How We Podcast

AI transcript
Hi everyone. Welcome to the A6NZ podcast. I’m Sonal and we’re here today because we’re doing our
500th episode of the A6NZ podcast. Episode 499 was with Margaret Wenmarkers, the head of marketing
at Andreessen Horowitz, who built the A6NZ brand. And for the 500th episode, we’re here to answer
for the first time. So it’s sort of a behind the scenes about how the podcast works. A lot of
frequently asked questions people constantly ask us. And that also I got on Twitter, our special
guest host and interviewer is A6NZ general partner, Connie Chan. Connie will also ask any
questions she’s interested in because she’s actually very into podcasting and investment
podcasting as well. So that’s a context. So let’s first start with the history of the A6NZ podcast.
Tell me about how it got started, how you got involved. It was actually created before I even
joined. There was already a culture of writing at Andreessen Horowitz before they even built
an editorial operation. I mean, there was a popular P-marker blog. There’s Ben Horowitz’s blog and
then book. And this all happened right before I joined. And they were also already writing blog
posts about announcements. And they also had done very few specific op-eds that we talked about
in episode 499. So that was a context. And then I believe the story was that Dixon, Chris Dixon,
we all call him Dixon, came in one day as like, we should do podcasting. And Dixon was an early
blogger. So my speculation is that I think he thought it was like the evolution of that sort
of type of communication. And so the podcast was pushed by Chris Dixon and Kim Milosevic was
hiring up the editorial team for Andreessen Horowitz. And she first hired Michael Copeland,
who got the podcast off the ground and was the host for the first year. I started producing it
behind the scenes, not hosting about three months in. And then I only started hosting about a year
in, but I’ve been very involved since about three months in. There’s been a lot of stuff behind
the scenes. So Lovelessut, what year was this? I think it was late 2013, early 2014. And it was
actually that exact same year I had gone to XOXO. And I heard a talk from Marco Arment about the
resurgence of podcasting because it’s been around for years, as you know. And he talked about how
brands and others, when they do podcasting, there’s a certain intimacy that comes with it.
And so it has a very similar feeling to blogging in that there’s parallels in the authenticity
and intimacy of communication. In fact, the way I think of it is that if you think about our history
of oral storytelling and how we can all sit around a fire in the olden days, that used to be an
experience of one-on-one. And now we can do the almost the exact same thing where you have the
feeling of a one-on-one intimacy, but it’s scaled to thousands and thousands. In our case, hundreds
of thousands of people. What gave you the inspiration that we needed to double down on podcast?
You invested so much of your thinking and your energy into this. What was it that audio was
unlocking for you? It’s funny that you say audio because I actually did not think of it as audio
then. But I love that you’re saying that because that’s how I do think about it now. At the time,
quite frankly, I was scratching a personal itch, which was I had come from Wired where I had the
opportunity to edit hundreds and hundreds of different thinkers, writers, famous thinkers,
emerging tones. But all in written form. All in written form. I’d never actually done audio, by
the way. I didn’t have that experience. And I got here and I kind of was like, oh my god, I love my
team. I love the partners, but I’m going to kill myself with boredom if I only have eight partners
where I had come from hundreds. And so for me, the podcast was a way to answer your question about
doubling down to bring more diverse voices onto the platform. Right. Because before our blog posts
were mostly just written by general partners. That’s right. And then with podcasts, now you
open it up to more internal voices, more external voices. Exactly. And in fact, the external voices
was built on what I did at Wired and building the expert opinion section there. And I had three
views on it. So first was, I think it’s really interesting that in our modern world of media,
we even have intermediaries at all to dilute the voice of an expert. Okay. So in your case,
perfect example, you wrote a beautiful WeChat piece that we worked on. I remember it ran the
exact same day that David Pierce at Wired wrote a piece about WeChat, which is also very good and
well done. Very different pieces. Yours was a first person, first principles, first party expert
take that was not based on reporting it, but in using it, observing it, bringing your own thinking.
It was what I described as an ethnographic kind of piece. That’s first person expertise.
David’s was reported. He talked to people at WeChat. He did interviews. He was coming out
as a reporter. Also great. But in my view, why was a venture capital firm focusing on reported
stories when we have a huge network? This is our defining thing, a network of networks, in fact.
So why wouldn’t we bring in experts on various topics, but not have them diluted in their expertise?
And so when they come on as guests, we have the first person versus the third person. That was a
huge important thing to me. And it’s also my bias for builders and makers. And so as you bring in
external folks, though, I mean, that puts a lot more pressure on how you program it,
how you research for it and prep for these podcasts. Talk to me what that’s like.
Yeah. Other questions people have on Twitter. One of the most common questions that came up
is how do we program the podcasts? Like, how do you even decide who to bring on?
So to give you some more context, I think of a podcast in three phases. There’s everything
that happens before, during, and after. I would say that the majority of the work is before and
after then during the podcast itself. Okay. So let’s talk about programming. And then I want
to dive into each of those sections. Okay. So in programming it, I think of every episode as an
op-ed or a feature story. And so just like an op-ed or a feature story, you think to yourself,
what is the argument or topic or angle and what is the take? What’s the differentiated
fresh view and then who are the people to have that? So do you take that same like editing
framework as a written author and then think, okay, I need to have one main argument and a
conclusion at the end? So obviously conversation is so much more organic than that. Okay. You
actually don’t really decide the argument upfront. I would actually even argue writing is organic.
Sometimes you kind of know what you want to talk about, but you just kind of go with it and figure
it out as you write it out. It’s not like we walk into a room and say, hey, I’m going to come on
the podcast and I’m going to argue X. That never happens. But what we do do is figure out, okay,
so let me think of a good concrete example. Let’s say we want to do a topic on emojis,
which is one of my all-time favorite podcasts. And there’s lots of different ways to take it.
Well, okay. I think it’s really interesting that emojis are pervading our culture and that yet
at the same time, people have to propose through proposals specific emoji to get into the set.
So what if we did a conversation with someone who proposed the dumpling emoji? So Jenny Aitley did
this. And then Fred Veninson, who actually translated Moby Dick into all emoji using
mechanical Turk. And so you have two people at very different kind of perspectives on it.
But here’s the thing they have in common. Very different takes. Both, however,
are first principles, non-derivative experts who are going at it at a first person way.
Secondly, through this lens, we can then bring in all the concrete and abstract,
tangential ideas of governance, open versus closed, proprietary systems, how to design,
Apple versus Android, Twitter, Facebook, and use that as a concrete way to have a really thoughtful
conversation. It’s funny because you think the podcast is about emoji, but it’s actually about
how innovation comes about when you’re trying to have a system across all. So how do you even decide
I want an episode on emojis? Oh, well, that’s just what editors do. And this is actually probably
the broader context for the editorial operation, which is you always ask yourself, what are the
topics we want to cover and how. And you may not know the exact how, but you have an idea of how.
And one of the things that I always tell people, if we were to take this up even a notch, the
editorial operation is about innovation. And Margaret talked about this in our past episode
499. And I’ve had a rule of thumb. People ask me this on Twitter. So I’m going to answer this question,
which is whenever I think about any kind of brand or lens for content, I want it to go through two
words. And the funny inspiration for this, by the way, is from Domino Magazine. They once did
a feature about how you can find your signature style. And there’s like a stylist who would come
in and say, Connie, you are urban warrior. And this is your two word word to describe your style.
Here at Andries and Horowitz, it’s innovation brand. When I was at Wired, it was informed
optimism that came from Chris Anderson. And when I was at Xerox Park, it was entrepreneurial
scientists. And my point is that you use that as a lens with which to decide what to run,
what not to run, and how to treat it, and even how to edit it. And that serves as a filter
for what makes a cut and what doesn’t. We are about telling the stories of the future,
building it, explaining it, and really how tech changes our world. So that’s a lens.
So on the programming piece, how do you actually choose which guest to bring on?
Right. So this is, again, going back to the same philosophy I had for the expert section.
I am looking for the expert, not a expert. And again, going back to this idea of an
individual op-ed or a feature story for every podcast, you ask yourself, if you’re doing
a feature story, who are the third party experts you would bring in? So similarly,
we look for either the expert or the next best expert or someone who has very specific expertise.
We don’t really love consultants and derivative experts and people who just talk about the thing
versus do the thing. And then I look for a complimentary expert. And this is sort of the
person who can add texture. We don’t want two people constantly agreeing. We also don’t want
them completely disagreeing. Sometimes people talked about in the early days, we should do podcasts
where you have pro-con. And again, it’s like, exactly. I love debates and Oxford style debates
in particular. But what I find, what I call the panel problem where podcasts becomes a conference
panel. I don’t know if you’ve seen this at every conference you go to. Inevitably,
the smartest people, four people, so smart on a single panel, it’ll be the dullest, dumbest
conversation. And why is that? It regresses to the mean. And to me, it’s a pure statistical
thing. It’s like in statistics, if you sample from the extremes of a data set, you essentially
regress to the mean. It is literally the exact same thing happening when you do that with experts.
So having a pro and a con, it’s actually a case of like negating the conversation. You want to
have a thoughtful nuance conversation. So I like to avoid what I call one note narratives. I don’t
want an expert who has just a single observation. They’re going to do it like 10 different ways.
Do you kind of give them a guidance on what you’re going to ask about? Tell me about the prep
on the actual figuring out what questions you want to run. Do you let people do a dry run as it’s
scripted? What do you do? So the process is that I tell them, this is actually baked until all our
emails and how they get on, that they are not supposed to prep. Now, people hate that because
they want to prep. Yeah, I’m sure everyone wants to know what you’re going to ask them.
Right. And in fact, I kind of realized early on like, oh, someone will just because you like
that. I mean, everyone else likes it. What’s the downside of prepping? That’s a great question.
So one of the things I learned when I was at park and I worked with a really good event producer
for this event that we were hosting O’Reilly Media’s Make. It was a first inaugural hardware make
workshop. And one of the event producers on that said, I never put two people in a green room before
an event because inevitably everything they say on stage will refer back to what they were talking
about right before coming on stage. And you’ve probably seen this at many events. And the audience
doesn’t have that exact same sharpness that they feel when they hear that idea. For me, when I record,
I start the recorder before the person even walks in the room and I stop it only when they leave
because the best stuff comes when it’s a little bit unfiltered. So when we prep to answer your
question, what I tell people is I don’t want you to actually tell me what you’re going to say,
because actually then the second time, if you say it, it’s going to be 10 times worse.
It’s much better raw and real the first time. Of course, the speakers get freaked out though,
because they’re like, what if I sound like an idiot, which means that you have to do a lot of
editing. Oh, we can come back to editing. I mean, it’s not by accident that you make a lot of us
sound a lot more eloquent than we do in real life. Well, first of all, it’s not an accident because you
guys are all also experts. Let’s just be very clear that one of the reasons this works is that A6NC
does have experts and one of the number one rules of thumb I use for all editorial written
podcasts or otherwise is the concept of what I coined a number of years ago called writer topic
fit jokingly WTF. And the idea being that the writer has to have the topic and fit for the
expertise. This is not credentialist. It could be earned expertise. It could be data, it could be
whatever, but they have to have that. So the people who are freaked out about their executives
coming on and not having any idea, we don’t send questions in advance. I like the conversation
to be very organic. One of the questions people asked on Twitter was, do you prepare their script?
And the answer is no, there is no script. What we do at the beginning of every episode is we,
and sometimes these people have been in met each other. Sometimes you’re meeting Christina
shoe for the first time and the three of us are doing a conversation about stickers and memes
and live streams. So in that case, what we do is we’ll spend literally five minutes at the beginning,
maybe less just talking about what we want to talk about, meaning topics, but not the
actual argument. Exactly. Because then I get mad and I say, no, no, no, because people inevitably
start sharing what they’re going to say. And I’m like, wait till we get to that part. So we do
that. And then we just go through. And then this is where the editing lets you then reorganize it
into an arc that makes sense. And by the way, by arc, I don’t mean it has to be linear,
like point A, B, C, D. In fact, I want it to be nonlinear, slippery, raw, with an edge, not always
clean and clear. At wired, I had a phrase, which I use for my op-eds, which is three turns of nuance.
Like I like that kind of thing. What I’m doing with the editing of the story arc is again,
just listening for how the listener is going to move through it. Do they hear the organization?
Do they have to work to follow it? Can they just naturally flow along and learn as they go?
Okay, so that means you’re recording how much footage an hour, an hour, and that gets edited
down to what? Anywhere from 20 to like 45 minutes. 20 to 45 minutes. Right. But I want to say something
about that. This is why there are what I consider two types of editors. There’s what I call shaping
editors who are people who love as much raw material to work with as possible and then to kind of carve
out the arc their own way. I think of this a bit like a sculptor who’s given a slab of marble
and figures out the shape. That’s what I like to do. And then there are editors who take what
they’re given and they do a really good job figuring out how to rearrange it, how to put it
together, think about it. That to me is a more straightforward type of editing. As a shaping
editor, I look for maximum optionality in my recording, in that hour recording period,
because I want to have enough material to carve out the thing that I’m working on. Whereas for
some of the other editors- So when you’re editing, it’s not just taking out ums and blank spaces?
Oh my God. No. And in fact, I want to just deconstruct a myth here because we get a lot
of flak for people saying, “You guys remove the brets and you guys do this.” And it’s actually
ironic, we never do that. What we do do in fact is add brets because our people talk too fast
and don’t pause. And so sometimes we have to manually slow them down. Mark Anderson does speak
very quickly. It’s not just him though, he does. By the way, why do you joke about him when you get
like a written transcript is for every written page, for every page of a transcript, you can
estimate about two minutes. And in his case, it’s like one minute, the density is so high,
but the rest of the firm talks pretty fast too. And so PSA for everyone, do not listen to our
podcast at 1.5 or 2x speed. You should only listen to it at regular speed. But back to your
question. So it’s not ums and aus. I mean, we do do some tics. So my rule about tics is we don’t
want to remove all those tics, actually. We do however remove tics that are a little too repetitive
to the point of being disruptive to the listener. And I, by the way, have this tick there. You’re
going to love this. So when I first started doing podcasting, I was only a behind-the-scenes person,
so I really was insecure about being a host quite frankly. Right. You didn’t host the whole first
year. I don’t really want to host because I just felt like, no, no, I’m a behind-the-scenes person.
What are you talking about? I’m an editor. I can’t be a host. So it’s funny because everyone hates
Son of their own voice. And I noticed all my tics. And the first one would be like, uh-huh, uh-huh, uh-huh.
Got it, got it, got it. Mm-hmm, mm-hmm, mm-hmm. Like, like, like, or yeah, you’re right. Mm-hmm,
various versions of this. And so what I would do is I would hear my tics and then systematically
decide conscientiously not to say them. Guess what happened? Well, they went away.
And another one popped up in its place. It was like a game of whack-a-mole. It didn’t matter
how many tics I got rid of. A new one just jumped right up in its place. And so my theory about this
is, and maybe it’s grounded in science, I don’t know, I’ve never looked it up, is that there’s
something psychological with how we use these tics, whether it’s an anxiety management or a natural
way of thinking or like some kind of dead space words between thought-out words, like subconscious
words even. Anyway, that is how people talk. So you edit those sounds or specific phrases and words,
but the actual content, is it moving things? So a lot of the people on Twitter asked about
how do you edit the podcast from, you know, the whole, it’s a story arc to rest. So the edits are
to optimize for what I call insights per minute. There’s three things to that. One, it’s that
if you have a non-cultiv personality show, and I talked about this before, how there’s a taxonomy
of types of shows, and cult of personality shows, which are very host and personality driven.
You mean like a Joe Rogan show? Like a Joe Rogan, that’s a perfect example.
The audience is following Joe Rogan, they almost don’t care who his guest is, of course they care
if it’s Elon Musk or someone else, but because of that combo, they’re willing to listen for three
hours to the two of them on air smoking pot. Basically, you don’t want to abuse the user’s
time. It’s not so much to abuse the listener’s time, it’s that ear shares and you mind share,
you’re competing for that share. So the show has to be differentiated, and if you want people to
listen to your show and it doesn’t have a cult of personality, then you have to make sure it’s
resourceful, and they have a high insights per minute. So their time and their payoff is worth it.
Insights per minute. Is that an actual metric, or is this like a new metric?
No, it’s a thing that I coined, but it’s not like we measure it formally, but it’s what we
listen for. And to answer your question about starting it off, it matters more in the first
five to 10 minutes, because it’s just like editing an article. So I call myself a chartbeat editor,
because when I was at Wired, I was obsessed with the leaderboard, and I, of course, like love
seeing where I was on that, especially because I had a flailing section that I wanted to take to the
top. And so I was very motivated by that. But then what I noticed in chartbeat besides a leaderboard
is you saw where listeners, or sorry, where readers dropped off. And that was super valuable
to me, because then I started seeing patterns of, oh, well, if people drop off here, I need to work
harder to really get the nut graph up here before the third paragraph. And the more I work to make
sure that every sentence is calculated to keep the listener, the reader engaged, the better the
piece. And then by the middle and the end, when they’re committed, you have a lot more room to
be loosey-goosey and fun. So in the podcast, that’s the exact same thing, the chartbeat model
that I brought here. This is how I learned editing just by doing it. So to me, the first three to
five minutes are incredibly important, because that’s the highest drop-off point. So we use the
intro as a technique. This is why we actually record our intros after the fact, not before.
Introing who the person is. We intro who it is, the topic, the range, and there’s various ways
of doing it. We experimented for a while with having snippets. We did all kinds of experiments
throughout the years. But what I find is that the intro is a tool to let you start the conversation
in medias res, which is the term from literature for starting it in the middle of the story.
And why that’s so important is if you don’t have a cult or personality show, like a Joe
Rogan and an Elon Musk, if you have seen someone new who’s really smart, but no one’s heard of,
if they start out with their personal story, that’s probably going to be boring because they’re not
bought into this person. They don’t know who it is. However, if you start with their advice
and the thing that you find resourceful and useful as a listener, then the listener is going to be
like, oh, that’s interesting and remain hooked. And then by the middle or the end, you can then
weave in their story. So the editing is about reflowing and rearchitecting that arc for that
type of journey of a listener through the entire episode. And, you know, back on this note of
scripting versus not scripting, one of the folks on Twitter asked, do you guys do these sort of
informal hallway style conversations? And the answer is that’s how they actually started.
But what I found is that if you’re really trying to grow the show, it was only when we started
editing it that it significantly, you look at the charts, it was like upward curve with the edits.
And that goes back to the fact that if you’re not a cult of personality, people don’t, I mean,
doesn’t it bug you to hear people just chit chat when you’re not, you’re kind of like, get to the
point. So we do do hallways to all conversations, but mostly they’re just kind of these organic
conversations that are working towards some point of view. So tactically, how are you doing this?
Are you using a document? Great question. One of the questions people had on Twitter was about
the technology stack we use. So a couple of things on this. So I’m embarrassed to say that
the way I edit podcasts is by starting with the transcript, because I’m a word person and do like
a rough paper cut. And I actually do this without listening to the podcast. After the paper cut,
the technical audio editor turns it into a first cut. And the reason I do it this way
is because I want to see the whole shape of the narrative without being distracted by the sound.
The problem with that approach is that when you see on a text is unidimensional and flat, whereas
in voice, it’s much more multidimensional. There’s multiple factors. Yeah, like I might be super
excited in one sentence. Right. And then you suddenly have up talk and down talk. And you
can’t put that right next to another sentence where I’m quiet. That’s exactly it. So that’s why
it’s kind of a dangerous method, which is why the tool descript is a really interesting one because
they actually democratize the process. So to me, the first round is about seeing the global arc.
The second round is about listening to it and really seeing how it truly works and flows. And
the third round is really about sort of polishing it and making sure it just has this ease of listening.
So when you’re editing, can you boil that down to principles that we should take away when we
think about editing stuff? Yeah. So I guess the number one thing I would say is the biggest
difference between text and audio is that audio is a living breathing organism. So every change you
make introduces a new interaction effect. It’s like you’re adding a new variable. And so every
time you decide like in this cut, I’m going to do this, when you listen to your next cut, it messes
something else up, which is why tools like descript are so important because they shorten the time
between what I call the design and manufacturing phase of designing something like a semiconductor
chip. The ability to have that sort of iterative feedback loop is critical, which is why all the
new editors are getting trained on descript. So the living breathing organism, the different
framework required then as well. Yeah. It means it’s not unit dimensional, which I mentioned,
it has multiple layers. And I described that for every podcast, there’s like five dimensions,
or five levers even that you can use. So one is obviously the content itself, like what the substance
of what people are saying. One is the energy of the individual speaker, their tone, their excitement,
do they sound flat? And that can’t be edited, can it? Well, not really. I mean, you can actually do
some manipulations like raise a voice a little to make someone not sound so flat, but you don’t
want to distort the voice. Right, right, right. The third thing is charisma. Sorry, that’s a charisma
of the speaker. So that’s not just their level of energy for how they talk, but their sort of
charismatic way of drawing people to their ideas. Is that editable? Not really. You can do other
things though, because what I find with charismatic speakers is that they often also talk in platitudes.
And so one of the things that I tell our editors is you actually don’t want to be efficient with
the words. You want to cut the platitude statement and then keep the specific wonky statement. So
then they don’t come off as like BS. And then the fourth one is chemistry, which is the interaction
of the guests all in the room. So what do you do with chemistry when sometimes we’re meeting
that person for the first time? I don’t think having a pre-meeting helps with that chemistry.
So I think a great episode for this maybe as an example would be me and David Yulovich
when we did a podcast about what time is it and he has a fun chemistry and the two of us are just
very irreverent. This is when he first joined. It was about six months in and we did a podcast
about his career and the editing side. I put his story at the end because I don’t know if people
know him that well. So we started with his advice for founders. So that was like one of the arc
decisions I made because by the way, the fifth variable is arc or narrative flow. And anyway,
we have such chemistry because we had this like fight in the middle of the episode about how to
pronounce GIF or GIF. Which way do you say it? Oh my God, I say GIF, not GIF, which is what he
says. Ugh, I don’t even get me started on that. But like, you guys listen to that episode if you
want to leave a drop on that fight. But I kept it because it conveyed a certain chemistry. So to
answer your question, you have these five levers and the job of the editor is to take the material
they’re given with, whether fully all over the place like my material or more linear arc like
others, and then shape it into what it needs to be and edit it. So often that means removing
redundancies, but not to the point of being so efficient that it sounds like mechanical. It
means tightening flow, insights per minute. And then this is a beauty now of this five framework.
If you have like okay chemistry, but great content, you can work with that by rearranging the order
because then you keep people hooked by the flow. If you have wonderful chemistry and energy, but
very little substance, you can shorten it. So you have the energy, but reduce the length because
the payoff is so low. So basically what I’m saying is you can use one of these five levers and
manipulate them to get more or less dial it down or up or down to get what you need. And of course,
there’s only so much you can do. But what’s really cool is Descript has a company called Liarbird
and they are doing synthetic audios. And what’s up with timing? Like, is there a sweet spot for
how long a podcast should be? Yep, this is so funny because what I found when I asked people,
like, what’s your ideal length for a podcast? Guess what? The answer was exactly proportional
to their commute or workout time. So if their commute was 15 minutes, that’s what they thought
that was a perfect length. If there was 45, that was great. Is there like a time that you aim for?
No, we don’t aim for a time. It seems like the sweet spot is somewhere between 20 and 30 minutes.
My philosophy about time and length, and this is so strong both for written and spoken content,
is I think discussions about length are so arbitrary, religious debate, it should be
as long as it needs to be. So if it’s gratuitously long, cut it. If it needs to be longer because
we’re going in depth and it’s so interesting, why would we cut that arbitrarily? But this is
the caveat to that. The payoff and the insight per minute has to be proportional to the length.
So if the person is listening for a long time, there’s not enough payoff,
that’s a complete waste of their time, they’re never going to trust you, you destroy that trust,
you lose them as a fan. So that’s the first thing. The second thing to your question about the ideal
length, I do believe that short form podcasting is a really important form. That’s one of the
trends in the space. And I had a moment, kind of an insight was, I had this realization that,
gosh, all these other people are doing news shows like the New York Times Daily and Vox Today
Explained. And again, it’s a reported model, third party experts, which is great. But why aren’t we
doing a first person where people who know this industry are commenting on it directly
and do that way? So I was like, I want to do a news show and news podcasting is actually a
growing trend. And then why don’t we combine it with short form? So then I thought, let’s do
it for 16 minutes because Andrews and Horowitz, A16Z, 16, why not? So it had 16 minutes.
Now I think this one must be much harder to edit though.
Well, first of all, I’ve not gotten it to 16 minutes. Sometimes it’s like 17. Sometimes it’s
19. I try to only have it below 20s. A couple of times I’ve abused it. So for the first episode,
I tried doing it where everyone just did one full take a few times. And then what I had found was,
this is actually true for you in particular, you said different things on each take. And I was
like, Oh my God, what she said there was so good. And what she said, the other take was also really
good. So I use the editing to then seem together the best parts of what you said. And if you again,
think about insights per minute and the fact that you have only this many minutes,
how do you add value for the listener? You want the highest insight per minute,
hence the editing. So yes, it’s now a more highly edited show. And I’m kicking myself
for it because it takes a lot of time. It’s a lot of work and it’s freaking aiming to be a weekly
new show. So I’m pretty burnt out.
Okay. So how do you think about frequency of programming? Does it have to come out at the
same time every week?
It’s funny to leave. That’s another area where people have a lot of theories. And when I first
joined someone from the outside said to me, it should be exactly every Friday at 3 p.m.
There’s all kinds of theories around this. And that’s all wonderful. Here’s the three
things I learned. The best content will always win. Time of day, all the other stuff aside.
As you know, I am a master of timing things. This is how I made my section successful and
writing the specific timing for zeitgeist and virality. But for podcasts, there is no such
thing. So someone asked on Twitter about the tools for creators and distributors of podcasts
for people seeking to start their own podcasts, because not everyone is a big brand like this,
is actually sub-stack, which people think of as a newsletter only tool. It’s really about
connecting writers and people with their audiences. So people are seeking a place to both host and
distribute what better place than within the email ecosystem because you really own your
audience when you own that. Here’s the thing. People may talk about a podcast on social.
Like you’ll see a ton of people talking about a certain episode, but the reality that they’re
actually listening to it is very little. And so I think that the evolution of social podcasting,
which I know you were interested in as well, it’s still too early. So technically podcasts have
what I call, quote, “slow burn virality.” They don’t just go viral overnight. It takes about a week,
like the first wave of listeners is in that first week, and then you kind of see it grow from there.
Kind of like how I watch TV now. Oh my God, that’s such a good point. Well, you and I did a podcast
on a podcast about podcasting with Nick Quaw. People should listen to that episode if they want
to hear our thoughts on the trends because we did talk about binge watching and other things.
Back on the timing thing, I do believe editorially and especially for written content. And this thing
that I dubbed the McCluskey curve after Mark McCluskey, my former colleague, got wired.
He’s now at Sports Illustrated, I believe. But anyway, he always talks about how you add value.
And I say it’s when you’re offering something new or differentiated or leading early in the
cycle, or you do it in the end of a news and discussion cycle where it’s after it’s very noisy
and you have a very fresh or differentiated take and kind of being in the middle of all the noise
is like the worst position to be in in terms of value and for the timing of it. So that’s like
literally my philosophy. One thing I didn’t mention in the cycle, because someone asked about
the cycle from ideas to publishing that a big focus for us is around more promotion. And you
know, that’s sort of the whole cycle of things. And so one of the experiments I had wanted us to
do was to start doing audiograms and people promote podcasts through video. You know, the whole
definition of what is a podcast is blurring. Tom Webster of Edison Research was sharing how
for many young people, especially those that are streaming, whether on video or audio and Spotify,
they don’t know the difference between whether it’s on video or not. And so when people say
subscribe in your favorite app, it could be YouTube. What about sound effects music? Yeah.
So on the sound effects and music, we tried. So Hannah did a great episode for Halloween a few
years ago where she had sound effects for the person who was talking about the why behind the
weird. I would like us to have music on our show. As long as it doesn’t come off as like corporate
overproduced slick, because that’s not our aesthetic at all. However, I do think we need that and it
adds more dimensionality. I mean, it just sets the tone in the beginning. So for Ben’s new show,
it’s Ben Horowitz and Chaka Senghor. They co-host this show called Hustlin Tech, which is guides
to technology for hustlers. But what it really means, it’s really helping people use technology
to help themselves, which is an amazing concept. Basically, for that show, I did add music and
it was interesting because I didn’t want like stock music. And so I asked Chris Lyons, you know,
who runs our cultural leadership fund, but I didn’t even know he’d put his own sample in there.
He sent me like six tracks, including all these stock things that had a more hip hop sound to it.
And it turns out the one we all liked happened to be his personal one. So I was so excited,
but it sets a tone for the show. And I do want that. The reason we haven’t been able to do it
though is because licensing for songs is very complex. It’s not just copyright. It’s like layers
and layers of record distribution labels. So music is like one of the things that you’re
experimenting with. It almost sounds like you’re treating this like a startup or like a product.
Totally. I had a moment of emotional where I got a little teary-eyed because I was solo
for a long time. First, it was me and Michael. Then I was solo for a long time. Then I hired Hannah
and then it was me and her. And then about almost a year ago, I hired Amelia to be our managing editor.
And she’s growing the team so that we can scale this. And I had proposed that we hire an editor
for every vertical so we can really just go deep and kind of channelize our insights there and
truly self-select the audience. So anyway, to that point, I kind of got emotional when I saw
that all these desks that had been empty around me had people sitting in them. I got kind of like,
“Oh my God, is this how a startup CEO feels? If this is it, I’m in. Count me in.” Because I’ve never
done that. Okay. What are some new experiments you’re thinking about? So in general, in podcasting,
I’m fascinated by audio fiction. That’s like a really interesting and important trend. However,
I cannot for the life of me think about what our version of audio fiction would be. So maybe I’ll
just maybe do something personally. So who knows? At some point, I can experiment with that.
I’m still waiting for your book, by the way. I know. So am I.
Okay. So experiments in our podcast. Okay. So there’s a blending of, as you talk about, you wrote
about this in your knowable post, that you get this found time with audio. One of the things that
I’m interested in is that if the blinds and definitions of podcasting is blurring and you
really talk about this more than anyone, Connie. So I feel like I’m preaching to the choir here.
But the idea that audiobooks and podcasts and educational content all kind of blurs together,
for that same reason, why wouldn’t people listen to blog posts in podcast form or email
newsletters in podcast form? So I asked some of the other partners to read out loud their blog posts
on air. I’ll probably read a couple. And I do want to experiment more with us doing more
content in the audio form in that way. There’s already tools like autumn out there and media
outlets, but just more like a voiced way and not so manual. There’s like a new set of tools
coming about in that one. What are some experiments that didn’t work early? Oh, that’s a good question.
So very early on, Michael Copeland recorded a conversation with like four kids from a youth
and tech conference. And he told me about the footage and how complicated it was. And I was
like, why don’t you make it into like a narrative where you narrate it? And he did a beautiful
job of turning it into sort of an audio narrative story. So I’d like to do that. And it’s not that
it didn’t work. It’s just that we haven’t invested in it because we were so building one style.
So I’d like to do more of that. Experiments that didn’t work. So back to the question you asked
me about the ideal length. So I thought, well, what if I experimented with interstitials, where we
could segment an episode. Interstitials, like music in the middle. So not music. Kind of like
a intermission almost like a pause point that the listener would know. Hey, if you want to take a
break, this is a good spot. My thesis at the time was that people on campus, some kids on campus
at Stanford told me they were listening to it like while walking a class. And they don’t want to
listen to a long episode because the commitment was so big. So I was like, well, what if it’s a
long episode, but I segment it for them like the way you have chapter turns. That didn’t really
work because then I realized very fast from other people like, duh, I don’t need that. My app holds
my spot. So it doesn’t really matter. So it’s kind of orchestrated and contrived. So that didn’t
really work. Although that might come back because as with all experiments, sometimes it’s a matter
of timing. You probably do a hundred little experiments that you just don’t even think about.
Sometimes yes. But the reality is that I actually think you should have more focus. And I have a
very specific focus and a strategy for the existing Zee podcast and what I believe and where it’s
going. So I have a very particular vision for that. And when you have a vision for, of course,
the big podcast, but also each particular episode, are you editing until it hits that vision? Oh,
for an individual episode, how do you want to stop editing? Oh my God. This is the hardest question,
believe it or not, because right now we’re onboarding some new editors. Yeah. And how do you
teach other people? Well, this is what I’m struggling with. And frankly, I read Ben’s book
recently and it’s beautiful, by the way. And how do I think about doing this culturally and thinking
about it? This is the exact challenge because well, I learned podcast by myself, like everyone can
do it. And I’m realizing that the way we do things is quite things that I take for granted
as implicit or not that explicit, or they’re very tacit things or mindsets that are really
unique and foreign. So to answer that question, the answer you can’t use when you’re scaling
is, well, I’ll know it when I know it. Like, you know, the line from Justice Potter Stewart
about porn, he said, I’ll know it when I see it. I mean, that’s how I think about investing sometimes.
Really? So it’s like instinct. So this goes back to that whole like view of like instinct.
But the reality is that instinct is trained by experience. But what I find is taste is very
difficult to train. And you can’t just say to someone, well, you’ll know it when you hear it.
That’s not a good enough answer. So it’s frankly not helpful. It’s not helpful. And it’s hard. I
don’t have a full answer yet. But I will say that you can figure out the bar by having some
principles for what you’re doing. So what are your editorial principles?
They’re things like I mentioned non derivative experts, true to the maker, the culture of
adding a very fresh and differentiated take. We don’t want to say what everyone else is saying,
the art of timing. Does this meet the bar? Is this really adding value to the conversation?
Is it signal versus noise? Is it more of the same? Is it spinning forward? I use that phrase all the
time, spin it forward, spin it forward, spin it forward. How do we do it? At the same time,
how do we make it concrete? Because our audience is not just people like big Fortune 500 companies
thinking about the future of tech or startups. I mean, I hear people talk about the podcast who
don’t work in tech. That’s to me is a bar of success right there. So actually,
I describe it as the podcast is influencing the influencers. And so to me, when media outlets,
reporters say like, oh, I listened to that episode, I love that because they may not write it up,
but it informs their thinking. One of my big principles is that we need to either provide
a framework for how to think about something, if not an answer, or tease apart height versus reality
and think about like how big tech trends like VRAR or whatever the topic is,
it may play out concretely, then it informs the influencers. Give me a concrete example,
like what’s the proudest moment of that? So one of my most favorite moments and stories about
the podcast is that a Senator, a US Senator was listening to the episode and this is a testament
to the network and they come across our content and the whole thing kind of reinforces like a
flywheel. And he heard the episode, it was about health data, and he literally had his staffer
reach out to us and the staffer quoted his line to us, like what he said. And he’s like, I can’t
believe this idea is not already being done already. I want to propose it in the upcoming
session as legislation. Can you please put me in touch with that founder? So you’re affecting
policy. I literally call it policy by podcast. Now, I don’t actually think that came about,
but that is one of my all time favorite stories. Wow, that’s awesome. And by the way,
in the early days, in terms of thinking about the audience, there was an incredibly strong brand
that Mark and Ben and Margaret built as a base and a foundation for sure. And the network is
a thing that continually reinforces it. But initially, I had to beg my contacts, people I
edited people in the book publishing industry to get them on the podcast because A, they didn’t
really know A, six and Z. B, it was a fledgling nascent podcast. It hadn’t had like an established
presence. And so I convinced one or two of the key publicists and book publishers because I ran
their excerpts in my section, I then got the, once I got one big name author in, then the rest
started following. And then I started getting pitched because you must get pitched books all the
time. Not only pitch books, we get at least five to 10 emails a week that are just pitching us.
Then of course, people, book authors going on podcasts became a thing because it actually
moves book sales. By the way, one author told me that he came on our show and we moved a thousand
books in like a few days because it’s a very self selected audience that’s, you know, listening
and very motivated and there’s no better way. But we try to break the script for book podcasts.
So one of my rules is again, going back to editorial principles of differentiation is
if you have someone like Yuval Harari who’s been doing the circuit and he’s like on every major
podcast show and he’s a really well read author and the person who put him on is my friend,
Rimjim Day. She’s one of the people who took a very early bet on us.
He has to talk about something different with you.
Well, it’s that he has to talk about his book, but we have to do it in different way.
So I want to go back to a topic you mentioned early on, which is how in that first year,
you didn’t like hearing your own voice. That’s why you don’t want to get on the podcast.
But now I have been with you in public where people run up to you and say they recognize
your voice. So how does that feel being a voice celebrity? I’ve seen people want to take photos
with you, but what does that feel like? It goes back to feeling that sort of vulnerability of being
a person who wants to be behind the scenes. I don’t know if you know this, but my first two years,
I didn’t even put my name on the byline of the podcast. In fact, people found me proactively,
which is crazy to me because I thought the goal of the host, because this is what editors do.
Editors do tremendous work. The shaping editors do tremendous work to shape a piece. They practically
co-write them. But I did not include myself on the byline because I thought it was my job to be
invisible as the host and the moderator. And I always view myself as a shepherd for the audience.
That is my job. Although it’s funny because over the years, then people started finding me. I
eventually added my name on the byline. It’s incredible when people come up to you because,
frankly, when you’re sitting in a room and this is what I love about podcasting, is that that’s
intimacy. People think they know you. I love that. They feel like they know me because I’m in their
ear, but there’s a huge asymmetry there. But anyway, it’s amazing and powerful and moving to see
your work in action. And I’m so grateful to our fans and to Andreessen Horowitz for letting podcasts
go off the ground. Dixon and Kim and Michael started it, but after a while, I don’t think
really people paid attention to it. I think Mark told me about a year and a half in.
I think it surprised us all. I think it did. People didn’t really think it’d be so big.
But like I said, we had to earn those listeners because it’s not like you have brand and they
come. It’s that you have the guest and then they listen and it gets better and this is where the
editing comes in. Yeah. I feel like I have seen you edit and work magic. Even our WeChat piece
that we did years ago, you made that into a completely different thing. I’ve seen you edit
ever since and it’s really funny because once you were in the zone, like one time I was watching you
on Google Docs edit and I really felt like I was watching a painter paint. I just saw these like
sentences moving around and it was like watching a paintbrush. That’s so beautiful. I love that
you’re saying that. Don’t make me cry on the podcast. Were there podcasts that surprised you,
things that made you cry? The podcast that made me cry, there’s actually been two or three.
So one was with Leila Gena of Sama Source. They were empowering people around the world
with micro work. They found a way for a woman who previously had no spending money to be able
to for the first time in her life by makeup. And that sounds so frivolous, but that made me,
it completely, I started crying. I edited it out, but that was one of the episodes that made me cry.
Another one that made me cry recently was Ben and Shaka when they interviewed Deshawn and Cherie
about Maven and there was a moment that just brought me to tears and you should listen to that
episode, but that was also another one of the episodes that made me cry. So I want to talk about
your policy on cursing on the podcast. Do you do it? Do you bleep it out? It’s funny because I felt
a lot of tension about it. I remember I once asked Ben and Margaret about it because I was like,
you guys think it’s bad that I cuss? Should I stop? And Ben was like, entrepreneurship is hard,
it’s a struggle. It’s meant to be hard. And then I started getting folks on Twitter, some being very
helpful and some being judgmental. You know, we don’t think you should cuss. You’re such a beautiful
voice and you sound so nice. Is it necessary? That would probably drive you to swear even more.
It did. And frankly, the reason I wanted to keep it is because I believe when women,
we’re asked to conform to so many things, no uptalk, no this, no that. There’s so many different
things. And I’m just like, you know what? I want to be me, raw and real. But here’s why I did finally
decide very recently to stop cussing on those shows. And no, I don’t bleep it out, but funnily
enough on the podcast you and I did with Nick Kwaab, a podcast about podcasting, someone on Twitter
totally teased me. They’re like, I think it’s hilarious that you bleeped out the name of a
company to protect their confidentiality, but you didn’t bleep out your F-bomb. So now I edit it out.
So I stop because they’re kids in the car. And in the beginning I was like, well, don’t listen.
But now, and we’ve talked about ear shares and new mind share, people need to listen to podcasts
with kids in the car. And by the way, the other trend I think is super interesting about podcasting
is these new wave of shows just for kids. It’s one of my favorite things. I feel like in podcasts
you’re taking something that’s previously just information and education and you’re forced to
make it entertaining. Ah, you’re right. Well, I think the job of the moderator for me is to be a
shepherd for the audience. And that means including stitching together statements, helping the listeners
follow along with the arc, summarizing and explaining. But to the entertainment part, I agree with you.
I do believe the future of podcasting is merging with entertainment. And that is going to be
interesting to see. Tell me a little bit about what software, what technology do you use?
Like the hardware, not just the software. Yeah. Yeah. So we use a zoom recorder and we use sure
mics and we use like standard. At one point we use hall mics for our clamp. Do you use a mixer?
Oh, no. So from the early days of the podcast, people would keep complaining on Twitter like,
“I listened to your podcast. It doesn’t sound good.” I thought musicians always use mixers.
Well, one of the negative legacies of podcasting tools is that a lot of them were grounded in the
music world versus made natively for podcasting, which is why you like Descript and other tools
came about. Basically, we got rid of the mixer and then our sound improved drastically because we
can manipulate more because you record separate tracks. How did you know to get rid of the mixer?
So this is my biggest, most invisible partner in crime who I want to give a shout out to is
our sound engineer, Seven Morris. I brought him in and I was like, “Please fix our sound.”
People keep complaining, “I cannot take no for an answer. You need to tell me what’s wrong
and I want it fixed.” I’m just like, “There surely is a solution.” It was the funniest thing
because what he did was he basically removed the mixer. He’s like, “You guys don’t need this.
That’s what people use for live events. You’re editing. You want individual tracks. You should
be plugging directly into the recorder, not having a mixer in between.” So that’s what we did. Our
equipment is not that expensive. I give a lot of startups advice on how to do this stuff and
the list of equipment is all under $1,000. We have sound panels and acoustic stuff, but not really.
We use really standard equipment. I think the primary thing in the tech stack is that we’re
now on Simplecast, which is our hosting platform. I think of them as the stripe of podcasting because
they have an API model. We’re getting a lot of features in that and particularly going to be
more important as we expand to more and more shows. The other thing is that, as everyone knows,
analytics for podcasts have been very, very broken and very difficult because the industry has not
standardized. What’s great is that Simplecast has been going through the process of IAB certification.
For me, the wishlist has always been completion. Where do people drop off? I want chartpeat-like
analytics for things. I want to know about audience overlap between shows. I want to know where people
are more engaged. Are there parts where they’re repeating and trying to listen again? There’s
so many million things I want to know. What’s also great is Amelia hired. I actually credit to Andrew
on this because he had suggested as part of the hiring plan. He suggested we hire a growth and
audience development person in addition to the editors. I’m really glad he did because Amelia
hired a wonderful growth and audience development person for us, Jared. He is very much thinking
about how to bring the promotion side. As we’re growing and thinking about, “Okay, we have our
main show. Now we have the 16 minutes thing. We have hustling tech. What is the future of our
podcast library look like?” I think, as a market has a great phrase, she describes our podcast and
the editorial as a platform, which I think is exactly right for the podcast. I think of us
as expanding into more of a network. Network meaning multiple shows. I want to try different
types of shows, but you also don’t want to abandon your audiences. I don’t want to arbitrarily start
a show. Then if we only have a few episodes, not keep that feed, which is why some of these
shows are starting off as series. Then we can break it out into its own show and its own feed.
Capsule collections in the beginning and then they break out. For some, we start from the get-go.
So 16 minutes, that I knew from the beginning would be its own show. But what we did was
we let it run on the main feed for the first 10 episodes. Someone on Twitter actually asked
that. I thought it was the cutest question. They said, “Why is it 16 minutes? It’s not on the main
feed anymore. This is the reason why.” Then we ripped off the Band-Aid and told people it’s no
longer on this feed. So now people only subscribe there. You see a big spike when you gave those
call-outs. That’s because I need to build a new show there. Let me just tell you how painful that
is because after building a show at this 500th episode and now starting at 15, it’s a very different
game. It’s like exercising new muscles all over again, but I love it because I miss that zero to
a second child. I guess maybe it’s very similar because you’re going back to scratch again in
some ways. That’s how you build that type of a network. Also, we will be verticalizing some of
the channels so that people can subscribe to feeds. We’re going to have a separate channel
initially for A6 and ZBio. That’s great because an audience can self-select. If people want to
talk about journal articles without people who don’t want to hear about crypto policy,
you don’t have to mix those. This is great because this is, to me, the future of media.
I’m a big believer in Kevin Kelly’s 1,000 true fans and then going from there. Initially,
when I started thinking of mapping out the territory of podcasts, I literally thought
about it as mapping sales territory. I want to conquer the open-source community. I want to
conquer Node.js. Let me bring someone on from there. I literally mapped it out geographically in
community-wise to grow and aggregate that. This is how to take over the world.
Software is in the world, but so is audio. I am so honored to be able to interview you today.
Happy 500th episode. Thank you. Thank you to everyone for listening. I also thank you to
our incredible, amazing team here. I want to thank especially our audio engineers,
Sevin Morris and Tommy Herron. The first editor I hired, Hannah Tidnam, who’s now been here for
three years. Now, Emilia and the rest of our team who’ve joined Das, Lauren, and Zoran,
who are starting to podcast. Thank you to Margaret especially. Thank you to Kim, who reached out and
hired me and also was funny. She actually told me she had never thought I would be into podcasting
and I was like, “Me neither,” but I’m so grateful to the firm. Frankly, it’s a miracle that they
would be so supportive of us doing this. I’m so thankful for that. Thank you, everyone,
and thank you, Sonal, for the A16C podcast.

“Hi everyone, welcome to the a16z Podcast…” … and welcome to our 500th episode, where, for the first time, we reveal behind-the-scenes details and the backstory of how we built this show, and the broader editorial operation. [You can also listen to episode 499, with head of marketing Margit Wennmachers, on building the a16z brand, here.]

We’ve talked a lot about the podcasting industry, and even done podcasts about podcasting, so for this special episode, editor-in-chief and showrunner Sonal Chokshi reveals the how, what, and why in conversation with a16z general partner (and guest-host for this special episode) podcasting fan Connie Chan. We also answer some frequently asked questions that we often get (and recently got via Twitter), such as:

  • how we program podcasts
  • what’s the process, from ideas to publishing
  • do we edit them and how!
  • do guests prep, do we have a script
  • technical stack

…and much more. In fact, much of the conversation goes beyond the a16z Podcast and towards Sonal’s broader principles of ‘editorial content marketing’, which hopefully helps those thinking about their own content operations and podcasts, too. Including where podcasting may be going.

Finally, we share some unexpected moments, and lessons learned along the way; our positions on “tics”, swear-words, and talking too fast; failed experiments, and new directions. But most importantly, we share some of the people behind the scenes who help make the a16z Podcast what it was, is, and can be… with thanks most of all to *you*, our wonderful fans!

Leave a Comment

AI Engine Chatbot
AI Avatar
Hi! How can I help?