15:37 It’s not that it’s not perceived as useful. It’s more that the price that you get paid
0:15:42 for that component of the patient journey is not where the money’s made. Like a cholesterol
0:15:47 test, a very basic test that’s pretty routine. The value of the test itself is probably a
0:15:52 few tens of dollars in terms of how much you get reimbursed for that test. But the outcome
0:15:56 of that test, if you are in fact diagnosed as having high cholesterol, the result of
0:15:59 that is that you’re put on a statin really for the rest of your life. And that’s really
0:16:03 the blockbuster category of drugs. And so that’s just an example of where the value
0:16:10 distribution is very skewed. CMS very recently just published their list price sheet for
0:16:16 how they intend to reimburse for the COVID test. You do a double take when you see it.
0:16:22 It’s literally in the $35 to $50 range per test that would be reimbursed by Medicare.
0:16:26 And when you think about the value of the information being provided by that test, especially
0:16:31 right now when that information could literally mean life or death, it’s kind of astounding
0:16:36 to think that it’s so little compared to these disaster stories that you hear about patients
0:16:41 in the ICU on ventilators, having to undergo anesthesia to get those ventilators in. There’s
0:16:46 just a huge skew in terms of where the dollars flow relative to where the perception of the
0:16:47 value might be.
0:16:50 I think what we’ve seen in diagnostics in terms of how they have been deployed across
0:16:55 the country is a great example of, unless you have very clear policy on who can get
0:17:01 tested and what it will cost, it makes it very hard for this to be deployed very widely.
0:17:06 And we can put aside for a second any issues that actually took place in the logistics
0:17:11 of rolling out tests. The incentives just weren’t clearly in place early enough.
0:17:15 I mean, what we’re talking about is all these broken models that show up when something,
0:17:20 do we need to wait for these moments of really kind of horrifying failure when there isn’t
0:17:25 a vaccine available, when we don’t have the right antibiotics or the right tests in order
0:17:28 to change some of these ingredients to prevent it from happening again?
0:17:33 Well, I think the specifics will matter. But I think generally speaking, when the next
0:17:36 pandemic comes, there are a couple of key things that I think would be helpful to have
0:17:41 in place beforehand. One, from a policy standpoint, it should be clear that, number one, there
0:17:47 won’t be any cost for any member of the public to get tested, assuming tests are available.
0:17:52 Number two, for those companies that are providing tests that need to in many cases massively
0:17:57 scale up capacity in order to meet demand, that they will be reimbursed in a meaningful
0:18:03 and timely manner. Number three, I think it would be helpful to have from a policy standpoint
0:18:07 coverage of treatments associated with whatever the pandemic is.
0:18:12 And that way, you don’t have the situation where somebody that’s underinsured or uninsured
0:18:17 is avoiding the healthcare system out of fear of cost repercussions. I mean, that’s obviously
0:18:21 a challenging thing to do in a normal environment. It can be a catastrophic thing to do in a
0:18:25 rapidly emerging pandemic environment. And so I think there’s certain things you can
0:18:30 do from a policy standpoint to make sure that both the public is responsive, but also industry
0:18:34 and those that have technology to address the situation are also responsive.
0:18:39 But how about from a market perspective? I mean, are there things that you can do to
0:18:45 fix these basic market dynamics of dealing with rare emerging diseases and last line
0:18:48 of defense antibiotics, things like that?
0:18:53 So if the policy were to be anyone that can provide a validated test against the pandemic
0:18:59 is going to be reimbursed X amount, and these are going to be the reimbursement criteria,
0:19:05 then you have incentive to very quickly scale up production to do that. If the policy is
0:19:10 that anyone who wants and seeks out testing, regardless of whatever that individual circumstances
0:19:14 would be, they can get access to that test and they won’t be charged at all for that
0:19:20 test. You’ve also created demand. In effect, what’s necessary in a pandemic is to ensure
0:19:24 that you have a guaranteed buyer, whether it’s, you know, you’re developing vaccines
0:19:28 that are going to be deployed broadly, whether you’re developing tests that need to be distributed
0:19:33 around the country or around the world, you know, whatever the situation may be, if you
0:19:39 need to scale something up for a one time event, you need to be sure from the supply
0:19:45 side that someone is going to buy that product. And really the only institutions that have
0:19:49 that kind of scale when you’re talking about pandemic level events are nation states or
0:19:54 countries. And so I think another key aspect of the right policy in place is essentially
0:19:59 have the backstop guarantee that the government is going to step in and buy product if you
0:20:03 develop it to address an unexpected event like a pandemic.
0:20:08 One of the big failures that we’re experiencing right now is that we who are told to stay at
0:20:13 home and avoid the hospital don’t have a means to get tested at home. And, you know, there
0:20:18 were, I think there was a set of unintended consequences that occurred when the FDA intervened
0:20:23 and put into place this emergency approval process for new tests that were being developed
0:20:27 for COVID did actually, and I think this was done unintentionally, but is one of the major
0:20:31 reasons why there’s been such a bottleneck in getting new tests developed unintentionally
0:20:36 created a need or requirement that the FDA does approve under these emergency provisions
0:20:41 every test that is being developed. And so we saw a bunch of innovative companies, startups,
0:20:46 et cetera, spin up home based tests that would work for patients in their home setting, which
0:20:51 would not require them to go to a lab or go to a clinic to get the testing done. But those
0:20:57 were unfortunately shut down because they did not comply with these emergency rules.
0:21:01 And I think things like that where there are sort of intersecting policies that may conflict
0:21:05 with each other between different government agencies have to be looked at differently
0:21:10 such that the next time this happens, we’re able to respond and have the market, you know,
0:21:14 sort of do its thing and be able to get new product out there in a way that doesn’t get
0:21:15 hung up in bureaucracy.
0:21:18 Yeah, it’s interesting because I feel like a lot of what we’re talking about is this
0:21:24 delicate balance of regulation, right, in a heavily regulated space where regulation
0:21:28 can push forward innovation in some ways, but then time passes and you don’t want that
0:21:33 to then become kind of a weight that you’re dragging along when events change or things
0:21:39 unfold. Like, how do you generally think about that balance between the heavy regulation
0:21:45 that we need in place for helping to sort of perhaps create market incentives or the
0:21:49 relaxation we need in others like in testing therapies? How do you think about how those
0:21:52 evolve over time and making sure that we keep that balance?
0:21:56 Yeah, I mean, we talked a little bit about this with regards to telehealth and the fact
0:22:02 that many of the laws both at the state and the federal level that have prevented or at
0:22:07 least created friction for telehealth to be very broadly adopted. Our laws that were put
0:22:12 into place like decades ago before anyone contemplated telehealth. And so, you know,
0:22:16 even though it’s unfortunate that this kind of event and pandemic is the forcing function
0:22:20 for us to revisit some of those things and actually undo them because the world has
0:22:24 changed so much because technology has changed so much because the way that we deliver care
0:22:28 is changing so much where the laws that have been put into place are, you know, maybe very
0:22:32 outdated. I think that’s just a general thing that this whole pandemic is shedding a light
0:22:36 on is the fact that so many of those rules that have been put into place, you know, likely
0:22:39 don’t make sense anymore in this day and age.
0:22:44 Look, I think the idea that we need to ensure that, you know, we do no harm, that things
0:22:49 that we develop to be an effective intervention against how we treat or manage disease are
0:22:54 in fact effective to ensure that we have the processes in place to make sure that we are
0:22:59 in fact not blind to issues around safety or efficacy. I think it’s very important.
0:23:03 But as we’ve talked about before, the world has changed in terms of, you know, what medicines
0:23:08 look like. You know, now we have gene therapies and cell therapies and digital therapies. And
0:23:14 I think the regulation has to evolve so we can properly evaluate those opportunities to
0:23:19 change how we treat disease. And we’re also seeing a lot of innovation that’s coming down
0:23:24 the pipe right now in terms of how we think about how do we monitor patients, right? Historically,
0:23:29 that was always done, you know, at the point of healthcare at a hospital or at an institution.
0:23:33 But it’s, you know, things like, you know, monitoring of patients with wearables emerge.
0:23:37 That’s going to change a lot of paradigm. So I think the reality is that, you know,
0:23:41 it is true that regulation is always lagging behind innovation. I think that will always
0:23:42 be true.
0:23:47 But I think what we need to do is find ways to close the gap between the innovation that
0:23:52 emerge and the regulations that help us govern those innovations. And to the extent that we
0:23:56 can further close that gap, we’ll obviously be in a better position, you know, for future
0:23:58 emerging health threats.
0:24:00 Thank you guys so much for joining us on the A16Z podcast.

0:00:05 Hi, and welcome to the A16Z podcast. I’m Doss, and in today’s episode, Jason Warner,
0:00:12 the CTO of GitHub, talks with A16Z General Partner David Yulovic about the best and worst
0:00:18 practices of remote development and, more broadly, remote work. Shelter and place orders have more
0:00:23 companies making the move to remote faster than ever. So is this our new normal? And is it the
0:00:29 future of work? In this hallway-style jam, the two discuss the tools and strategies for communication
0:00:35 and alignment, what does and doesn’t work, from demos to burn down charts, and end on some reflections
0:00:41 given the current crisis. They begin, however, by talking about where location does, or maybe we
0:00:48 should say doesn’t come in. I’ve been a remote executive now for just over 10 years, and I myself
0:00:53 have never lived in San Francisco. I wonder how many people list their location on LinkedIn as the
0:00:59 San Francisco Bay Area, but don’t actually live here. I’m sure it overcomes the Silicon Valley bubble
0:01:04 bias of everyone having to be here, which I guess is what we’re talking about today. I think my career
0:01:08 would look very different if I didn’t put San Francisco on at one point. That’s really interesting.
0:01:14 I wonder if that’s going to change in the future. I hope it does. I really do. You can get more
0:01:19 efficiencies in your business. You can have better access to talent. And Silicon Valley has a bias
0:01:25 towards them being the best of the best. But the world has moved on from that. And I think what
0:01:30 you’re going to find too, and I think we will find this in time in this next decade, is that
0:01:36 the access to talent is the only game that really matters in the next 10 years.
0:01:41 The aperture will open for roles that previously people thought couldn’t possibly be done remote,
0:01:46 that this particular role had to be in the office. And then now, of course, that role is working
0:01:50 because it’s remote by definition because we’re in this moment in time. Jason, what are the dynamics
0:01:55 that you’re observing, I think both in this moment in time and then more broadly?
0:02:01 So I think the most common thing that I’ve seen when people try to go distributed and remote as a
0:02:09 company is that some of the worst human behaviors or tendencies tend to get amplified. So as an
0:02:15 example, if you are a person who is prone to gossipy type of stuff, you might seek out more
0:02:21 gossip. If you’re prone to anxiety and you worry about your job, someone not replying to you in
0:02:27 an instantaneous basis might give you heartburn. However, the worst one is micromanagement. If
0:02:33 you’re a boss who micromanages people when you’re in the office, it gets amplified 100 times remote
0:02:41 because you tend to “want to check in to see how things are going” or reopen decisions. Typically
0:02:46 to think about what the behavior mechanism of micromanagement is, it’s lack of either transparency
0:02:52 inside of trust. So figure out which one of those it is and fix that. How you fight that as a tendency
0:02:57 is you really have to clear up your lines of communication. What are you doing for why? How
0:03:02 are decisions being made and ratified? Who has the decision-making authority? How are we tracking
0:03:07 progress? Do you have the right structures, meetings, cadence, or updates that you need
0:03:12 to feel confident that things are moving in a direction? I think the ability to go
0:03:17 to a remote workforce for me as a leader means the ability to hold people accountable who aren’t
0:03:21 directly in front of me and know how to communicate over various communications mediums that don’t
0:03:26 involve just walking over to their desk and annoying them, which is manager 101. Sometimes
0:03:31 when you see somebody face-to-face all the time, you can write off that quick one-line email that
0:03:34 might be a little bit snarky or they might misinterpret because you’re going to see them
0:03:38 the next morning. But when somebody’s working remote, you start to put more context around
0:03:43 things and you start to say thank you and please, you can’t just see them in the morning and say,
0:03:47 “Oh, hey, sorry about that email last night.” I found that to be one of the most interesting
0:03:52 transitions for me personally. I got a lot better about communication in a remote company,
0:03:58 distributed company, really fast. And as you said, more context, a little bit more empathy
0:04:03 about how we read. And next thing you know, you have a whole bunch of autonomous leaders
0:04:07 emerge out throughout the company. Something I want to advocate for when people think about
0:04:12 distributed work is the types of communication channels that they need. Basically, if you’re
0:04:17 going to optimize for the fullness, you’re going to want three. But if you have to choose one,
0:04:23 choose async memorialized communication. So we will always have email. That will never go away.
0:04:27 And I don’t consider that part of the tool discussion. That just is. However, when I say
0:04:30 async memorialized, I mean things where you basically ratified decisions and threads and
0:04:35 you’ve locked them and they will stay there as your institutional memory. So that’s the first.
0:04:41 I think the second one is video calls. And then the third one, which is the one that most people
0:04:47 reach to first is something like Slack or IRC or any of those things. I think that should be your
0:04:52 third. That should be the optimization that you have and you never treat it as institutional memory.
0:04:55 I don’t think those tools are set up to be institutional memory. Backscroll is the worst
0:05:02 way to go find out the context to a conversation. I do think that you can use it for in real time,
0:05:07 slight communication between a couple of different people. But then once the decision’s made in that,
0:05:10 it should be ratified and memorialized in a different system.
0:05:15 I agree completely with the memorializing sort of in situ and the tool that you’re using. So if
0:05:19 it’s code and product decisions, maybe that’s in GitHub. And one of the things that’s been
0:05:25 interesting to observe over the last few years is that many, many tools, especially SaaS tools,
0:05:30 they’ve sort of been including more and more multiplayer support. But what we’ve observed is
0:05:34 that they’re actually adding more and more of the collaboration into the tool themselves. So for
0:05:41 instance, people that use Figma, they are commenting and talking about the Figma files directly in Figma.
0:05:46 Now, as it moves over to marketing and product marketing and other people in the organization,
0:05:50 away from the designer and into engineering and other aspects, all that comment history of like,
0:05:54 how did the design end up the way it is? Or why did we make this decision the way we did?
0:05:59 Is in that file and it stays there forever. And that’s super powerful. The downside is people
0:06:04 sort of have more places to look for commentary and things. The good thing is that the metadata is
0:06:10 now sticking around with the content itself, which is quite powerful. One comment about Slack that
0:06:15 I’ve observed with the companies that I work with that are distributed is that Slack ends up becoming
0:06:20 a notification mechanism for them. So, you know, obviously it has that water cooler like capability
0:06:25 that a lot of people use to talk about things real time and sort of synchronously. But it also
0:06:30 takes on this notification mode where all these tools where things are being memorialized or
0:06:35 being discussed sort of in situ, then get sort of centralized back into Slack for people that are
0:06:39 sort of just boy years and want to know what’s happening but aren’t maybe engaged in the day to
0:06:45 day. I have seen the exact same thing, which is Slack in particular is it’s a pinch point,
0:06:50 essentially every other system flows into it, updates it, then goes back out. And then people
0:06:56 stream all that stuff together. Just as basically GitHub and Slack can become platforms for exchange
0:07:02 of information, I imagine you’re going to see a lot of new tools emerge and basically say, okay,
0:07:08 hit this button, it’ll sync to X, Y or Z system. In fact, this may seem like a subtle point,
0:07:13 but turns out to be an important point that collaborating and discussing and working through
0:07:18 something in one tool is not the same as memorializing a decision in a way that
0:07:21 the organization can sort of keep that memory of it and make it easily accessible.
0:07:27 When you memorialize a decision, it only counts if everyone else knows about it. The other thing
0:07:31 is when things start to get memorialized, it allows new people to get up to speed more quickly. And so
0:07:36 that ends up being sort of a tailwind that just benefits all future people that join the company.
0:07:42 And so as these tools are creating sort of the collaboration inside the individual tools as
0:07:48 opposed to outside the tools, I do think it actually becomes more important to define where
0:07:53 decisions get memorialized, whether it’s in sort of like a wiki page inside of GitHub or
0:08:00 Confluence or Notion or whatever it is. So my approach has typically been you can choose Y tool
0:08:06 or Z tool if you want to do your local development practices, whether it’s tracking story points
0:08:11 or using Kanban style or so if you’re choosing Trello or Asana or something like that and you
0:08:14 want to track your stuff locally for your team there and you want to experiment with it. I’m
0:08:19 not going to stop you from doing that. However, the organizational institutional memory is going to
0:08:24 live in X. You can do yours locally and then you have to actually memorialize decisions or all those
0:08:29 things in the system of record. And that’s how we’re going to know what’s going on. I find that
0:08:34 works better as organizations get larger. If you’re smaller, I usually just say we’re all going to
0:08:39 live in X. Sorry, we’re not going to entertain other tools at this point. When people ask me about
0:08:44 habits of remote companies and I said, well, they’re just really habits of great companies.
0:08:50 And I want everybody who logs in in the morning to know exactly what they’re working on, why and
0:08:55 for whom. And that starts at the top and starts with context communication. And then you have to
0:08:59 have the swing all the way back up is are we actually doing it? Are we making progress? Are
0:09:03 things actually advancing? And the mechanisms by which you achieve those things are going to make
0:09:09 you a great company, distributed or co-located. And if you think about another aspect of leadership
0:09:14 in general is everybody in the organization is making the set of decisions that only they can
0:09:18 make. So if you’re the CEO, there’s a set of decisions that nobody else in the company can
0:09:23 make only you can make them. So you do, you make those decisions. And then every other decision
0:09:27 that somebody else can make is delegated to that person and et cetera, et cetera, all the way down.
0:09:33 But that only works if you are transparent enough and sophisticated enough with your
0:09:37 mechanisms to say, I’m going to make these decisions. Here’s our goals. Here’s our
0:09:40 measures. Here’s how we’re going to measure progress, all that sort of stuff at each one
0:09:45 of those levels. I paid attention to a project on GitHub called Home Assistant. And they’ve now
0:09:51 incorporated a ton of sort of automation, not just around is this code fitting a certain code
0:09:56 formatting policy, but also they now have these mechanisms like does this fit with our values?
0:10:00 And somebody has to sign off on that for the issue to get closed. Does this support all the
0:10:05 internationalization that we’re looking to do? And somebody who’s approved elsewhere quote unquote
0:10:09 in the organization has to sign off on that. And the issues get sort of redirected. And there’s a
0:10:14 bunch of bots that coordinate and facilitate all this. And so the mechanism of working in a
0:10:19 distributed way is being facilitated by technology in a way that I’ve never seen before. And I really
0:10:24 think that was driven by the open source community as you embrace source code revisionary kind of
0:10:28 paradigms and apply the dollars on the things that anybody can do a poll request and have it
0:10:34 reviewed and evaluated and considered or being able to fork the organizational code to improve it.
0:10:40 That feels like it breaks down these weird hidden power structures that sometimes can
0:10:44 emerge in very old or very large established corporations, which is basically like, hey,
0:10:48 we need to go do this. You’re like, oh, the only person who touches that bit of code or can make
0:10:53 that decision is X person. And you have to talk to the right people to go find out who that person
0:10:59 is. Instead, open source has basically built mechanisms for that. When you put an issue into
0:11:05 GitHub and it has this keyword into it, somebody automatically gets tagged who is responsible
0:11:09 for that. And then they’re notified and they have to go comment on the issue and have
0:11:15 discussion about it. I think the feedback loop mechanism of lowercase a agile is incredibly
0:11:20 important, which is it’s the only time that I’m going to find out how things are going is when
0:11:26 it’s released or at the end of a sprint or at the end of a quarter, the classic project management
0:11:31 red, green, yellow type of approach as probably not the right mechanism. So I’ve just introduced
0:11:37 a couple of different things. One is for remote stuff, obviously you want to have somebody responsible
0:11:42 for updating the progress of various projects, products or whatever across the organization,
0:11:47 particularly the ones that are interdependent upon other teams or organizations. The other is I like
0:11:54 to do review meetings now. And in those meetings, I’m getting updated on very large or ambitious
0:12:00 projects on a regular basis. And particularly I’m trying to edit in that meeting, not author.
0:12:05 So it’s really a time for me to either course correct or get updates. I also like the organization
0:12:12 to feel the same sort of progress. So I ask teams to give regular updates to the organization on a
0:12:19 weekly, bi-weekly or at most at absolute most monthly basis. And I ask them to do a write up
0:12:24 with a video demo of the progress that they made in that time period. I have not personally found
0:12:30 things like burn down charts to be that effective. I just think we as an industry are not great about
0:12:34 estimating progress. I agree. Things like burn down charts are sort of useless because they get
0:12:39 gamed, but just demoing very, very regularly creates those loops and those opportunities for
0:12:44 people to course correct and edit. I personally love demos. I get excited by them. I think the
0:12:47 organization gets excited by them. And I think there’s a forcing function. If you do it on a
0:12:52 regular time interval, you have to think a little bit of like, okay, how are we actually going to
0:12:59 showcase progress here? Because one of the worst habits I’ve seen of very large engineering teams
0:13:03 or large scale engineering efforts, large scale, maybe infrared ones or operating system ones is
0:13:10 to say we can’t actually demonstrate progress for months, which we all know from the way that
0:13:16 software can be written is that is never true. So the job now becomes how do you actually demonstrate
0:13:21 progress in a shorter amount of time? But also this is not just an engineering thing. It can be all
0:13:24 parts of the organization. So when people are listening to this podcast and they’re saying,
0:13:29 oh, that’s great for demoing some code or, you know, a new UI somebody built. It’s absolutely not
0:13:34 just for that. It’s for anybody in the organization. I’ve been lucky that superhuman Rahul Vora, the
0:13:39 CEO, lets me sometimes sit in on like their Friday demo sessions and afternoon demo sessions.
0:13:42 And they do a really nice job because you’ll even see the recruiting team come in and say,
0:13:46 hey, look, this is how we’re changing how we’re doing recruiting or this is how we’re changing
0:13:50 how we’re sourcing. I’ve seen the marketing team talk about how they’re thinking about some of the
0:13:55 growth of the business. So it really goes way beyond engineering. It’s a moment about celebrating
0:14:00 wins. It’s not a moment of critiquing. There’s other venues and channels and much more smaller
0:14:05 settings and directed settings. That is an incredibly important point. You know, we’re
0:14:10 currently in an environment where you can’t meet candidates face to face. And then just long term,
0:14:13 there’s going to be lots of situations where you’re interviewing wanting to hire somebody
0:14:17 who just it’s impractical to meet face to face. Are there things that you do to make sure that
0:14:22 you’re really doing a good job interviewing? Is this a non-issue? Is this an actual issue?
0:14:28 Yeah. Given that I’ve been doing this for a while, I find that it’s not so much about remote or
0:14:31 co-located. It’s actually about interviewing. And we’re just not that great at interviewing in
0:14:36 general. But for some reason, we have a better sense when we’re able to see someone in person.
0:14:41 So what I really say is it’s incredibly important that if you’re going to communicate to somebody,
0:14:46 you do it over video. And you can read the body language and you can have all those conversations.
0:14:51 And I’ve learned over the last 12 years of doing this to raise my voice at the right time,
0:14:56 to use the right intonation. If I’m on video, I overexpress with my eyes and my face and things
0:15:00 of that nature because you’re trying to send signals. And then I guess the other piece is
0:15:06 really, do you have to do things to make sure that the person being interviewed is prepared for a
0:15:10 remote style interview, maybe new for a lot of people? I just talked to a couple of different
0:15:16 companies and they had literally never talked to a candidate outside their physical building
0:15:23 until just recently. And here’s just some really tactical recommendations for this. One is, I think
0:15:28 from an interview process perspective, have a singular person who owns that candidate’s interview
0:15:32 process. If it has to be handed off to multiple people, it’s going to fall apart and get worse.
0:15:38 Once you have that ownership, identify what the days will look like, how it’s going to get done,
0:15:44 try to get to the conclusion as quick as possible. I think candidate experience from, “Hey, I’m
0:15:50 interested to yes/no decision.” If that lags, it’s an indicator of poor process. One small
0:15:57 recommendation, I see that a lot of companies do one-on-one interviews with candidates. And over
0:16:04 my career, I’ve really disliked one-on-one interviews except for executives. And I’ve
0:16:09 recommended that you at least go in two-to-one with candidates in interviews. And the reason why
0:16:14 is that if you look at interview processes, it’s about context sharing your experience with the
0:16:18 candidate with others. And this weird mechanism happens where you’re trying to say, “I liked
0:16:23 them. I did not like them for these reasons.” But you can’t corroborate any of that with somebody
0:16:28 else. I’ve always found that when you have at least two people from the same company interviewing
0:16:32 a candidate, it can go back and forth. It changes the interview process entirely and you get much
0:16:36 better results. Let’s also not forget the last most important part of interviewing, which is
0:16:40 references, which obviously can happen in a remote environment and often do. Maybe with the
0:16:43 exception of engineers where you really want to make sure that they can code and write code,
0:16:48 I think references are as important and oftentimes more important than even the interviewing.
0:16:51 There’s a lot of people that are great at interviewing that are just not great at doing
0:16:55 work and working with. And there’s actually a lot of people that are bad at interviewing that are
0:16:59 great to work with and do great work. And so references are often the way that you cut through
0:17:03 that and find out the truth. It’s less important for individual contributors,
0:17:08 but for people that are going to manage other people, the references have to be glowing.
0:17:12 I do like to give projects to people, very lightweight ones, but the project is not to critique
0:17:18 the style or code necessarily. It’s really to get to the way that they’re thinking, particularly
0:17:24 when it comes to senior folks. However, I still think it’s more about interviewing and we’re just
0:17:30 not that great at the ability to identify good talent. When you’re hiring engineers, I tend to
0:17:34 ask them to break down problems. I ask them to see if they can help me work through why they
0:17:41 made decisions. I like to see how much ownership they’ve taken on projects. And I don’t mean ownership
0:17:47 that they were given and then ran with ownership that they’ve more seen inherently implicitly and
0:17:52 then made explicit via their actions. We’re in the midst of a crisis that nobody who is currently
0:17:56 living has really ever experienced before. It’s both acute from a health standpoint.
0:18:01 It’s a massive shift for people economically and creates a lot of uncertainty. It may be a
0:18:05 lot of shifts for people in terms of the viability of the company that they work for or the changes
0:18:09 that a company might work for to survive. And it’s very hard to know what people’s personal
0:18:15 situations are. I got an email actually yesterday that there’s two executives and both of their
0:18:21 spouses actually work in emergency rooms. And I can’t really even imagine what it’s like in that
0:18:26 household where they have kids at home and their spouses on the front lines in the emergency rooms.
0:18:30 And so I think in this moment in time, you just have to give people a lot more latitude, a lot
0:18:35 more leniency, a lot more flexibility. If they’re new to the organization, you have to give them
0:18:39 forward credit for what you expect out of them when things are normal. And if they’ve been with you
0:18:44 for a long time, you certainly owe them the latitude and ability to be flexible and different
0:18:49 people cope in different ways. And I just think that managers and leaders who have been in a mode
0:18:53 of we got to grow or we got to hit our deadlines, we got to hit our numbers, all that stuff is on
0:18:57 pause right now. This is a good moment for lots of people to figure out, are they spending their
0:19:02 time on the right things? You’ve had this really, really focused growth mindset. You’ve been focused
0:19:07 on growth that is the only priority. And you can use this moment in time to say, is our product
0:19:12 roadmap the right product roadmap? If we know that the top line may not grow the way we had
0:19:16 expected it to, would we change our product priorities? Would we build something more exciting
0:19:21 or interesting or different? So just really pause and think about what are the priorities
0:19:25 for me while my day job is investor and meeting with startups and companies. And I’m doing a lot
0:19:29 of that because we’re open for business and we’re investing. I’ve spent a lot more time as both a
0:19:34 career counselor to friends, a lot more people reaching out, trying to figure out how do they
0:19:38 lead their teams. Yeah, this is not business as usual. We are not living in business as usual
0:19:43 times. As the venture community talks about, Hey, think about this time period over the next
0:19:50 whatever it is, year 18 months or 24 months. And what’s the scenario look like if your ARR
0:19:57 drops by a third or two thirds or whatever and start modeling those? Well, right now in unprecedented
0:20:02 times, I say the same thing about productivity. What does it look like when you’re at two thirds
0:20:08 productivity or 50% or then a one third of what you expected and start planning for some of those
0:20:16 scenarios? You know, the way that I’ve approached this at work is to say, one, take care of yourself.
0:20:23 Two, let’s talk openly about what’s possible and what’s not possible right now. And three,
0:20:27 understand why the work would or wouldn’t be important in the context of everything else
0:20:33 that’s happening. And I personally have a career and philosophical principle about what I do is
0:20:38 I try to run to the point of highest leverage. And I’m in a very interesting privileged position
0:20:43 at GitHub is that I run the office of the CTO. It’s a small staff at this point. And we’ve entirely
0:20:47 focused on COVID efforts. We’ve got some data scientists, we’ve got engineers, and they’ve kind
0:20:50 of said, you know, what right now is pause on some of the more exploratory work that we’re doing,
0:20:58 we want to go lend some effort and hands over to these in a personal story. But in 2009, my wife,
0:21:04 she’s a GP, a doctor, and she was at Mayo Clinic, and she got sick. We think it was H1N1 with
0:21:10 pneumonia, and she was intubated in the ICU for six days and in the hospital for 10. And here we
0:21:15 are, we’re finding ourselves in an epidemic that looks a lot like what she got. And we were young
0:21:20 and we had one kid and we realized that our life had been an autopilot for a little while. We didn’t
0:21:25 like it. And it caused us to have that moment where we reflected and said, are we actually doing
0:21:32 what we want to be doing? And we changed our life a little bit, we went to Australia. And I would
0:21:38 encourage everyone to use these moments to personally reflect on if your life is on autopilot.
0:21:39 Thanks, Jason. Thank you.

0:00:04 – Hi and welcome to the A16Z podcast, I’m Hannah.
0:00:07 We’re talking today about what is happening to oncology
0:00:09 and to patients going through cancer treatment
0:00:11 during the outbreak of coronavirus.
0:00:12 How treatment is affected,
0:00:15 what kind of clinical decisions oncologists
0:00:17 are having to make, what kind of new tools
0:00:20 and what happens to oncology as a whole
0:00:23 when it’s forced in this moment to go so virtual.
0:00:25 Joining myself and Venita Agarwalha,
0:00:28 physician and general partner at A16Z
0:00:30 are Dr. Bobby Green, community oncologist
0:00:33 and chief medical officer at Flatiron Health,
0:00:34 who is the first boy seal here,
0:00:38 and Dr. Sumit Shah, oncologist and head of digital health
0:00:40 at the Stanford Cancer Center.
0:00:43 We’re here today because coronavirus is now disrupting
0:00:46 the entire healthcare system,
0:00:48 not just because of the burden of dealing
0:00:49 with the actual disease itself,
0:00:50 but because of everything else
0:00:52 that’s had to grind to a halt.
0:00:56 One of those areas where we really worry about things
0:00:57 coming to a total stop like that
0:00:59 is of course cancer treatment,
0:01:01 which can often feel like a race against the clock,
0:01:04 even under the best conditions.
0:01:07 So can we just start by talking about the biggest issues
0:01:09 that your cancer patients are facing right now?
0:01:12 – I’ve been sort of, I guess, surprised
0:01:15 how much more resilience I’ve actually seen
0:01:17 among a lot of my patients.
0:01:19 And maybe it’s just because they’ve lived
0:01:20 through so much uncertainty
0:01:23 and gone through so much as part of their diagnoses
0:01:26 that this is just one more thing.
0:01:29 But I do think people have been remarkably resilient
0:01:33 and in fact, in some ways a lot more so
0:01:35 and maybe more laid back about this
0:01:38 than some of my healthy friends and colleagues
0:01:41 who haven’t had to deal with any health crisis.
0:01:42 There’s been a little bit of confusion
0:01:45 about what does it mean for my treatments
0:01:46 and uncertainty about timing,
0:01:47 ’cause I think none of us really know
0:01:48 how long is this gonna last
0:01:51 and how long is the world gonna look like this.
0:01:54 And then also the other thing I’ve sort of seen,
0:01:57 and fortunately I haven’t had any patients
0:01:59 that I know of who have been laid off.
0:02:03 But I think uncertainty about jobs and people,
0:02:04 most people have their health insurance,
0:02:07 as we all know, if they’re not on Medicare
0:02:09 through their place of employment.
0:02:11 And that’s a concern as well.
0:02:13 – Yeah, I totally echo what you said about resilience, Bobby.
0:02:16 I had a patient last week who told me,
0:02:19 who told our team that they can empathize
0:02:21 with what everybody’s going through
0:02:23 because it’s exactly what they’ve felt like
0:02:24 every time they had chemotherapy,
0:02:27 that they’re suddenly susceptible to infection
0:02:29 and that they have to be careful.
0:02:31 And that was like their dominant reflection
0:02:33 is that I understand why everyone is afraid.
0:02:36 And I’ve felt that fear myself,
0:02:41 which was an incredibly sort of empathic
0:02:43 and resilient thing for somebody to say
0:02:45 you could have been worried entirely about their own care.
0:02:47 – It feels like from the outside,
0:02:49 you’d imagine that would almost double
0:02:51 that you’d get sort of extra doses of that fear
0:02:54 instead of almost being like inoculated against it
0:02:55 because you’ve gotten so used to it.
0:02:57 That’s really inspiring.
0:02:58 – Yeah, I will agree with you guys
0:03:00 that I think most of our patients
0:03:01 have been incredibly resilient
0:03:03 and understanding of the current time.
0:03:05 But I’ve also had a fair number of patients tell me,
0:03:08 Dr. Shaw, my cancer can’t shelter in place.
0:03:09 What do I do right now?
0:03:12 Which is a very poignant point as well.
0:03:14 So I think a lot of the patients are wondering
0:03:16 about the timing of chemotherapy.
0:03:17 Should I initiate chemotherapy?
0:03:19 Should I delay chemotherapy?
0:03:21 What are the risks of doing so?
0:03:23 Am I going to put my body at higher jeopardy
0:03:26 for becoming immunocompromised?
0:03:28 Or am I going to leave my body at higher risk
0:03:30 for this coming back in the future?
0:03:33 So these are very difficult decisions to kind of make.
0:03:34 And we don’t have a whole lot of data
0:03:37 to help us support that decision-making capacity.
0:03:39 So a lot of this is done on a case-by-case basis.
0:03:44 – I think a lot of providers are discussing the Lancet paper
0:03:49 about a relatively small cohort of 18 cancer patients
0:03:52 who got infected with coronavirus across China
0:03:56 and looking at outcomes and results there.
0:03:58 Our data that are emerging like this factoring
0:04:01 into your decision-making and into the decisions
0:04:03 that cancer centers across the country are making,
0:04:05 that data was of course limited,
0:04:08 but it did suggest increased morbidity,
0:04:11 even among patients who were not actively immunosuppressed.
0:04:14 How are you guys thinking about data like this
0:04:18 and also about generating data within your centers?
0:04:20 – We do think that cancer patients
0:04:24 are probably more susceptible to viruses in general.
0:04:26 They’re also more likely to get more serious complications.
0:04:29 So I think we do have to be very ginger
0:04:31 about what types of treatments we’re giving our patients.
0:04:32 And we know that for patients
0:04:35 who are at higher risk for complications,
0:04:37 we may be able to give them more support
0:04:40 in terms of medications that can maybe decrease the risk
0:04:43 of their immune system being compromised.
0:04:45 So I think it is a valid concern.
0:04:47 And I think that most of our treatments
0:04:52 can affect patients in terms of their immune system.
0:04:53 And we have to be very cognizant
0:04:56 of how we’re treating these patients.
0:04:59 – So let’s get into those kinds of clinical decisions
0:05:01 that patients are facing right now.
0:05:04 There must be an enormous amount of gray area.
0:05:07 Is there any kind of broad framework in place
0:05:10 or is it really on a case-by-case basis?
0:05:12 – We broadly categorize treatments
0:05:15 in terms of curative intent versus palliative intent.
0:05:18 Curative intent meaning that you’re initiating chemotherapy
0:05:21 or immunotherapy or any surgery, for instance,
0:05:24 to be able to cure that patient from that cancer.
0:05:27 Palliative is working as thinking about treatments
0:05:30 more in terms of improving patient symptoms
0:05:31 or helping them to live longer.
0:05:34 But we know that for the majority of these patients,
0:05:37 they won’t be able to be cured from this condition.
0:05:39 So there have been some larger frameworks
0:05:42 about saying that for patients
0:05:45 who are undergoing curative intent chemotherapy,
0:05:46 that we should go for with that
0:05:47 because the risk of recurrence
0:05:50 could be causing a great deal of harm for these patients.
0:05:52 However, for palliative patients,
0:05:54 maybe we should be a little bit more ginger
0:05:55 about starting chemotherapy,
0:05:59 which may dampen their immune system in the short run
0:06:02 while we’re facing a higher risk from coronavirus.
0:06:04 – I think almost everything we do in oncology
0:06:09 is looking at risks and benefits of various treatments.
0:06:11 And I think most decisions we make
0:06:12 are based on that calculation.
0:06:14 And what’s so interesting now
0:06:17 is you have a whole new set of risks with COVID, right?
0:06:19 Some of that’s like you take a risk
0:06:21 by walking out of your house in the morning,
0:06:23 which certainly wasn’t the case.
0:06:25 You take a risk by walking into clinic.
0:06:27 We take for granted that if someone gets sick
0:06:29 and needs to go to the hospital,
0:06:30 we have a good hospital for them,
0:06:33 but are the hospitals gonna be overloaded in three weeks?
0:06:34 If someone needs an ICU bed
0:06:36 or needed a ventilator for something,
0:06:38 is that gonna be available?
0:06:39 So, I’ve spent a lot of time
0:06:42 just trying to think about these risks,
0:06:44 a lot of which are very uncertain
0:06:46 and how those play into those treatment decisions
0:06:49 about starting someone on adjuvant therapy
0:06:52 or starting someone on palliative therapy.
0:06:54 – Several of the guideline organizations
0:06:56 that guide cancer care in this country
0:07:01 have, like ASCO, have put out relatively broad statements
0:07:06 outlining the role of coronavirus
0:07:09 factoring into decision-making
0:07:12 with respect to both curative intent
0:07:13 and palliative intent chemotherapy
0:07:16 with respect to timing of bone marrow transplantation,
0:07:20 for example, recommending that such a procedure
0:07:22 potentially be delayed for patients,
0:07:24 thinking about adjuvant therapy,
0:07:27 but most of the guidance has left a lot
0:07:28 of room for interpretation.
0:07:32 And so, I’m curious, how are you personalizing that guidance?
0:07:34 – I think ASCO has done a really nice job
0:07:38 of responding to this and putting out information,
0:07:41 but at the end of the day, at least from where I sit,
0:07:43 most of the recommendations have been
0:07:46 use your clinical judgment and take into account,
0:07:50 again, going back to that risk-benefit framework.
0:07:52 So, I’ll give you a couple examples.
0:07:55 I saw someone this week
0:07:58 who was an early-stage lung cancer patient
0:08:02 who I had seen at the end of February,
0:08:04 and we planned to give him adjuvant chemotherapy,
0:08:08 and we just had our second discussion today,
0:08:10 and we went from the conversation
0:08:12 about what’s the benefit of adjuvant chemotherapy
0:08:15 and should we go through with this in February
0:08:18 before any of us were really thinking of coronavirus,
0:08:21 and then we re-had the conversation twice
0:08:23 over the last week and a half,
0:08:24 and the framework shifted, right?
0:08:27 Like, it made sense to do at the end of February,
0:08:29 both to me and to the patient,
0:08:30 and it no longer made sense,
0:08:33 because the absolute benefit that we’re gonna see for this,
0:08:35 when you compound it with the risk of,
0:08:37 he’s gonna need to come into the office frequently,
0:08:40 it’s gonna be harder and harder for him to stay isolated,
0:08:43 what happens if he has a side effect or toxicity
0:08:44 that puts him in the hospital,
0:08:47 what are the additional risk factors for him
0:08:50 being immunosuppressed because he’s having chemotherapy,
0:08:51 it just didn’t make sense to do that.
0:08:55 Now, he was a stage 1B lung cancer, a situation,
0:08:57 and he was high risk, but it’s a situation
0:09:00 in which you would go back and forth to begin with
0:09:03 about whether to give adjuvant chemotherapy.
0:09:06 I’ve had patients on bone modifying agents,
0:09:08 these are drugs that are supportive drugs
0:09:12 that are typically used in a variety of diseases,
0:09:14 but while there’s benefit, the benefit accrues
0:09:17 over a long period of time, and I’ve pushed those off
0:09:19 because, again, I don’t know that there’s
0:09:21 a right or wrong answer, but in my judgment,
0:09:24 the risk of them stepping out of their house
0:09:27 and coming into the clinic probably doesn’t make sense.
0:09:29 – Most practitioners would probably continue
0:09:31 with the chemotherapy for patients who are young
0:09:34 and otherwise fit and doing well with treatment.
0:09:37 For patients where you’ve had a deep remission,
0:09:40 now one year, two years beyond treatment,
0:09:42 for a lot of those patients, we can probably
0:09:45 safely stop treatment at that time,
0:09:46 and hopefully the patients will continue
0:09:48 to remain in remission.
0:09:49 So it really depends on where the patient is
0:09:52 and what kind of response they’ve had,
0:09:54 but it also depends on the type of cancer they have.
0:09:56 We know that there is a significant amount
0:09:58 of heterogeneity between cancers,
0:10:00 so not all cancers are created equal by any means.
0:10:02 So in each situation, we have to kind of do
0:10:04 the risk-benefit calculus and make sure
0:10:06 that it’s in the best interest of the patient.
0:10:09 – Are questions like this coming up at our tumor boards,
0:10:13 who is the group that you’re able to engage
0:10:16 in real time on such difficult decisions
0:10:18 for individual patients?
0:10:20 – Well, Vinita, as one of my seven Twitter followers
0:10:22 you may have seen…
0:10:26 – I saw that you crowdsourced that, well done.
0:10:30 – Yeah, I crowdsourced the early stage lung cancer question
0:10:31 today.
0:10:34 We have a multi-disciplinary lung tumor board, Vinita,
0:10:36 where these questions have come up.
0:10:39 A lot of curbsiding other docs, that’s been my experience.
0:10:40 I think what’s really interesting
0:10:43 about the problems we’re facing is, you know,
0:10:46 there’s sort of the art and the science of medicine.
0:10:48 And this is really one of those circumstances
0:10:51 that the art of medicine and judgment
0:10:55 and how to apply sort of knowledge about data
0:11:00 to great areas of uncertainty really comes into play.
0:11:02 And it’s been intellectually challenging to do so.
0:11:04 – I’m glad you brought up Twitter because I’m wondering,
0:11:08 is that a viable kind of tool for you guys
0:11:12 for crowdsourcing, for even anecdotal data advice
0:11:14 decision-making in this area?
0:11:17 – I’ve personally found the discussions on Twitter
0:11:21 about this to be really helpful and really informative.
0:11:24 So, to me, yes, you know, you can only give,
0:11:26 you have to be a little bit more general
0:11:29 than you would like to be for PHI reasons, obviously.
0:11:31 But I find it very useful.
0:11:33 – I actually think that Twitter is probably the best source
0:11:34 of medical information right now,
0:11:37 as an academic and an oncologist.
0:11:40 The majority of my data that I’m actually receiving,
0:11:43 I’m receiving in real time from my Twitter feed,
0:11:46 as opposed to waiting for publications to come out.
0:11:48 So it’s actually been very, very helpful
0:11:50 to have access to Twitter.
0:11:52 And it’s just been a tremendous communication tool
0:11:53 from experts around the country
0:11:55 and in the world in general.
0:11:57 We’ve been using our tumor boards
0:11:59 to discuss a lot of these cases
0:12:02 as we alluded to earlier, that the data is very gray.
0:12:04 We really don’t have a lot of information
0:12:06 to base these decisions on.
0:12:07 There have been consensus statements
0:12:10 that have been put out into large publications
0:12:12 from thought leaders across the world.
0:12:15 And we’ve been using those as a framework
0:12:18 by which to make our decisions at Stanford as well.
0:12:21 But to still very much a gray area is important
0:12:23 that we have employees share decision-making
0:12:26 with our patients to make sure that they’re also,
0:12:28 feel that they’re a part of this conversation.
0:12:31 – It strikes me that a lot of the decisions
0:12:32 that you’re talking about making,
0:12:35 you already are at a place of understanding
0:12:37 to a large degree what you’re dealing with,
0:12:39 what kind of cancer, how it tends to behave.
0:12:43 What about the patients that just found a lump
0:12:46 and were coming to you for like the very first step?
0:12:48 How do you, what are the guidelines there?
0:12:52 Like begin, wait it out for two weeks, you know?
0:12:54 – The answer to this in general will really vary
0:12:56 depending on where you are in the country,
0:12:58 what kind of health system you’re in right now
0:12:59 and the where we are in the pandemic as well.
0:13:01 ‘Cause we know that resources are shifting
0:13:06 on a daily basis based on local prevalence rates.
0:13:09 So while I think it’s true that most non-essential
0:13:11 procedures and surgeries are being postponed,
0:13:15 the suspicion for cancer does increase
0:13:18 the prioritization of certain scans or procedures.
0:13:22 So for a newly diagnosed or a new breast mass
0:13:25 that you may feel in the shower, for instance,
0:13:28 that would actually take prioritization to have that worked up.
0:13:32 For men who have an increase in PSA
0:13:34 over a slow period of time,
0:13:37 they probably don’t need a prostate biopsy right away.
0:13:40 So it really will vary on the type of cancer
0:13:42 and the type of patient as well.
0:13:45 But in general, patients with cancer suspicion
0:13:48 will probably be at a higher prioritization
0:13:50 for getting their treatment done.
0:13:54 – So at Flatiron Health, among our practices
0:13:57 which use our electronic health record, ONCO EMR,
0:14:00 we’re able to track patient volume.
0:14:05 And we saw last week a 22% drop in office visits
0:14:09 across the network of practices
0:14:13 that also included a 16% drop in visits
0:14:16 related to chemotherapy all in the past week.
0:14:19 I think those are partially shock of the system.
0:14:21 Let’s reevaluate and see who needs to come in
0:14:23 and who doesn’t, but it was pretty impactful.
0:14:26 I mean, our numbers across our network
0:14:27 are very, very consistent.
0:14:30 And then there was just this big drop last week.
0:14:31 I think a lot of that is just gonna be,
0:14:33 whoa, hold on, we have to figure out what we’re doing.
0:14:35 I don’t think necessarily you’re gonna see
0:14:37 that much drop in chemotherapy,
0:14:39 but I think there’s gonna be a lot of really interesting data
0:14:41 that’s gonna come out of that
0:14:45 to try to understand how this impacted cancer outcomes.
0:14:47 – And to your point, Bobby,
0:14:51 about kind of this being a sort of unprecedented shock
0:14:53 to the oncology care system
0:14:57 for us to really see what kind of an impact happens
0:14:59 on visit volume and treatment volume.
0:15:03 And some of that may even extend to outcomes
0:15:08 and it may be a really sort of fine grain sensitivity analysis
0:15:12 that we sort of have an opportunity to later look back at
0:15:14 and say, well, what really happened
0:15:18 if surgery was delayed by this period of time
0:15:19 and how did outcomes change?
0:15:21 And what really happened if adjuvant therapy
0:15:23 was delayed by this period of time?
0:15:25 How did outcomes change?
0:15:28 I think, I hope that to some extent,
0:15:31 some of that may be a silver lining
0:15:34 in terms of learning from this crisis.
0:15:35 – The other sort of interesting thing
0:15:39 that’s come out of this is it makes you spend a lot of time
0:15:41 thinking how much you really need,
0:15:45 things that you thought you really needed, right?
0:15:49 Like, usually this patient comes in and gets blood drawn
0:15:52 and yeah, I’m gonna do this telemedicine visit
0:15:53 and we’re gonna skip the blood draw
0:15:55 and I find myself saying, don’t worry,
0:15:57 we can do it again in three to six months,
0:15:58 which sort of raises the question,
0:16:01 do they really need that blood draw to begin with?
0:16:04 – Yeah, or that scan or that physical exam.
0:16:07 We’ve all felt sort of the sadness
0:16:11 when we hear a patient describe their five-hour drive
0:16:14 and then their hour-long wait in the waiting room
0:16:16 and then their five-hour drive back
0:16:18 because it’s crazy to try to get a hotel room
0:16:19 on the day of their visit
0:16:22 and this kind of logistical nightmare
0:16:26 that many patients undergo in order to seek cancer care
0:16:30 or seek second opinions or seek clinical trial evaluation
0:16:32 and I think we’ve all wondered, well,
0:16:35 could some of this be happening more efficiently
0:16:37 and in a more patient-centric way
0:16:40 if we were to embrace technology in various ways
0:16:44 and sometimes a crisis is an opportunity for us
0:16:46 to embrace that tech stack
0:16:48 and I think we’re all seeing it happen.
0:16:51 I’ve been sort of floored and amazed
0:16:54 at how much of the Stanford oncology clinics
0:16:58 are now sort of operating in the telemedicine sphere.
0:17:00 I’d love to hear how you guys are managing this.
0:17:03 Which patients are you bringing into clinic?
0:17:05 Which are you managing via telemedicine?
0:17:07 – I mean, I think the whole remote care
0:17:10 telemedicine virtual medicine,
0:17:11 I did my first telemedicine session
0:17:14 and I finished it and my thought was,
0:17:15 where have you been all my life?
0:17:19 – Your patient might have had the same thoughts.
0:17:22 – Yeah, right, like that was easy.
0:17:25 The easy answer is routine follow-up.
0:17:27 So the patients who were coming in for routine follow-ups
0:17:29 who didn’t want to reschedule or push back,
0:17:31 I’ve done those over telemedicine
0:17:33 and that’s been relatively easy
0:17:36 and relatively straightforward.
0:17:38 I’ve also had a couple other patients
0:17:40 who weren’t routine follow-up
0:17:42 but I wanted to try to keep out of the office
0:17:44 thinking about a couple of patients, for example,
0:17:47 with a disease called chronic lymphocytic leukemia
0:17:49 who were on relatively new therapies,
0:17:51 were coming in to get their blood counts checked
0:17:53 so they weren’t just routine follow-ups
0:17:56 but given the changing circumstances,
0:17:59 I felt relatively comfortable doing a telemedicine visit,
0:18:00 making sure they’re okay
0:18:03 and pushing back their lab results for a few weeks
0:18:05 where you obviously can’t do it
0:18:07 is on people who need a treatment.
0:18:11 So people who need medicines to boost up their blood counts
0:18:14 to keep their blood counts from getting too low,
0:18:15 there are ways you can give those at home
0:18:18 but for patients who are getting it in the office,
0:18:21 sometimes it’s just not easy to get that quickly done
0:18:26 and then the regulations around using FaceTime and Skype
0:18:29 and other non-HIPAA compliant platforms
0:18:31 has been lifted at least for the time being.
0:18:33 So I’ve had a couple of circumstances
0:18:36 where I’ve just FaceTimed patients to do this.
0:18:38 – We’ve really dramatically scaled up
0:18:41 the virtual clinics in our clinics in the last two weeks
0:18:42 which is quite ironic
0:18:44 ’cause we’ve actually had virtual capabilities
0:18:45 for over a year now
0:18:48 but it’s literally taking a pandemic to do this.
0:18:49 The utilization of virtual
0:18:52 was around five to 10% of our clinic visits
0:18:53 over the past year
0:18:54 but now in the last week
0:18:56 is now greater than 60% of our visits
0:18:57 are actually now all virtual
0:18:59 which is quite extraordinary.
0:19:04 Yeah, and we have this alignment for the first time actually
0:19:05 where we actually have an alignment
0:19:08 between providers, patients and now even payers
0:19:13 as CMS changed their laws this past week
0:19:15 to allow for reimbursement for televisits
0:19:17 and virtual medicine.
0:19:21 So that’s really changed the landscape completely
0:19:22 and so now we’re seeing a huge rise
0:19:25 in our ability to deliver virtual care.
0:19:27 – So we had a podcast recently
0:19:30 where we talked quite a bit about using virtual medicine
0:19:32 and telemedicine tools for primary care
0:19:36 and sort of triaging symptoms from your home.
0:19:39 Are there particular pressure points
0:19:42 that you’re noticing from the specialist point of view
0:19:44 where things aren’t working so well,
0:19:45 where there are sticking points
0:19:47 or where the data flow gets messed up?
0:19:49 – Part of the thing in oncology
0:19:52 is you administer therapies to patients
0:19:53 and a lot of those therapies
0:19:57 it’s sort of hard to do in a remote setting.
0:20:00 I think from the oncology standpoint, that’s a barrier.
0:20:03 The other area which I’ve sort of seen
0:20:05 and again this has been in a limited experience
0:20:09 is there are often difficult discussions you have
0:20:10 and I think we’re all accustomed
0:20:14 to delivering those difficult discussions in person
0:20:16 and the ability to have physical contact
0:20:20 but I think that’s been a tough part of remote oncology.
0:20:21 – Wow, yeah.
0:20:24 – So much of oncology is an art form as you were saying
0:20:27 and it’s really our ability to connect to a patient
0:20:28 which makes oncology so special
0:20:30 and which is why so many of us went into this field
0:20:31 in the first place.
0:20:34 But it’s very difficult to do that virtually
0:20:36 as much as I would love an emoji.
0:20:38 I think an embrace after giving someone good news
0:20:41 is much more wanted by us
0:20:46 and we also know that the physical exam, as you were saying,
0:20:48 is also very limited for these virtual visits.
0:20:50 We do have some limitations
0:20:52 especially where the physical exam
0:20:53 can be a little bit more important
0:20:56 such as gynecological cancers
0:20:58 like endometrial cancer or cervical cancer.
0:20:59 – What could be better?
0:21:03 Like if you had to brainstorm a feature list
0:21:07 for the platforms that you’ve tried, what would help?
0:21:08 – On our current platform,
0:21:10 we can’t share screens nearly as easily
0:21:14 so I was trying to tell a patient about a lung nodule
0:21:16 and he wanted to see it actually on his CT scan.
0:21:18 So I actually had to take a mirror
0:21:22 and show him his lung scans
0:21:24 through the reflection on the mirror
0:21:26 which I thought was extraordinary in 2020
0:21:28 that we can’t do this quite yet.
0:21:30 – How about the sharing of information
0:21:33 like provider to provider or specialist to specialist?
0:21:34 That’s something that our partner Julie Yu
0:21:36 brought up on the last podcast
0:21:39 as being still not seamless with the data flow
0:21:41 of these telemedicine tools.
0:21:44 – Well, I mean, I’ll give you an example today
0:21:46 that happened to me as one of my patients
0:21:49 who I didn’t think I was gonna need to examine
0:21:51 said, “Oh, I have this thing on my back”
0:21:53 and turned around and tried to show me this thing
0:21:55 on his back and it was,
0:21:57 which as an aside, there have been a lot
0:22:00 of sort of comical technology related things
0:22:02 that have happened in the last week and a half too.
0:22:04 – Like what, like I can’t even,
0:22:06 would you put your phone on your cabinet
0:22:08 and like turn around and take your shirt off?
0:22:10 I can’t even really imagine how that.
0:22:12 – Well, like this was one trying to turn around
0:22:15 and show that my patient trying to show me a picture
0:22:17 of his lower back and calling his wife in
0:22:19 and not being able to see where he was.
0:22:22 Another patient who just wasn’t really used
0:22:24 to using the phone on his camera
0:22:27 and he kept putting his camera up against his ear
0:22:29 so I could see the inside of his ear.
0:22:33 But not that, you assume that everyone
0:22:35 in the world uses technology like you do
0:22:39 and you quickly find out that is not the case.
0:22:43 Like people who aren’t sure what to do with a hyperlink.
0:22:45 So that’s been sort of interesting,
0:22:47 but I’ve had some nice laughs with my patient
0:22:50 around this as well.
0:22:53 – So some of this are sort of growing pains
0:22:56 and rollout pains, but maybe in the future,
0:22:59 if a cancer patient was expecting a certain fraction
0:23:02 of their visits to be over telehealth,
0:23:05 maybe that adoption curve would look different.
0:23:08 – Yes, so I mean, I’ll just give you one example
0:23:09 for a feature that would have been nice.
0:23:10 So this patient who I just told you about
0:23:12 who had this thing on his back,
0:23:14 I couldn’t get a clear enough picture on the video.
0:23:17 He actually took a picture of it with his camera
0:23:18 and then emailed it to me.
0:23:22 I looked at it, I thought I knew what it probably was.
0:23:25 I then went through a long and arduous process
0:23:27 of communicating with his dermatologist
0:23:29 who he was supposed to see next week
0:23:31 and sending him the image.
0:23:32 Boy, wouldn’t it have been nice
0:23:34 if I could have taken the image
0:23:36 right from the telehealth platform,
0:23:39 sent it over to his dermatologist
0:23:42 and messaged his dermatologist on the same platform.
0:23:43 Hey, can you take a look at this
0:23:46 and let me know what you think?
0:23:48 The biggest problem from my perspective to solve,
0:23:50 and this is not just telehealth,
0:23:53 but I think everything is, you know,
0:23:56 real time or fast communication between clinicians,
0:23:59 whether it’s colleagues or second opinions
0:24:01 or docs you refer to.
0:24:04 That’s the biggest pain point for me,
0:24:07 which ultimately could be solved partially with this.
0:24:10 A couple of the observations that I had
0:24:14 watching some of these telehealth visits take place
0:24:18 were just kind of how open and comfortable
0:24:22 a lot of the patients seemed chatting over a video visit.
0:24:25 My sense is that some of that is because applications
0:24:28 like FaceTime and other video chatting applications
0:24:29 are just so much more prevalent today
0:24:32 that a lot of patients don’t feel like
0:24:35 it’s quite as awkward as you might have anticipated.
0:24:39 It was actually interesting to get a glimpse
0:24:42 into how they’re functioning and what they’re doing
0:24:44 and the fact that they’re running in from the kitchen
0:24:46 or, you know, just you kind of get a sense, actually,
0:24:49 of a patient’s mobility and comfort level
0:24:53 with their ADLs, activities of daily living
0:24:54 in a way that you can’t sometimes get
0:24:56 when they’re sitting on an exam table.
0:24:59 So I think it’s been interesting to see
0:25:03 that we might actually learn about patients
0:25:06 in a way that is hard to do when they come to the clinic.
0:25:09 Yeah, you know, it’s funny to say that.
0:25:12 I mean, I’ve found one of the really useful things
0:25:15 is seeing people in their own environments and in their homes.
0:25:18 One thing I always like to do with patients
0:25:22 is I’ll ask them to bring in pictures of themselves
0:25:24 either when they were younger or before they were sick.
0:25:28 And I was telehealthing with one patient
0:25:30 and she was sort of sitting at a desk and behind her,
0:25:34 she had like a million family photos all hanging on the wall.
0:25:36 And I was like, and I didn’t do it,
0:25:38 but I was so tempted to say, hey, could we, you know,
0:25:42 pull a couple of those off and let’s take a look at them.
0:25:45 But you do, you really get insight that you don’t.
0:25:47 And I’ve had like, I literally had two patients,
0:25:49 30 seconds into the conversation, say to me,
0:25:51 oh my God, I forgot to put on makeup.
0:25:54 And, you know, you just realize you see people different
0:25:57 than you see them when they come into the office.
0:25:59 One space that we’ve all heard a lot of discussion
0:26:03 about actually in the context of coronavirus therapies
0:26:06 are clinical trials that are now actively enrolling.
0:26:09 And I think a lot of people have started thinking
0:26:10 about what a clinical trial is
0:26:13 and have heard the word more than they might have before.
0:26:16 But for cancer patients, this is the norm, right?
0:26:18 A lot of our cancer patients are always thinking
0:26:23 about a trial in the future or they might be on a trial now.
0:26:26 We’d love to hear how clinical trial operations
0:26:29 are affected by shelter at home orders
0:26:32 for so many of our non-essential workforce.
0:26:35 How is that playing out for patients on trials
0:26:38 or patients considering trials?
0:26:40 The shelter in place is obviously
0:26:43 hampered enrollment considerably.
0:26:45 We’ve had a tremendous decrease in enrollment
0:26:46 over the last couple of weeks,
0:26:48 which is very understandable.
0:26:50 Patients just don’t want to come to the hospital
0:26:51 nearly as much.
0:26:53 Adding to that is that a lot of trials
0:26:56 are actually holding new recruitment to trials as well.
0:26:59 So a lot of the crew was being held right now.
0:27:00 For patients currently on trial,
0:27:02 they are allowed to continue on treatment
0:27:04 and they are encouraged to do so.
0:27:06 They are allowing for more deviations,
0:27:09 meaning that patients can skip treatments
0:27:11 if they feel that they need to
0:27:13 in order to protect themselves from the virus.
0:27:16 So we’re starting to see that sponsors or trials
0:27:19 are a lot more lenient than they were in the past.
0:27:22 A lot of the trials that are with oral medications
0:27:24 are still being continued
0:27:27 and these sponsors are also feeling a lot more lenient
0:27:29 about shipping drug home
0:27:31 so that patients don’t have to come into the hospital.
0:27:34 I also agree that this is going to cause us to
0:27:37 reconsider the way that we do a lot of our clinical trials.
0:27:39 Is it really necessary that you get
0:27:43 that certain esoteric lab on day 52 of a clinical trial
0:27:45 and make the patient come in for four hours away to do that?
0:27:47 I think this will make us realize
0:27:49 that a lot of the things that we’re doing
0:27:52 are probably not as important as we used to think they are.
0:27:55 – The FDA announced that they’re working
0:27:59 on providing guidance to sponsors and trial sites
0:28:02 to enable sufficient regulatory flexibility
0:28:05 to allow trials to continue through this period
0:28:06 to the extent possible,
0:28:09 while of course keeping patient safety paramount.
0:28:12 What are some of the tactical ways
0:28:15 in which you think this guidance could play out
0:28:18 and in which you think there might be more flexibility
0:28:20 than there was before?
0:28:23 – I think the flexibility ultimately helps.
0:28:26 There is so much concern about not following protocols
0:28:31 exactly to the T and deviations to that.
0:28:34 But I think that flexibility is gonna get people
0:28:37 ultimately more comfortable with having patients on trial
0:28:38 during this time.
0:28:43 To me, it’s certainly it’s understandable
0:28:44 why things are dropping off.
0:28:47 I think it’s also in many ways tragic.
0:28:50 We at baseline don’t put enough patients on clinical trials.
0:28:53 It’s such an urgent need and it’s just disappointing
0:28:56 even if understandable to see why that’s dropping off.
0:28:58 And you probably saw, I think Pfizer made an announcement
0:28:59 that they’re stopping accrual
0:29:01 except for life-threatening conditions.
0:29:03 I don’t know if life-threatening applied
0:29:05 to every cancer trial that they’re doing or not.
0:29:08 For trials where, assuming you have enough staff
0:29:10 to keep taking care of patients,
0:29:13 if you have a trial that doesn’t require visits
0:29:15 outside of the standard of care,
0:29:18 I’d hope we’d be able to see those continue.
0:29:22 One thing that was really personally very exciting to me
0:29:24 is we have a lot, community oncology
0:29:26 does a lot of clinical trials,
0:29:28 something that I think not everyone appreciates.
0:29:32 And we’ve seen a lot of continued accrual to trials.
0:29:34 We work very closely with one particular trial
0:29:37 that we’ve helped do data collection for.
0:29:41 And we’ve seen practices even in the last week
0:29:44 accrue four, five or six patients to this.
0:29:47 Docs, despite adversity, get that clinical trials
0:29:50 are important and are continuing to try to do it,
0:29:51 even in difficult times.
0:29:53 – It’s interesting because, going into this,
0:29:56 I sort of, as an outsider, sort of naively thought
0:30:00 that the broadest level, the advice
0:30:02 or the kind of thinking would be pause
0:30:05 what we can pause safely.
0:30:07 But it actually sounds like what you’re saying
0:30:11 is keep doing everything that we can keep doing safely.
0:30:13 It’s more the spirit.
0:30:15 That’s my perspective, especially,
0:30:17 I mean, it also depends on the trial, right?
0:30:19 So if you’re doing a clinical trial,
0:30:21 which are often the case in cancer,
0:30:24 whether it’s a phase one trial or a phase two trial
0:30:27 for people with bad cancers who have run out of options,
0:30:29 it’s difficult to continue those.
0:30:31 But I think it’s appropriate a lot of the time
0:30:33 to think about how can we make it work.
0:30:37 – We are doing a international cancer registry
0:30:40 right now on patients with coronavirus.
0:30:42 And this was an effort that was largely led
0:30:44 through Twitter, actually, by recruiting other physicians
0:30:47 from other institutions to capture all this data.
0:30:50 And I do think that clinical trial data,
0:30:52 especially randomized clinical trials,
0:30:55 are gonna be more difficult to do in the current era
0:30:57 because of the regulation.
0:30:58 There should be an importance placed
0:31:00 on what we call real world evidence.
0:31:03 And this type of data is gonna be very informative
0:31:05 in the next several months as well,
0:31:06 as we’re gonna get limited data
0:31:08 from randomized clinical trials.
0:31:10 – Let’s sort of go back to where we started
0:31:13 and think about what happens to oncology as a whole
0:31:18 when it’s forced in this moment to go so virtual.
0:31:19 What do you think is gonna stick?
0:31:23 And what do you think we will let fall by the wayside
0:31:26 when we finally get to move out of this moment?
0:31:29 – I do think that virtual clinics are here to stay.
0:31:32 I also think that we’re gonna see a large shift away
0:31:36 from hospital-based care and more towards home-based care.
0:31:39 Do you really need your infusion at Stanford
0:31:41 or at your cancer center,
0:31:44 as opposed to the confines of your living room?
0:31:47 I think if you look at certain examples,
0:31:49 for drugs that we use in lymphoma,
0:31:51 for instance, a drug called Rituximab,
0:31:54 it’s an IV medication that has now been formulated
0:31:56 to be subcutaneous.
0:31:57 And you can imagine a scenario
0:32:01 where you had a digital safety lock on that syringe
0:32:04 that could be activated by your provider
0:32:07 if after a virtual consultation to make sure
0:32:10 that your lab’s look okay and that you were feeling okay.
0:32:12 So I think a lot of the stuff that we’re doing right now
0:32:14 in the clinics can certainly be done at home,
0:32:19 kind of furthering the capabilities of virtual medicine.
0:32:20 – What’s the incentive to keep doing that
0:32:23 after the coronavirus goes truly away?
0:32:25 – Patients in general actually prefer this.
0:32:28 It’s amazing that after these virtual consultations,
0:32:29 I’ll often tell a patient that,
0:32:31 “Well, I’ll see you back in three months,
0:32:32 hopefully here at Stanford.”
0:32:33 And they say, “Well, Doc, actually,
0:32:35 this worked out pretty well.
0:32:37 Why don’t I just see you back on my cell phone
0:32:38 in three months instead?”
0:32:41 I do think that we’re gonna see a greater use
0:32:44 of digital health and public health interventions as well,
0:32:48 and the rise of public-private partnerships.
0:32:50 If you look at the countries that were the most successful
0:32:53 in containing this epidemic,
0:32:55 they’re all countries that employ technology
0:32:58 as a very large part of their response.
0:33:00 You look at South Korea using cell phone data
0:33:04 to be able to contact tracing for patients
0:33:06 who are infected with coronavirus.
0:33:08 Even Russia was using artificial intelligence
0:33:12 and facial recognition to be able to enforce their quarantine,
0:33:14 not suggesting that we do that by any means.
0:33:16 But then you can even look to China
0:33:18 that was using artificial intelligence
0:33:22 to be able to diagnose COVID just from chest x-rays.
0:33:24 So I think that we’re gonna start seeing
0:33:27 a lot more digital health and public health infrastructure,
0:33:30 and I think that’s another thing that’s here to stay as well.
0:33:32 – Specifically around telemedicine,
0:33:33 the genie’s out of the bottle,
0:33:36 and it’s gonna be hard to sort of put it back in.
0:33:38 One of the things that CMS did
0:33:40 that I think was really, really helpful
0:33:41 was expanding the number of codes
0:33:43 that you could use for telemedicine.
0:33:45 I like to think that’s gonna continue
0:33:48 because that just made the whole process
0:33:51 much, much easier for clinicians to do.
0:33:53 I like to think that all of us having
0:33:56 to go through the experience of really asking,
0:34:00 do we really need these things that we always think we need?
0:34:01 Just like sort of value-based care
0:34:03 has pushed us in a direction
0:34:05 to second-guess things that we used to do.
0:34:09 I hope that this does as well.
0:34:10 And then lastly, one of the things
0:34:14 that this whole episode has caused me to reflect on.
0:34:16 I’m fortunately healthy, my family’s healthy,
0:34:19 and I was basically, we were all given news
0:34:21 a couple of weeks ago that you have to stay at home,
0:34:24 you can’t go out, your life’s gonna be disrupted,
0:34:26 and if you get this disease,
0:34:28 there’s like, for me and my age group,
0:34:31 there’s a couple percent chance that I’m gonna die,
0:34:33 and that was really, really hard news
0:34:35 for me to take as an individual.
0:34:37 And in my job as an oncologist,
0:34:39 I give people news 10 times worse than that
0:34:43 every single day, and they deal with it a lot better
0:34:46 than I have dealt with this.
0:34:48 So I like to think that for all of us
0:34:49 who’ve been lucky enough to be healthy,
0:34:51 it gives us a little bit of perspective
0:34:52 of what our patients go through.
0:34:53 That’s wonderful.
0:34:56 The silver lining, in my view,
0:34:58 from sort of a technology perspective,
0:35:02 is that it’s the pressure test of unprecedented scale
0:35:07 for our system to navigate how to incorporate technology
0:35:08 in different aspects of care,
0:35:10 how to keep in touch with patients
0:35:12 when they can’t come into the clinic,
0:35:14 how to make complex decisions
0:35:16 that require coordination in real time
0:35:19 with uncertain data between different specialists,
0:35:21 because this is a time when,
0:35:23 under this type of real pressure,
0:35:24 that’s when these future lists are generated,
0:35:26 and that’s when we realize what we need,
0:35:28 and what we realize we need now.
0:35:30 And so I think that this pressure test
0:35:35 is just going to actually be an incredible learning opportunity
0:35:37 for a variety of sectors of digital health.
0:35:40 The other piece that I think is interesting,
0:35:43 kind of zooming out of oncology in particular,
0:35:45 is just the increased awareness
0:35:46 that’s occurred for diagnostics
0:35:48 and infectious disease therapeutics.
0:35:53 I think it’s given our public, our funding agencies,
0:35:57 companies, investors, everyone across the board,
0:35:58 a deepened level of appreciation
0:36:02 for how important anti-infective agents
0:36:04 and preventive agents really are.
0:36:09 And so I hope we’ll see renewed investment globally
0:36:10 in technology in that domain.
0:36:11 – That’s really inspiring.
0:36:13 Well, thank you so much for all joining us
0:36:15 on the A16Z podcast,
0:36:16 and thank you for everything you’re doing
0:36:18 for your patients every day.
0:36:19 – Thank you so much, guys.
0:36:23 (upbeat music)
0:36:33 [BLANK_AUDIO]

0:00:05 Hi and welcome to the A16Z podcast. I’m Hannah and in this episode we talk all about the remarkable
0:00:12 story of one American family, the Galvans, Mimi, Don, their 10 sons, and two girls,
0:00:18 out of whom six sons were afflicted with schizophrenia, following them from the 1950s to today.
0:00:22 Robert Kolker, author of the book and previous author of “Lost Girls,” writes,
0:00:29 “They lived through the eras of institutionalization and shock therapy, the debates between psychotherapy
0:00:34 versus medication, the needle in a haystack, search for genetic markers for the disease,
0:00:38 and the profound disagreements about the cause and origin of the illness itself.”
0:00:44 And because of that, this story is really more than just a portrait of one family. It’s a portrait
0:00:49 of how we have struggled to understand this mental illness, the biology of it, the drivers,
0:00:54 the behaviors and pathology, the genomics of it, and of course the search for treatments that might
0:01:00 help. Also joining Robert Kolker and myself for this conversation is Stefan McDonough, Executive
0:01:05 Director of Genetics at Pfizer World R&D, who is one of the genetic researchers who worked closely
0:01:11 with the Galvans. We start by talking about our attempts to understand and treat schizophrenia
0:01:17 from Freud to lobotomies to the entrance of Thorzine onto the scene, where that understanding
0:01:22 of the disease finally began to shift, especially with new technologies and the advent of the human
0:01:28 genome project, and where we are today in our understanding of the disease, how to treat it,
0:01:33 and where our next big break might come from. What really struck me about this book was that
0:01:41 it was this huge story, not just about one family and this particular disease of schizophrenia,
0:01:46 but also kind of a portrait of our entire effort to understand mental illness period,
0:01:53 and not just how we understand it, but how we experience it and how we try to treat it.
0:01:59 Let’s go back a little bit and talk about schizophrenia itself. I’d love to hear where
0:02:05 you think our modern understanding of the disease really began. You describe a key moment in 1903
0:02:11 where we shift from thinking of it as a religious ailment into something else, or where would
0:02:16 you begin that story? That’s around the time of the dawn of psychiatry as we understand it today.
0:02:21 Obviously, there are glimmers beforehand of people believing that mental illness is physical and
0:02:27 not spiritual or religious, but by the turn of the century, there was an entire field emerging,
0:02:31 and there was a nature-nurture debate over what schizophrenia was that really,
0:02:37 in many ways, continues today in a different form. Back then, the debate was between Freud,
0:02:43 who believed that therapy could cure schizophrenia, that schizophrenia was something that happened in
0:02:48 the nurture side of things, something that happened in your childhood caused it, perhaps bad parenting.
0:02:54 On the other side were a lot of other psychiatrists, including the ones who named schizophrenia,
0:02:59 who believed it had some sort of physical property, but could never put their finger
0:03:05 on what it was. Into this debate come the Galvans, who by the 1950s and ’60s are starting to become
0:03:10 mentally ill at a time where most psychotherapy believes it’s the parent’s fault, and medical
0:03:16 psychiatry is sure that drugs can hit whatever is happening genetically, but they really have no
0:03:21 clue how those drugs work or what genes are really at play, and this continues for decades.
0:03:27 The story of the family starts in the 1950s, but you describe some developments before you
0:03:34 get to this with Dr. Frome Reichman and Gregory Bateson and a couple of other characters that,
0:03:39 to me, felt like kind of key points as our understanding of the disease was developing.
0:03:45 In 1948, Frita Frome Reichman, who was a psychoanalyst then living in America, came up with a term
0:03:51 called the Schizophrenogenic Mother, which she believed was a certain type of mother or father,
0:03:58 in some cases, who was so bad at parenting, was so torturous in the way that they dealt with their
0:04:04 children that the child then would somehow create their own imaginary reality in order to escape from
0:04:11 that parent and become schizophrenic. That was the split. The split was from internal to external.
0:04:15 Exactly. Sometimes people think schizophrenia means split personality,
0:04:20 but it really never did, and it’s really a split between your perceptions of reality and of what’s
0:04:26 happening inside you. Frita Frome Reichman was doing this at a time where lots of psychoanalysts
0:04:32 were blaming mom and dad for lots of things. Of course, by 1960, you have the movie Psycho,
0:04:36 the greatest homicidal maniac in all of cinema. His problem is his mother, and everybody says,
0:04:42 “Oh, yes, that must be what happened.” It seems strange now, but when you think about it back then,
0:04:48 people like Frita Frome Reichman and Gregory Bates, they were doing battle against eugenics
0:04:52 at the time of modern feeling that you could breed out schizophrenia and that you should
0:04:59 sterilize or even euthanize mentally ill people. They were also doing battle with people who were
0:05:04 committing lobotomies and insulin shock therapy and electroshock therapy. They were doing battle
0:05:09 with a medical field that was treating schizophrenic people as subhuman. They felt like they were on
0:05:15 the side of the angels. It was really interesting to me that the kind of duality from Reichman,
0:05:21 her position, on the one hand, having more compassion than anyone had ever really had before
0:05:25 for the people suffering from this disease, but on the other hand, having so little compassion.
0:05:29 For the mothers, it was just a very interesting split there.
0:05:35 Yes, indeed. I think she and a lot of other therapists of her generation were threatened
0:05:40 by changes in society. Women are working after the war. The family unit is being threatened
0:05:46 in some way. The sexual revolution is about to happen. Any major changes in society then began
0:05:51 to be attached to the idea of mental illness until the researchers who went after the Galvin
0:05:57 family who came up in the late ’60s and early ’70s at a time when a lot of established psychiatry
0:06:02 was telling them still that parents were the problem and that working mothers in particular
0:06:07 were the problem. One of the things that really struck me, Bob, was you start in a really interesting
0:06:17 place. You begin with a story of training a falcon and Mimi, you call her a refined daughter of
0:06:23 Texas aristocracy by way of New York, clutching a live bird in one hand and a needle and thread
0:06:30 in the other, preparing to sew the bird’s eyelids shut. Can you tell me why did you start this huge
0:06:36 story about mental illness with this one incredibly vivid and surprising moment?
0:06:43 This took place maybe a week or two after Mimi and her children moved to Colorado Springs in
0:06:48 the early ’50s to join their husband who had just moved there for the Air Force. The whole thing was
0:06:53 unfamiliar to her. It was out of her comfort zone and then to be thrust into this situation where
0:07:00 she was suddenly getting into falconry and having to sew eyelids shut, this was as foreign as it
0:07:05 came to her. The point of the story really is that she accomplishes it. She winds up training and
0:07:12 disciplining the falcon and it winds up becoming almost an allegory for how she approaches the rest
0:07:16 of her life, including the raising of children. She thinks if she tries hard enough, does all the
0:07:21 right things and does them all in the right way with discipline and pressure, she will get the
0:07:26 results she desires. And we all know that with children that isn’t exactly true, but in the
0:07:34 Galvin family’s case, it’s tragically true. 12 children, six of whom had acute mental illness.
0:07:41 And before we get into this particular family and how they dealt with this, I just want to talk
0:07:46 also a little bit about up until that moment, the different treatments that had matched up to our
0:07:51 understanding of the disease that we had tried. You mentioned insulin shock in the 1930s, lobotomizing
0:07:56 attempts and things like that. Can you kind of map out those early therapeutic attempts?
0:08:01 Well, lobotomy is about severing nerves in your frontal lobes. It’s an extreme measure and I
0:08:07 think most people would consider it barbaric now, but it was intended to impair you just enough
0:08:11 so that you would stop hurting yourself mentally. That seemed to be the justification for it at the
0:08:16 time. But the other procedures that you mentioned, things like electroshock therapy and insulin
0:08:22 shock therapy all sort of operate on the same principle, which is to somehow induce enough
0:08:29 stimulation, enough of a seizure, even almost a medical coma, so that you shock the patient into
0:08:35 focusing and not being so distracted or drawn away by whatever is going on with their brain chemistry.
0:08:41 And sometimes it would seem to work at least temporarily, and other times they would decide
0:08:46 that the person needed to be shocked every day. So then in the 1950s, a major development on the
0:08:51 drug side, you talk about the entrance of thorazine onto the scene, which dominates the next
0:08:55 second half of the century and still has a huge legacy in how we handle this
0:09:03 mental illness. Can you talk about what brought thorazine in and what that moment was like?
0:09:08 Like a lot of pharmaceutical advances, it happened sort of sideways or by accident.
0:09:12 There was a French surgeon who was trying to come up with a battlefield anesthetic,
0:09:20 and he did some combining of traditional anesthetic and narcotics and found that the people he was
0:09:26 testing it on, it almost induced a happy coma on them the way that he described it. This drug
0:09:32 eventually was thorazine and even now, really, thorazine is the great advancement pharmacologically
0:09:38 for schizophrenia and still is. And then there’s an atypical version or variety of a psychotropic
0:09:44 drug called chlozapine. And my understanding is that those two drugs really are the coke and pepsi
0:09:50 of this field and that any drug out there is sort of a derivation of one or the two.
0:09:56 There are people who were in such extreme condition and harming themselves so much
0:10:00 that certainly drugs like this could be helpful to them to keep them alive.
0:10:05 But I think the other sad fact is that they aren’t cures and that it’s been decades now
0:10:10 and there really has been no revolutionary drug for schizophrenia since the 1950s and 60s.
0:10:14 The best clue to actually understanding how to attack a disease is have something that cures
0:10:19 the disease, especially something like schizophrenia where the etiology, we may be starting to
0:10:24 understand some of the underpinnings with it, but people don’t come to us at birth and say,
0:10:29 “I’m going to have schizophrenia. Please modify me in some way.” We wait until the symptoms develop
0:10:35 and so exactly as Bob said, it was the cornerstone advance in the field to find a therapy that you
0:10:41 can take as a pill that did in some ways make some patients better. Then you can just simply
0:10:46 reverse engineer that therapy and try to find out on a molecular level what is it doing.
0:10:50 And then that leads to understanding of disease, to dopamine receptors, to serotonin receptors,
0:10:56 and so on. So let’s go back now to that moment in the 1950s which is basically where the Galvan
0:11:03 story starts as well. By the 1960s, the oldest of the 12 Galvan children were starting to go
0:11:08 off to college and as we know schizophrenia’s onset is quite often in late adolescence. And so
0:11:13 the oldest son, Donald, the star of the family, the football star, the guy who dated the general’s
0:11:20 daughter and who was a master falconer and repeller on the cliffs of central Colorado,
0:11:25 he had really had secretly felt quite alienated from mainstream life and really was struggling
0:11:30 in many ways privately and that struggle went public by the middle of college. He ran into a
0:11:36 bonfire and didn’t know why, he tortured a cat and killed it and didn’t know why. He ended up in
0:11:42 student health services for many different reasons until finally psychiatrists got involved. And this
0:11:49 was a panic moment for Don and Mimi Galvan, the parents, because they knew that first of all they
0:11:55 would be judged because at the time if your child had a psychiatric problem it obviously must have
0:12:01 been the parent’s fault. And so they went shopping for a good opinion because back then really what
0:12:06 illness you had psychiatrically really depended on what doctor you visited. Some would say give him
0:12:12 Thorazine, others would say give him a lobotomy. So they went and they found a doctor who recommended
0:12:17 he could go back to college and that he would just grow out of it. And then he got worse and worse
0:12:24 and until finally he had a moment of violence with his young wife, that was it for him. He went off to
0:12:28 the state mental institution for a few weeks and then spent the rest of his life
0:12:34 really at home with Don and Mimi, with his parents, almost as a revolving door between the state
0:12:40 institutions and home. It struck me that every story kind of showed the different lenses that
0:12:47 we’ve used to try and understand this disease. Each had some of these common traits and some
0:12:54 wild discrepancies and differences. A lot of the stories began with we sort of always knew,
0:12:59 but some of them became completely out of left field. And then some of them even seemed like
0:13:05 they come from a cultural moment like Michael who goes to live on the farm in the 1960s.
0:13:09 That’s right. Michael’s sort of a hippie and he is rebelling and then that gets confused with
0:13:16 mental illness for a time. He insists that he is not mentally ill. Peter was very, very
0:13:21 oppositional as a kid said no all the time and then he had psychotic breaks. So you could say
0:13:26 they saw that coming. Joseph had a detachment from reality, it seemed now and then. And so
0:13:31 everybody was sort of waiting for him to finally have a psychotic break and he did in the early
0:13:36 80s. But then there were surprises. Matt, who was a talented ceramic artist suddenly one day out of
0:13:43 nowhere, smashes something that he made and strips naked in a friend’s house and suddenly
0:13:50 he becomes medicalized as well. So it’s interesting how some seem explainable and others do not.
0:13:55 At the same time as we’re hearing this story of how our understanding of the disease changed and how
0:14:02 this one family that manifested so many of those efforts to understand and manage it,
0:14:08 it’s also a story of technology, the developing technology that we have to understand biology
0:14:13 and to understand our brain. Another key moment was when suddenly we start being able to see the
0:14:21 brain through scans in the 70s. Can we go to how this story evolved when that kind of technology
0:14:29 came on the scene? Yes, by the 70s researchers were able to have some glimpses into the brain
0:14:35 thanks to technology, thanks to MRIs and PET scans and CT scans and the like. And with the
0:14:41 sequencing of DNA, it became possible to think about being able to actually study the genetics of
0:14:46 any sort of population of people with any sort of illness or disease. Technology drives biology
0:14:50 across multiple areas. I see that story again and again in whatever field I’m in. You look at the
0:14:55 beginning, the first MRIs, those could only be structure. This might tell you is the structure
0:14:59 of the brain different, but it doesn’t tell you how it functions. It feels like it’s a story about
0:15:04 a family, but it’s also really a story about kind of modern genomics and going from understanding
0:15:11 something as a inherited disease in some way to dialing into a way different level of understanding
0:15:17 about genetic information. So what was the genomic story of what we understood of schizophrenia
0:15:23 pre-Galvin’s? Why was the family such a turning point in the context of the human genome project?
0:15:28 And then what did we learn from them? I write about Robert Friedman, who’s at the University
0:15:34 of Colorado Hospital, and Lynn Delisi, who is at the National Institute of Mental Health.
0:15:38 Lynn’s story really intersects with Stefan’s, my fellow guest on this show.
0:15:44 And before she met Stefan, she was a pioneer in studying families like the Galvans. And the
0:15:49 Galvans were the biggest family she ever found in those early years. And she was convinced that
0:15:54 families were the best way to take a look at this illness because you weren’t searching for a needle
0:15:59 in a haystack. You had a much smaller haystack to look through. They were a closed petri dish of
0:16:04 shared genetic data with a lot of incidents of schizophrenia. Multiplex families like the Galvans,
0:16:09 with lots of schizophrenia in them, have something to teach us. And she amassed the largest collection
0:16:14 of family DNA for this purpose. But there came a time in the ’90s when the human genome project
0:16:19 was underway, when everyone thought that once the genome was sequenced, they’d be able to do
0:16:25 entire population-wide studies. And that anyone with schizophrenia or with any other complex
0:16:31 genetic disorder would sort of stick out like a sore thumb, you would find the smoking-gun gene,
0:16:38 or genes. And then you’d have a target to medicate with a drug. And bingo, we’d all be cured by the
0:16:45 time dinner came. But the problem is that with complex conditions like schizophrenia, it only
0:16:51 complicated things so much more. They found one genetic irregularity for schizophrenia, and then
0:16:57 another, and then another, and then another. Until now, there are far over 100 genetic issues.
0:17:04 Unfortunately, each one of these irregularities contributes just a fraction of a percentage of
0:17:09 the probability that you might get the illness. And so it winds up being meaningless. It strikes
0:17:15 me as just as fluid and complicated and long a list as the list of symptoms that over the last
0:17:21 century have been associated with the disease, the manifestations of it. Exactly. And it’s not
0:17:27 helpful clinically. It might be helpful for future research, but at the moment, it just makes the
0:17:33 mystery more mysterious, which is what makes it so interesting that when Stefan recognized that
0:17:38 families might have something to offer and wondered, hmm, who out there has been studying
0:17:43 families? And lo and behold, there was a woman who had been doing it all this time, and the two of
0:17:47 them teamed up. Stefan, can you talk about what it was like to enter on the scene in that moment and
0:17:54 what the genomic aha was for you there? Yeah, the technology had not been there to really analyze
0:17:59 the families that Lynn collected. She collected them so much before the technology was available to
0:18:04 really look in fine depth at the genome and find, is there something different? Back when I was in
0:18:10 grad school studying genetics, it was mustard weed and fruit flies and sort of model organisms.
0:18:15 When the genomic revolution came along, so much computational power came to be developed. The
0:18:21 technology just kept developing to be able to sequence entire genomes rapidly and inexpensively,
0:18:26 comparatively inexpensively. People sort of thought with a disorder that is so strongly
0:18:33 heritable as schizophrenia, there must be something there. And again, what was turned up in surveys
0:18:38 of tens of thousands of schizophrenia, looking at all the genetic variants they carried versus tens
0:18:44 of thousands of people matched controls as best they could, say for ethnicity and other factors.
0:18:49 No question there are differences. Those have led to some hypotheses, like sort of a general
0:18:55 overall role of the immune system. But in terms of discovering what is the driver
0:19:01 for a disease like schizophrenia, it just simply didn’t work that way. And we still don’t understand
0:19:07 why. Where are we starting to make progress there and understanding kind of the biology
0:19:13 and the drivers and potentially how to treat them? You talked about looking at when a drug
0:19:18 like Thorazine works and trying to work backwards from that to understand more of the biology.
0:19:23 Are we still there or does our understanding of the underlying genomics shifted a bit?
0:19:30 Shifted a bit. I think we’ve gone from a picture of, again, as Bob said, evil spirits or dreams
0:19:36 or some environmental influence or a mother to sort of a holistic picture of the brain whereby
0:19:42 following Thorazine, we would say, okay, well, dopamine is dysregulated or glutamate is dysregulated
0:19:46 and these are chemicals neurotransmitters for how nerve cells communicate with one another.
0:19:53 Now we’re verging towards a sort of cellular synaptic view. Another of the technology that
0:19:59 developed probably in the 80s and especially 1990s was the ability to really look in very,
0:20:05 very fine detail in sort of millisecond scale resolution at how nerves communicate with one
0:20:10 another. You can stick an electrode on one nerve and stick an electrode on the other and really
0:20:15 record how they’re communicating. So this is, I think, how our overall picture of biology is
0:20:22 evolving into as to what it means how we cure the disease. Classic analogy, in order to fix a
0:20:27 broken TV set, you have to know what makes it work in the first place. And we’re still not there,
0:20:32 but we’re getting closer. What is the technology that’s coming onto the scene now that is changing
0:20:38 our potential understanding moving forward? I think one of the areas that’s exciting now
0:20:43 is you can actually take a skin cell from somebody and treat it with appropriate biological factors
0:20:49 and it will differentiate into something that to first approximation might be a human neuron.
0:20:54 I haven’t seen therapies come out of it yet. In fact, it may be another blind alley as with
0:21:00 all areas of research. But there is the hope that if you take a skin cell or a group of skin cells
0:21:06 from somebody with schizophrenia, perhaps that mutation is genetic. Perhaps that mutation then
0:21:12 is still carried in the skin cells and their nerves might look different. So this is a possible angle.
0:21:16 It’s a bit of a risky one for many reasons. You never know if you’re actually dealing with a neuron.
0:21:20 What do you mean by it’s hard to even know if you have a neuron?
0:21:26 Well, you’ve taken a fibroblast, a skin cell, you’ve treated it with appropriate factors and
0:21:31 certainly it elongates, it starts sending out processes and if you can stick an electrode
0:21:36 in it, you can see that it’s electrically active. Does that mean that it really is close enough to
0:21:41 a human neuron in a human brain that has developed through its entire life, through the entire life
0:21:45 of the individual and has been exposed to all the different environmental influences?
0:21:46 Right.
0:21:51 The question isn’t, is one thing like the other? They are like on some level, unlike on others.
0:21:56 Is it enough alike that you could actually try to turn a therapeutic on it and try to do your
0:22:01 modification now you’ve got sort of a disease in a dish? And that’s an open question.
0:22:06 Bob, where would you see the Galvanes their story if it was unfolding today? Can you talk a little
0:22:12 bit about how it kind of mapped to where the understanding is, where their story ended?
0:22:16 Well, the two separate teams who studied the Galvanes, each have come forward with some really
0:22:23 interesting advances, both of which offer a lot of hope. Lynn, Delisi, and Stefan sequenced the
0:22:29 genome of the Galvan family and found one irregularity in a gene called shank 2. This is not a
0:22:34 silver bullet or a smoking gun. It’s not like the shank 2 gene is the ketoschizophrenia. However,
0:22:41 assuming that it is the player that really did its trick on the Galvan family, it is a gene that’s
0:22:45 highly related to brain function and could, with further study, point the way to understanding
0:22:51 how schizophrenia works, how that TV set works, as Stefan had said before. And so that’s exciting.
0:22:57 And more broadly, in terms of drug discovery, families like the Galvanes can be almost sort
0:23:03 of test kitchens. You can look at how their genetic code might interact with certain potential
0:23:08 therapies and see perhaps how it might go with the broader population. Then with that second
0:23:14 set of researchers led by Robert Friedman over in the University of Colorado, he, with help from
0:23:19 the Galvanes and other families, identified another genetic area called churn a seven. And churn a
0:23:25 seven is related to the vulnerability theory of schizophrenia, which is that perhaps one is
0:23:31 oversensitive or has a sensitivity issue to stimuli. It looks at schizophrenia as a developmental
0:23:37 disease, one that really begins in utero even though it manifests itself much later. And over the
0:23:43 years, he struggled to find a way to perhaps make the churn a seven area more healthy or more resilient
0:23:51 and less vulnerable. And he has a hypothesis that there actually is a safe nutritional supplement,
0:23:58 choline can strengthen brain health generally of the unborn child, but also perhaps cross your
0:24:05 finger 16 times, perhaps many years from now prove to make the children more resilient,
0:24:11 less vulnerable to psychosis. And they’re doing longitudinal studies right now using choline,
0:24:15 and if it shows any promise at all, he has the Galvanes and other families like them to thank.
0:24:20 Stephanie, it would be very interesting to hear from your side of the kind of story of the pharma
0:24:26 industry attempt to manage this as well. Where would you see those attempts after Thorzine?
0:24:30 Then where did we go next? And what was the sort of industry response? Where are we today
0:24:34 in the possibility? Yeah, there was a quite productive age where drugs like Cypraxa were
0:24:39 developed, where we’re looking for simply animal behaviors that were related to schizophrenia.
0:24:44 And here is where having sort of a toolkit is quite valuable in a sense, because if you know,
0:24:49 for example, if there’s some odd behavior that an animal is showing that Thorzine mitigates,
0:24:53 then without even knowing the receptors involved, perhaps you can test drugs and animals
0:24:58 for other drugs that mitigate those behaviors and perhaps don’t have side effects.
0:25:03 So the problem, of course, is that rats don’t get schizophrenia. They don’t even have sort of the
0:25:07 massive cortical structures in the folding that we think is where the
0:25:10 higher processes that are affected in schizophrenia reside.
0:25:14 So to your point about cells in a dish, I mean, it’s really a problem of models, right?
0:25:19 It’s a problem of models. How do you, before doing a clinical trial in humans,
0:25:25 how do you get confidence that your drug is going to work? And I think in the 1990s, there were
0:25:29 a number of very good efforts based on sort of synaptic studies. People have known, again,
0:25:34 going back to some of the early pharmacology that dopamine was involved, that glutamate was involved.
0:25:39 Now we started to identify with the human genome project and just molecular cloning in general.
0:25:45 We started to uncover what the molecules are that regulate glutamate and regulate dopamine.
0:25:50 And a number of clinical trials were done on these as well.
0:25:54 What still does stymie the field today is, if you take the overall disease,
0:25:59 what is your model? What do you test it on that gives you confidence that you can test
0:26:01 this safely in humans and that it will have some effect?
0:26:07 How do people even do that? I mean, are there any other tools before you begin clinical trials in
0:26:11 humans when there’s a disease that really doesn’t present anywhere else outside of humans?
0:26:16 Schizophrenia is a tough one. It’s very tough. Now nothing is easy, but for example,
0:26:21 tumors do grow in animals. And you can implant a patient-derived x-plant,
0:26:25 a patient-derived tumor into an animal, and perhaps test therapies there.
0:26:31 Or you do have cancer cell lines, tumors will actually, cell lines will actually grow in a dish.
0:26:35 And so something that kills those could reasonably be called to be acting on the tumor.
0:26:38 And we simply don’t have that equivalent for schizophrenia.
0:26:43 So what was the next moment where there was sort of a something that seemed
0:26:47 on the pharmacology side like a real viable treatment that we were,
0:26:50 you know, that people were getting excited about? And where are we now?
0:26:54 People were excited about metapotropic glutamate receptors. That’s a particular type,
0:26:58 a subtype of glutamate, which is the main excitatory neurotransmitter in the nervous
0:27:03 system in humans. People were excited about sort of finer manipulations of dopamine receptors.
0:27:07 And again, by reverse engineering some of the atypical antipsychotics,
0:27:11 you could find out that serotonin receptors also had involvement.
0:27:15 Now each of these is going to be a broad family of many, many genes.
0:27:22 So can you do more finer manipulations of these? Not every advance in drug therapy has to be a
0:27:29 totally new mechanism. Schizophrenics and other CNS disorders are famous for going off their medications.
0:27:34 So if you can perhaps make a medication that just simply lasts longer
0:27:39 and can be given maybe every month or even at less duration under a doctor’s supervision,
0:27:43 that’s a significant medical advance. And this is an engineering challenge.
0:27:49 I started life as an engineer and drug discovery is really biological engineering.
0:27:53 I’m not saying it’s easy, but we do know how to make drugs last longer in the body.
0:27:56 There’s a very interesting story in there with Nicotine.
0:28:02 The receptor that Robert Friedman in Colorado had identified with help from the Galvin family
0:28:07 and other families like them was a nicotinic receptor. And strictly speaking,
0:28:13 that’s a receptor that when medicated might actually help with focus and concentration.
0:28:18 I mean, there’s a stereotype of schizophrenic patients actually getting some relief from chain
0:28:23 smoking because it focuses their mind. And there’s a hypothesis related to nicotine,
0:28:28 and there was for a time that if you could somehow drug this receptor a little bit to help it along,
0:28:34 that perhaps this would prevent delusions or even prevent psychotic breaks.
0:28:39 And Robert Friedman did try for a while to work on a drug for that, and he reports anecdotal
0:28:45 excellent results from many patients. But it was a drug that you had to take several times a day,
0:28:50 and that was something that the pharmaceutical companies couldn’t bring through trials to make
0:28:56 into a once a day drug. So it went away. So he decided to go after the nicotinic receptor
0:29:04 in utero through choline. Especially in the 1990s, there was a lot of companies and a lot of academic
0:29:10 researchers investigating nicotine and nicotinic receptors. And again, there did seem to be a
0:29:15 clear link to schizophrenia. Perhaps schizophrenics are self-medicating by smoking. If so, perhaps
0:29:21 you can make sort of a subtype of nicotine that gives you some benefit or perhaps even some more
0:29:26 benefit. And again, as Bob said, that perhaps lasts long enough in the body to be practical to
0:29:32 be taken as a drug. So in this case, there was a biological challenge there, no question. But it
0:29:39 became also an engineering challenge, as all drug discovery does. Nicotine is quite a non-selective
0:29:43 molecule. Well, everything it hits is called a nicotinic receptor. But your body has something
0:29:49 like 14, 15 genes for individual subunits that together come together to form a receptor for
0:29:54 nicotine. And they all mix and match in very unpredictable ways and ways that still are not
0:30:01 well known. So the challenge was quite formidable. People did go ahead for technical reasons.
0:30:06 It turned out to be easy to make sort of a subform of nicotine that would only hit Alpha 7
0:30:12 receptors. Not easy, but not impossible either. People had good reason to think that this might
0:30:18 work. No question, it was a huge downer for patients, for the field, for everybody when
0:30:23 this entire class of drugs just sort of didn’t seem to come to nothing. But we learned.
0:30:30 And the negative result often is just as informative as the positive result we do learn.
0:30:37 Can I ask how incentivized is the sort of pharmaceutical industry right now to find
0:30:43 other alternatives to things like the class of drugs that, you know, Thorazine and some others
0:30:48 that you’ve mentioned? I mean, because those do work to some extent, yes?
0:30:53 To some extent. So I’m not a clinician. About 50% of the patients respond well to
0:30:59 atypical antipsychotics. But this doesn’t touch sort of the cognitive and the emotional problems.
0:31:04 And one of the things, one of the many things I’m grateful to Lynn for was really taking me
0:31:10 to visit her patients so that I could really see there’s no question something is wrong,
0:31:16 just sort of a very emotionless, flat affect. The cognition is fine. Clearly,
0:31:21 these people are very articulate. They’re very bright in many cases. But something’s
0:31:27 badly wrong. So to your question, what is the incentive for pharmaceutical companies?
0:31:32 It’s a huge incentive. I think lots of people would love to do it because schizophrenia is
0:31:39 1% of the population. This is across populations, across cultures. So it’s a huge opportunity
0:31:43 to make therapies that help patients. For what is not a rare disease.
0:31:49 For what is not a rare disease. And if you go beyond that, again, as Bob’s book so amply
0:31:56 demonstrates to the toll on people’s lives, it’s far beyond that 1%. We just don’t know how to do it.
0:32:02 Not for the broad schizophrenia there. And this is where I came from my angle
0:32:07 to sort of look at perhaps there might be subtypes of schizophrenia defined by genetics,
0:32:11 where you really would have one particular form of schizophrenia.
0:32:18 So Stefan, if you, as a researcher, if you could wave a magic wand right now,
0:32:24 you know, you mentioned better models. What are the things that if you could wish something here
0:32:29 tomorrow in the form of a new technology or a new capability, what would that be that would
0:32:34 really push us forward into a new chapter? I’ll go way afield. But if we could monitor
0:32:40 the brains of a schizophrenic with sufficient resolution with high resolution, right now we
0:32:46 get about a millimeter voxel with the best bold fMRI experiments. While they’re actually having
0:32:51 a psychotic break, the resolution is still, of course, could be made finer and finer. We still
0:32:56 can’t get down to the level of a single cell. But now with the blood oxygen level dependent,
0:33:02 magnetic resonance imaging, we can get a measure of function in somebody’s brain in real time.
0:33:05 Difficult to do, takes a lot of equipment, takes a particular stimulus,
0:33:08 but one could perhaps hope that this will lead to more insight.
0:33:13 Oh my gosh, how fascinating that we’ve never seen. We actually have no idea what’s really happening.
0:33:18 I mean, consider the logistics. You can’t consent somebody and get them to sit in a machine and
0:33:24 then wait for them to have a psychotic break. Yeah. What would you be looking for?
0:33:31 We need mechanism. If the field as a whole could say, here is a particular area where the excitability
0:33:37 is abnormal, an area of brain tissue that is abnormally excitable, or a particular receptor
0:33:43 that is abnormally excitable, that gives us a good place to start. That gives us mechanism.
0:33:48 And then perhaps we could study what do existing drugs do to that. What is missing with existing
0:33:55 drugs? That’s fascinating. It almost sounds like you need like a wearable MRI, a very high resolution.
0:34:04 Silicon Valley, go to it. What about things like CRISPR? If you do start defining some very specific
0:34:09 narrow, very entirely genetic cause, is that a possibility as well?
0:34:17 So we’ll give a possibility. Say if we knew that a baby galvan or their modern-day counterpart babies
0:34:23 had a variant in a gene that we thought because of families like the galvanes that we had good
0:34:28 reason to believe would make them develop schizophrenia. Can we get in there and change
0:34:36 that one nucleotide to the wild type? It’s conceivable, but again the challenge here is
0:34:43 the engineering challenge. We can do it in a dish, but trying to get just that one gene edited,
0:34:50 trying to get it in just that one nucleotide changed in every cell in the brain and no changes in
0:34:56 any other nucleotides in the brain and delivering something that will actually cross the blood
0:35:02 brain barrier and then doing it on an infant. How would you even test this? Very, very difficult.
0:35:08 Bob, you describe when Lynn Delisi first met the galvan family and you write this incredibly
0:35:12 profound line that really stuck out for me. As she walked through the door of the house at
0:35:17 Hidden Valley Road, she couldn’t help but recognize a perfect sample. This could be the most mentally
0:35:23 ill family in America and you really dove into every element of what that meant for them into
0:35:31 this family’s innermost suffering and struggles. It was just so intimate on some level and also
0:35:36 such a big story on another level of not just schizophrenia, but the way we struggle with
0:35:43 all mental illness, including trauma and depression. What was the kind of big takeaway for you for
0:35:50 having lived inside this family’s mental illness for several generations, really?
0:35:56 Well, I really think the day, hopefully not too long from now, that all this research yields real
0:36:02 rewards will be the day that this family’s sacrifice will finally find its true meaning.
0:36:08 But also, this is a story about experiencing unbelievable and mysterious tragedies one after
0:36:13 the other and coming out the other side. There are members of this family who have found a way
0:36:16 through this and found meaning in life when everything seemed to be going against them.
0:36:22 It’s really about the value of family, in my opinion, this story and the hope for the future.
0:36:26 Stefan, on the research side, what’s coming that we should be aware of that might be
0:36:29 bringing hope to the next generation of Galvans?
0:36:34 It’s tough. I could answer this for Parkinson’s. I could answer for Alzheimer’s. I could answer for
0:36:39 any number of diseases. Schizophrenia is really, really tough. We need some help. We need something
0:36:44 to break in the academic world and that break will come from studying people. There are ever better
0:36:50 ways of studying what is really going on with human biology in people who are kind enough and
0:36:55 selfless enough to volunteer themselves and their families for research. Not just their genome,
0:36:59 we can measure the protein circulating in their blood. There are ways to do this massively in
0:37:04 parallel. Stem cells, organoids, I think it’s too early to see if these are going to be helpful,
0:37:09 but no question, these are a way to explore. People can do longitudinal studies for how people
0:37:14 are changing over time. Imaging lets us look into the brain better than ever before and these
0:37:20 technologies just keep accelerating and improving. We need somebody to set the goalposts. There’s
0:37:26 a lot of advanced technology. Technology will drive biology. We are focused on human subjects.
0:37:32 Something’s got to break. That’s amazing. Thank you so much for both of you joining us on the
0:37:44 A16Z podcast and here’s hoping we get to that break soon.

0:00:05 The content here is for informational purposes only, should not be taken as legal business
0:00:10 tax or investment advice, or be used to evaluate any investment or security and is not directed
0:00:15 at any investors or potential investors in any A16Z fund. For more details, please see
0:00:21 a16z.com/disclosures. Hi, and welcome to the A16Z podcast. I’m
0:00:27 Doss, and this episode is all about what the numbers, both financials and KPIs, do and
0:00:32 don’t tell you about your business. Our guests for this episode are A16Z General Partner and
0:00:37 Managing Partner Jeff Jordan, who previously ran several businesses and took a company
0:00:42 public right after the 2008 financial crisis. David George, who runs our late-stage venture
0:00:47 operation, and Caroline Moon, who leads our financial operations practice and helps companies
0:00:54 with their own best practices. She’s also a former CFO. In our conversation, we cover
0:00:58 the most common mistakes people make when it comes to understanding numbers. When investors
0:01:03 think when they look at a company’s profit and loss statement and why, how investors
0:01:07 use metrics to determine if a business is healthy, and how some founders may use them
0:01:13 to navigate times of crisis. We begin, though, with the basics of the three core financial
0:01:18 statements, the income or PNL statement, the balance sheet, and the cash flow statement.
0:01:24 The first voice you’ll hear after Caroline’s and mine is Jeff’s, followed by David’s.
0:01:28 Especially in the early days of a startup, they’re just going to do cash accounting,
0:01:30 and that’s just literally how much cash you had in the beginning of the period and how
0:01:35 much you had cash at the end of the month. That’s not the same thing as what a PNL really
0:01:41 should show, because your PNL paints the picture of how your business did in a particular period
0:01:46 of time and measurement, whether that’s quarterly or yearly, or the cash statement then reconciles
0:01:49 that with what did you actually collect. Everything that happens on that cash flow
0:01:52 statement then ends up on your balance sheet.
0:01:57 The reason why it’s important to be able to present it in that fashion, it’s called generally
0:02:03 accepted accounting principles, so GAP accounting, is because that’s how everyone understands
0:02:08 that the comparisons are apples to apples when you look across companies.
0:02:12 When you are trying to figure out how a business is doing, what are the financials that you
0:02:13 look at?
0:02:20 Typically, the early investing, you don’t emphasize financial metrics that much, because
0:02:24 usually, there isn’t a mature go-to-market organization. I tend to focus much more on
0:02:31 KPI-type metrics, users, daily to monthly users, engagement, and things on those lines,
0:02:34 and then the financials tend to emerge over time.
0:02:40 Yeah, I would say I care most about two very high-level topics at the later stage. The
0:02:47 first is, can you demonstrate that you can have very persistent growth? Secondly, how
0:02:53 profitable will you be when you reach scale? I spend less time for later-stage high-growth
0:02:58 companies staring at their balance sheet than I do KPIs, income statements, and cash flow.
0:03:02 The main thing I look for in the balance sheet is the comparison for how much traction they
0:03:06 have on the income statement and the cash flow documents relative to the amount that’s
0:03:10 been invested in the company. For me, the most important balance sheet metric early
0:03:15 is how much capital is the company deployed to get to where they’re going.
0:03:19 How do you guys know when a business is truly profitable?
0:03:24 I do think you go to the unit economics and really understand them, but this is often
0:03:29 a lot of art as well as a good amount of science. Some of the most frustrating interactions
0:03:34 I’ve had with companies are where they’re presenting that their unit economics work,
0:03:38 but the business isn’t working. And so I had one where, okay, we’re capital efficient,
0:03:43 the unit economics are working, we acquire users, they’re profitable in three months,
0:03:48 and the company was hemorrhaging cash. It turns out the unit economics actually weren’t working.
0:03:53 The cash flow statement was the arbiter of truth and the analysis that the company had
0:03:57 done on unit economics was wrong. Yeah, I agree. I think lifetime value is one
0:04:01 of those traps that people fall into. They’re assuming, oh, our customers are going to stay
0:04:05 with us for five years, three years, so we’ve got plenty of time to do the payback,
0:04:09 but that’s a key driver to whether or not your unit economics work.
0:04:14 There’s nothing that’s less consistent in the market than how lifetime value to
0:04:20 cost of acquisition of the customer. LTV to CAC is defined. What I always counsel companies and I
0:04:26 like to see is very transparent calculations of what goes into the LTV side and the CAC side.
0:04:33 So the LTV to CAC metric that I like to look at is for the LTV, so the lifetime value side,
0:04:38 I always use gross profit, not revenue. And then I like to use a shorter duration than
0:04:43 founders typically like to use. So I like to use three years. Often founders present five years,
0:04:47 and the point I make on that is that five years is too uncertain and long of a period of time,
0:04:52 whereas three years is much more visible, and then use actual retention statistics that you’ve
0:04:57 experienced in the past to project those three years. The thing that I really try to emphasize
0:05:03 to founders when they talk about these kinds of metrics is, look, this is not about necessarily
0:05:07 showing investors. This is how you have to run your business. What am I spending on sales and
0:05:12 marketing? What am I spending on my R&D? And how much am I spending on G&A? And is that
0:05:17 the right level of investment that I should be making in my company? So you need to be as honest
0:05:21 with yourself as possible as to what all these things cost you and what you’re really generating
0:05:25 in terms of revenue, because if you can’t be honest with yourself, you can’t run your business.
0:05:30 What are some of the other really common mistakes or things that founders do in presenting numbers
0:05:34 that you’d want to help them correct or you’d like to see them do differently?
0:05:39 You know, one tell for me where business is probably struggling is when they come up with
0:05:46 North Star metrics, you know, KPIs, and then when they come back to report on them a quarter later,
0:05:50 they’ve changed. And then they come back a quarter later and they’ve changed again. And what I found
0:05:55 a little bit of pattern recognition is when the KPIs change all the time, it’s largely because
0:06:01 they’re not working. And the company’s trying to navigate through it. For me, you pick your metric,
0:06:07 you report on it. And ideally, your understanding of the business improves over time as your metric
0:06:12 and your models are either validated or unvalidated. That leads this interesting question, I think,
0:06:18 of the psychology and how you look at your numbers. So how do you manage your own psychology
0:06:21 so that the numbers are a tool, not this obsession where it’s like,
0:06:26 my obsession is I want to reach my KPIs. So I’m going to keep adjusting my KPIs. So I do.
0:06:30 You know, the reason they’ve so defined the three financial statements is it’s kind of truth-seeking
0:06:36 and trying to fool your investors or lack of better word or your board. You know,
0:06:42 I don’t want to let them know how bad things are. By not telling the truth to your key constituents,
0:06:47 you often run the risk of not telling the truth to yourself. And so I’ve had a couple founders
0:06:53 where sometimes they fall prey to it themselves, where they believe their own machination,
0:06:57 and then the board and investors can’t help them based on the truth.
0:07:03 What do the best founders do, especially in challenging moments, like when the finances
0:07:07 and the numbers maybe aren’t going your way, or you know that you do have to tell a difficult truth?
0:07:13 They typically acknowledge it. They take it as, okay, the truth isn’t what I wanted it to be.
0:07:17 So now, what can I do to change the business to improve the truth?
0:07:21 The only thing I would add to that that I’ve observed from some of the best founders of
0:07:28 later stage companies is they’re very careful not to drown themselves in KPIs. So you can actually
0:07:34 inundate yourself with KPIs, but the very best ones pick out a very few handful of metrics that
0:07:39 they think are the most important drivers of their business. And if they see those divert from where
0:07:44 they would like them to be, they dig in from there. So for example, Ali from Databricks always
0:07:50 focuses on the productivity of a sales rep because he believes that indicates health
0:07:55 of the business in many different ways. So how well is the sales organization actually functioning?
0:07:58 What are the market dynamics? What’s competition? How is the product performing?
0:08:03 And you do get a real force for the trees. I have companies that will present you 50 pages
0:08:09 of metrics based on the last quarter and you just drown versus what in here is really important?
0:08:13 What are the key ones? Are the one or two that matter the same for every company or does it
0:08:18 depend on the nature of the company and the stage that they’re at? I think to some degree,
0:08:22 some of them are the same. For instance, retention should matter to any business model.
0:08:26 You spent money to acquire your customer base. How long are you hanging on to them?
0:08:31 Yeah, I find they are consistent by type of business. So marketplace metrics typically
0:08:36 have a lot in common with each other, but they’re very, very different than e-commerce metrics.
0:08:40 The key e-commerce metrics typically center around the efficiency of customer acquisition
0:08:46 and LTB to CAC. A lot of marketplaces I work with don’t spend a penny on customer acquisition.
0:08:52 And so it’s got organic distribution or something like that. So comparing across models can be
0:08:58 challenging. Comparing within models can be very helpful. So for B2B companies, for example,
0:09:03 the efficiency with which you spend a sales dollar, whether it’s on a rep or marketing
0:09:07 or bottoms-up sales, inside sales, outside sales, is always one of the most important things that
0:09:12 you look to. Things change. Markets are unpredictable, which is something I think we’re seeing now
0:09:18 more than ever. How do you use finances to make better, faster decisions, especially in uncertain
0:09:24 times? You know, I started my career in finance. I ended up as CFO at the Disney Store. So it’s
0:09:28 near and dear to my heart. The typical finance function is conceived of as kind of keeping score,
0:09:34 the accounting control function, just reporting back. For me, that was necessary but not sufficient.
0:09:39 The finance function has access to all of the key data. And so I look at them not only to keep
0:09:45 score, but to score points to make the business better by leveraging their access to the information
0:09:49 and to the trends and to the unit economics to improve the business.
0:09:57 A good finance leader needs to work with the CEO to make sure that the company has
0:10:04 enough money to not just survive but thrive. So that is becoming super intimately familiar
0:10:08 with the business, not the financial statements, not the accounting that goes into developing
0:10:11 these things, because those just represent what’s happening at the business level.
0:10:17 They really need to understand how everything works and then where are the levers that you can
0:10:22 change, that you can pull on, that you can push on to accomplish the things that you want to do
0:10:27 as a business in the timeframe that you need it all backs up with the cash that you have on hand.
0:10:32 When I was managing businesses, I always had a mental model of how the business should work.
0:10:37 And that mental model typically, ideally, was consistent with the financial model.
0:10:43 eBay, back when I managed it, was a perfect economy. And eBay as a platform attracted every
0:10:48 leading finance professional who was into micro because it was one of the most pure examples
0:10:54 of a perfect economy. If there was an increase in supply, prices fell. If you change the fee
0:11:00 structure, behavior changed. And so it’s when businesses diverge from my mental model that
0:11:05 you really needed to pay attention. It’s like, why is the conversion rate going down? My god,
0:11:10 I’ve never seen it go down like that. That’s a big warning indicator for me. So I would
0:11:15 typically be pretty comfortable running the business until anomalies emerge. And then I just
0:11:19 would need to understand the driver of the anomaly. And I can’t emphasize enough how important it is
0:11:24 for companies to understand their bottoms up for how revenue is generated. I see a lot of people do
0:11:30 tops down forecasting. So the last quarter, we had whatever, a million dollars in revenue, 10
0:11:35 million dollars revenue. And then you go, okay, and historically, we’ve grown 50% or 100%. And so
0:11:41 we’re going to model something similar to that for the next year. And so that’s our number.
0:11:45 And that’s got no intelligence built into it whatsoever. What you have to do is double click
0:11:50 on that and go, okay, so we made whatever $10 million last year, how was it made? What was the
0:11:55 makeup of that customer base? Who’s likely to still be here? Who’s going to spend more? Who’s
0:12:00 not going to spend more? Who’s going to completely leave the platform? In marketplaces, you often
0:12:05 get two shots at bottoms up because you typically can build a model based on the supply,
0:12:09 or you can build a model based on the demand. Get an example at eBay. We would look at the
0:12:15 behavior of sellers and we had this many sellers growing this fast, doing this kind of behavior,
0:12:20 and then you just kind of roll them together and come up with a revenue estimate. Then we’d sanity
0:12:26 check it with, we have this many buyers buying this frequently, spending this much and coming onto the
0:12:31 platform at this rate. And then you’d run up that number. And ideally, the two would inform each
0:12:37 other. So one of the best CEOs who I worked with, who I partnered with was George Kurtz from Crowdstrike.
0:12:42 He had an exceptional business. One of the things when we were working together that we came to
0:12:48 realize was his gross margins were a little lower than most other software companies that we were
0:12:54 working with. He actually made the decision in one quarter based on that to try and experiment
0:13:00 where he made gross margin actually be part of the calculus for sales compensation for the reps
0:13:06 in that quarter. And his gross margins over the last three years have actually gone from 35% to 70%.
0:13:12 So a very operational tactical decision that can have a massive impact on the value of the business.
0:13:17 So I wanted to go back to a point, I think, Jeff, that you brought up of having this mental model
0:13:22 of your business and hoping that that matches the financials and how then you have those red flag
0:13:27 moments. And I think a lot of companies right now are having a red flag moment because of a
0:13:32 lot of circumstances very beyond their control. I’d love to hear, what are you telling founders
0:13:37 right now when it comes to how to think about their financials? This is one of the most significant
0:13:41 disruptions I’ve experienced. I’ve had a long enough career that I’ve experienced a bunch of them,
0:13:47 the bubble bursting in 99, 2000, the financial crisis of 2008, 2009, which by the way, we took
0:13:54 OpenTable public in May 2009. So the future isn’t dictated. But a couple of things come to mind.
0:13:59 One is cash is king. The income statement, throw it away. Just look at the cash flow statement.
0:14:03 How much cash do you have? How’s the burn? How are you adding or using cash over time? So
0:14:08 cash becomes completely king. Throw out your forecast because the forecast is now meaningless.
0:14:13 It was based on a bunch of assumptions that no longer hold. So throw the financial print out
0:14:17 and start looking really hard at things like year over year, which typically doesn’t lie.
0:14:23 And then just do tons of sensitivities. And you got to do it decisively. I always like this thought
0:14:31 exercise of how bad could this possibly get? Just let’s take the absolute worst. How bad could it
0:14:37 get? Because I think people tend to do the opposite. They iterate down of like, okay, I’m going to,
0:14:42 we’re down 5%. I’m going to plan down 10%. But if it’s going down 5% per day,
0:14:47 plenty down 10% just met your plans out late in two days. And so I found it helpful both from a
0:14:52 business, prudent cash management perspective, also from a mental perspective. Don’t let this
0:14:57 just continue to erode. And I get more and more depressed every day. Get really depressed one
0:15:02 day, look at reality and then try to change it. Yeah, I agree. So I was a CFO at a company
0:15:07 called Adbride in 2008. And I think that at first we didn’t want to believe that it could get that
0:15:12 bad. But we were an advertising network. And so unless you were Google and even they were impacted
0:15:16 by this, your customers weren’t going to advertise anymore. The marketing departments were decimated.
0:15:21 So there were situations where we were like, some of this is just going to become zero.
0:15:26 Contracts that were signed are now just getting outright canceled. So we made the decision to
0:15:31 cut really deep and as quickly as possible. Because we knew that even if we got it wrong,
0:15:37 at least we could then rebuild the company and do it only once. And then your employees then
0:15:41 are told, hey, we’ve made this big decision. Here’s what we based it on. Here’s our cash
0:15:45 position. Here’s what we’ve sort of expecting in terms of worst case scenario. You bring them
0:15:50 into that circle of trust of what’s happening at the company. But there’s asymmetric potential
0:15:57 issues. If you underestimate how bad it’s going to get and don’t deal with the situation quickly,
0:16:01 the outcome is very well be you lose your company. Yeah, death by a thousand cuts.
0:16:05 Yeah. And if you overreact and it doesn’t end up being as bad as it would have been,
0:16:11 you might have suboptimized your company for some period of time, but it’s alive.
0:16:16 So for me, the mistake is to underestimate the potential versus overestimating.
0:16:20 Yeah, I want to go back to something Jeff said, which was this notion of throwing your forecast
0:16:26 out the window. Very much agree with that on the top line on your revenue. But you have this whole
0:16:30 base of costs that are under your control, those are your operating expenses. And so
0:16:36 we’ve spent a lot of time focused with our companies running sensitivities of, hey, this is
0:16:41 your operating expense budget. And what’s an operating expenses are your salespeople, your
0:16:47 marketing people, your CFO function, your HR function, your engineer’s product. Those are all
0:16:54 people and costs that you have as a base. What happens to that cost base in order to preserve
0:17:00 cash under various scenarios of revenue decline? And so I think that’s the way that you have to
0:17:05 be managing your business on a very, very granular level. And especially since companies,
0:17:10 especially startups, they staff in advance of growth. And so you have to really be honest with
0:17:16 yourself. I want to just also chime in and just say, look, these are all very, very hard decisions.
0:17:22 And I think Caroline and Jeff, especially because you’ve been in the seat of operators
0:17:27 during really, really trying times, you’re probably pretty diagnostic about it. But suffice
0:17:32 to say it’s hard decisions, you know, people’s jobs, decisions not to be taken lightly.
0:17:37 And that’s why there are cases where it is a death by a thousand cuts because people are reluctant
0:17:43 to do those layoffs, make those cuts. And believe me, you don’t sleep when you have to make these
0:17:48 decisions. It’s so tough. So I don’t take that lightly at all. But when you’re running a company,
0:17:53 number one is making sure the company can make it through to the other side. And so you have to
0:17:58 make these really tough decisions. And believe me, I understand how difficult that can be.
0:18:00 But you can’t kick the can down the road on some of these things.
0:18:06 Everyone in the organization knows that the proverbial shit is at the fan. And so if the
0:18:13 leader is unwilling to acknowledge that with the team, that for me creates a crisis of confidence.
0:18:20 I always found it way better just to call it what it is, share it, try to enlist the team. And
0:18:24 do you agree with this version of reality and try to get agreement? And then it’s like, okay,
0:18:29 what do we do? But denial and trying to hide it from your team is a failing strategy completely.
0:18:34 And I understand the human psychology around that because I think people don’t like to give
0:18:39 bad news. And so I think the natural impulse is to hide those things. But these are the moments
0:18:44 where you have to actually be the most transparent. Talk about why you’re doing what you’re doing,
0:18:48 how much cash you’ve got left, how much you want to preserve. And what I find is when you do that,
0:18:52 when you bring everybody into the fold, they all become part of the solution. So they understand
0:18:58 that cash is king, and they’ll figure out ways to be even scrappier than they might have been
0:19:03 otherwise. Jeff, you’ve lived through some crises already. Go back to a time when you were facing
0:19:07 a crisis where things were rapidly changing. You were having to make some of these difficult
0:19:13 decisions. What was a day in a life like then? And what were you doing, especially with regards to
0:19:20 the financials? I got a good one for you. So OpenTable in mid-2008, the board decided it’s
0:19:25 time. Let’s go public. The market wasn’t good. But for a variety of internal reasons for the
0:19:29 company, we decided, okay, we have to, quote unquote, get the puck on the ice. And so we got
0:19:36 ready for the IPO. And we did our bake-off in, I think it was August 2008 and did it like on a
0:19:42 Thursday. And Friday, we informed the six banks whether they got on the offer or not.
0:19:46 We told Lehman on Friday they didn’t get the offer. They went out of business on Saturday.
0:19:51 We told Merrill Lynch, they did get the assignment to take OpenTable public,
0:19:57 and they traded to Bank of America on Sunday. So over that weekend, the number of people dining
0:20:03 in fine dining restaurants in America went down by 15% in one weekend. So we had the org meeting
0:20:07 Monday morning. I walk in, sit down, all the bankers there, all the lawyers there, and this is not
0:20:15 going well. Our business is in free fall. The bank just changed. The consumers terrified. And so
0:20:20 it was pretty clear we could not proceed with the IPO at that point because we couldn’t predict it.
0:20:25 But then we just said, okay, what can we predict? And so we put it on hold for three or four months.
0:20:30 And it turned out that the consumer kept dining in restaurants at 85% of what they had the prior
0:20:34 year. And all of a sudden, we got confident that the business was predictable at that point. And we
0:20:38 restarted the process and went out. And it ended up being a successful offering.
0:20:42 How were you looking at the financials during that time? How did those come into play as you went
0:20:47 through that? We were watching the year-over-year change in reservations of people dining and
0:20:52 reservations made daily, just like, okay, where’s this business going? Because if it kept falling,
0:20:57 one of the scenarios we were concerned by is more and more people would stop eating as they got more
0:21:03 and more nervous about the economy. And we’d go from revenue of X to revenue of like 0.3X. And the
0:21:08 business would have been hugely stressed at revenue of 0.3X. So we were watching that the
0:21:15 one key North Star metric of diners per night, year-over-year, maniacally. And that ended up
0:21:19 giving us the confidence to restart the offering. How does a startup or a founder right now approach
0:21:24 contingency planning around their finances, especially if you’re a high growth startup that’s
0:21:29 been going through cash quickly and been pretty aggressive with your risk taking until now?
0:21:36 For me, you don’t scenario plan constantly. But when a shock hits the system like this shock has
0:21:44 hit this system, hit the world, is you want to plan quickly, even if it’s bluntly. If I was running a
0:21:50 business in this environment, I would get the expected outcome. Maybe it won’t go there quickly,
0:21:55 outcomes are slightly better, but also just what is the worst case? Where could this go? And then
0:22:02 you build your response if each of those comes true. And for me, you put much more time into the
0:22:08 plan, what if, than you do in the building the sensitivity scenarios. One of the less productive
0:22:14 activities is making that sensitivity beautiful and accurate. And it takes two months to come out
0:22:19 with and the company’s out of business. Yeah, just take the bluntest assessment. David and his team
0:22:24 has done this for a few of our internal companies. Yeah, what we did is we basically took every
0:22:30 company’s financials and started with revenue and said, okay, let’s start with your budget.
0:22:36 And then let’s run sensitivity analyses for your revenue. Let’s assume you hit your budget,
0:22:42 that you’re flat, that you don’t grow, that you declined by 25% or that you declined by 50%.
0:22:48 And then we compared that with a company’s operating expense budget. And across all those
0:22:54 different scenarios, if you run your current budget of operating expenses, if you assume you
0:22:59 don’t grow your operating expenses, and then if you assume you decline your operating expenses by
0:23:05 25% or 50%, what is your cash runway in each of those scenarios? And we plug that in for each of
0:23:11 our companies and gave it to them. And I think it’s just a helpful way for them to put some parameters
0:23:17 around, hey, if things get really bad, this is what our runway is. And often it helps them just to
0:23:23 start thinking about, okay, how do I contingency plan in the event of flat revenue? I had never
0:23:27 even thought about that before. If that happens, I only have this much runway. Maybe I should take
0:23:32 action. And another thing that I would say to you, a lot of times companies are building things as the
0:23:37 plane is in the air, and they solve their problems linearly by throwing bodies at it. This is an
0:23:42 opportunity to be able to potentially refactor your code base, to shore up infrastructure,
0:23:48 to build internal tools to make your teams more efficient, so that when you do come out on the
0:23:53 other side, that you are primed and ready to just hit the ground running and run in a million miles
0:23:57 an hour, because you have now built the foundation that you need to be able to really scale your
0:24:02 business, a company called AdBright. We were an ad network. This was before Amazon Web Services was
0:24:07 really a big thing. And so you had to have your own data centers, which means you had to buy
0:24:13 equipment. So we were a very capital intensive business. And what we realized was that we weren’t
0:24:17 going to be able to afford anymore to be constantly replacing our servers, because we just did not
0:24:24 have the money to do it. And our CTO had been playing around with this thing called AWS and
0:24:29 brought it to us and said, one, we can’t afford to upgrade our servers, even though we need to.
0:24:34 And two, this is going to probably in some ways improve our gross margins, because now
0:24:39 we can flex up and down when we need the capacity. So can we give it a try?
0:24:46 This was cloud-based. Anything was still pretty new. This was 2008, I think, AWS launch in 2006.
0:24:50 So we became a data customer of theirs. At the end of the day, when we came out of the crisis,
0:24:56 we were pretty much a cloud-based ad serving company. We deprecated all of our data centers
0:25:01 when we just moved everything to AWS. So what bottom line advice right now are you giving to
0:25:07 founders? One anecdote about the 2008 credit crisis, when housing prices dropped like 30
0:25:11 to 40%, if you were to interview people on the street who owned homes, you asked them, hey,
0:25:15 what do you think the US residential market has done in terms of real estate values?
0:25:20 They would across the board say, oh, it’s down 30 to 40%. And then they would be asked,
0:25:25 all right, what do you think happened to the value of your own home? And they say, oh, nothing,
0:25:28 nothing at all. It’s still fine. It’s not down at all. And it’s like, well, that’s not how averages
0:25:33 actually work. And so no one believes that it’s going to happen to them. But believe me, it is
0:25:39 happening to them. And that is the thing that I want founders to understand. You are not going to
0:25:43 be impacted asymmetrically compared to everybody else. You’re not going to be that outlier more
0:25:49 likely than not. So we’ve had a lot of advice in here on confront reality decisively, plan for
0:25:56 the worst case, scenario plan the worst case. And psychologically, that is pretty darn challenging
0:26:01 on the founder. I mean, I’ve lived it, I understand there. So that brings the point that it’s just
0:26:06 incredibly important for the founder to manage their own psychology. And I think probably the
0:26:10 best resource I’ve read on that is Ben’s book, The Hard Things About Hard Things. You flip from
0:26:17 peacetime to wartime, people are looking for you to lead, and you’ve just got to take the horn.
0:26:23 But I always was most uncomfortable with my personal psyche when things were going great.
0:26:30 I mean, when OpenTable was trading for like 21 times forward revenues, which is an absurd valuation,
0:26:36 I was jumping out of my skin. But once you confront the fact that, okay, we’re in one of those
0:26:42 moments and I need to lead out of it, I actually found it after an absolutely miserable X hours.
0:26:48 I found it motivating. We can get over this, let’s show them what we can do. For me, the CEO and
0:26:54 founder needs to confront reality quickly and then they need to lead. And you can lead your company
0:26:58 through these things and get to the other side, then things will get better again. But the biggest
0:27:03 thing is manage your psychology actively. I just want to thank you, David. Thank you, Jeff. Thank
0:27:07 you, Caroline, for joining us on the podcast today. Thank you. Thank you.