Ben Horowitz and Ali Ghodsi: How to Run a Billion-Dollar Business

0:00:02 I was like, maybe they’re right, maybe we should just sell.
0:00:04 And I remember having that conversation with Ben,
0:00:06 which is he said, hey, you can do whatever you want.
0:00:08 You can sell, you’re going to make a lot of money,
0:00:10 and you’ll be super successful in life.
0:00:12 But, you know, if you’re like me,
0:00:13 you’re going to look back the rest of your life
0:00:16 thinking, you know, I missed that one shot.
0:00:17 That was the one thing.
0:00:18 I should have taken it all the way,
0:00:20 and now I’ll never know how far I could have taken it.
0:00:20 Could have been.
0:00:21 So do you want to live with that,
0:00:23 or do you want to just have the money?
0:00:24 You know, I’ll support whatever you want to do.
0:00:26 I really couldn’t care less.
0:00:29 In 2016, when Databricks was on the brink,
0:00:31 the board wanted a new CEO.
0:00:34 Co-founder Ali Gazi was ready to go back to academia.
0:00:37 Instead of leaving, he took the job
0:00:38 and turned an open-source project
0:00:40 into a $100 billion company.
0:00:42 On this episode of Boss Talk,
0:00:44 I’m joined by Databricks CEO Ali Gazi
0:00:47 alongside Ben Horowitz and Sarah Wang.
0:00:49 We unpacked the 2016 crisis,
0:00:51 the Microsoft deal that changed everything,
0:00:54 and how Databricks built intensity without burnout.
0:00:57 We also talk about giving feedback that actually lands
0:00:59 and why not selling might be the boldest call
0:01:00 a founder can make.
0:01:01 Let’s get into it.
0:01:05 Excited to bring back Boss Talk.
0:01:07 This was a series that you guys did
0:01:08 a few years ago on Clubhouse.
0:01:09 That was a big hit.
0:01:10 Yeah, we had fun.
0:01:11 It was Ben’s idea.
0:01:13 Yeah, excited to bring it back.
0:01:15 So in the spirit of Boss Talk,
0:01:16 let’s talk about the first time
0:01:17 that you became a boss
0:01:19 in terms of running Databricks.
0:01:21 Let’s talk about the moment in 2016
0:01:23 when things weren’t as smooth
0:01:24 as perhaps they should have been
0:01:25 and we were looking for a new CEO.
0:01:27 And Ben, you recommended Ali.
0:01:29 What did you see in Ali?
0:01:30 Well, that’s not really what happened.
0:01:32 Why don’t you tell the disclaimer his story?
0:01:32 Uh-oh.
0:01:36 So what happened was Scott Schenker called me,
0:01:40 who was kind of the professor in the background,
0:01:42 kind of founder type character,
0:01:43 very smart guy,
0:01:44 very nice person.
0:01:47 and he said the guys,
0:01:49 the other co-founders,
0:01:54 really think Jan probably shouldn’t be CEO anymore
0:01:57 and then Ali should be CEO.
0:01:58 And I was like, okay.
0:02:01 And I didn’t know Ali that well at that time
0:02:01 and I said,
0:02:05 you want to swap out one professor
0:02:06 for another professor
0:02:09 because the one professor you have,
0:02:12 you feel isn’t running it well enough.
0:02:13 One professor wants to swap out
0:02:15 one of his professors for another professor.
0:02:17 Yeah, it’s like a professor who?
0:02:19 And I’m like, okay,
0:02:21 are you sure you want to look at anybody
0:02:22 from the outside?
0:02:23 Oh, no, no, no, no.
0:02:25 All the founders feel like it should be Ali.
0:02:27 And I think Martin Mikos
0:02:28 was running around at that time
0:02:30 and I was like,
0:02:31 well, maybe you should talk to him
0:02:33 and see what an outside guy looks like.
0:02:35 And I said, look, I’ll talk to Ali.
0:02:38 There was another factor that was going on,
0:02:40 which was the company was in real trouble.
0:02:41 Like it was not going well.
0:02:45 And so the idea of bringing in an outside CEO,
0:02:48 we wouldn’t have been able to recruit the caliber.
0:02:50 Like now you could recruit anybody
0:02:51 to be CEO at Dataverse.
0:02:53 But then the choices weren’t that good.
0:02:55 So I sat down with Ali
0:02:58 and I would say the one thing I remember
0:02:59 about the meeting was
0:03:04 I was surprised about how clear he was.
0:03:06 he wasn’t sure exactly what to do about everything,
0:03:08 but he knew all the problems
0:03:10 and then he kind of knew who he was.
0:03:11 And so I thought,
0:03:15 well, we should give this a shot.
0:03:16 So then I took that back to the board.
0:03:17 And of course,
0:03:18 the other people on the board
0:03:21 who were outsiders were very skeptical.
0:03:23 And I said, we’ll give him a year deal
0:03:24 and we’ll see how it goes.
0:03:26 First of all, like kudos to Jan
0:03:27 for building the company originally
0:03:29 and Ben investing and believing in us.
0:03:30 And then also,
0:03:32 I kind of couldn’t have done the CEO job.
0:03:35 Ben basically babysat me
0:03:36 the first couple of years.
0:03:38 It’s a short babysitting job.
0:03:39 So I did know what was kind of wrong
0:03:40 with the company
0:03:41 because I had been there for two, three years
0:03:43 and I had seen from inside
0:03:44 what we should change
0:03:45 and what the issues were.
0:03:47 But we had an open source project
0:03:49 that actually became very successful
0:03:50 thanks to those first two, three years.
0:03:53 Apache Spark became a worldwide sensation
0:03:55 and we could pride ourselves
0:03:57 on the number of downloads of the software.
0:03:58 Well, and the Spark conference.
0:03:59 Yeah, yeah.
0:04:00 Spark conference.
0:04:01 Now the data and AI conference.
0:04:01 Yeah.
0:04:03 We would say we have 5 million recurring revenue
0:04:05 because the ticket fees for Spark Summit
0:04:06 was like 5 million.
0:04:07 And then they would say,
0:04:08 well, how is that recurring?
0:04:08 And we would say,
0:04:09 well, the conference is every year.
0:04:11 So it’s recurring revenue.
0:04:12 But the problem was that
0:04:13 as it is often with open sources
0:04:14 that everyone’s just downloading
0:04:15 the open source version.
0:04:16 Actually, your biggest enemy
0:04:18 is your open source project.
0:04:20 The main thing you have to fight
0:04:21 in the market is,
0:04:22 hey, why can’t I just download
0:04:23 the open source version?
0:04:24 Amazon is offering it.
0:04:25 The cloud vendors are just offering it.
0:04:26 I’m just going to use that.
0:04:28 So this was the biggest challenge
0:04:29 that Databricks had at the time.
0:04:30 And we needed to do
0:04:32 very serious, aggressive pivots internally,
0:04:34 which were going to be very, very painful
0:04:34 to lots of people,
0:04:36 like to the whole ethos
0:04:37 of the company internally.
0:04:39 And so I kind of knew that
0:04:40 for almost a year.
0:04:41 So when I got the shot,
0:04:43 that’s what we started doing.
0:04:45 The strategy was like
0:04:47 make Spark the biggest open source thing.
0:04:48 I mean, I can remember it
0:04:49 on all the slides now
0:04:50 and then Databricks
0:04:51 would have the best spark.
0:04:53 But Databricks never had
0:04:54 necessarily,
0:04:55 we didn’t do a lot
0:04:56 to make it the best spark
0:04:57 or not differentiated enough.
0:04:59 And that was kind of
0:04:59 the first thing
0:05:00 Ali did on the product side.
0:05:01 And then he hired
0:05:02 Ron Gabrisco,
0:05:05 which that was transformational
0:05:07 because that kind of dragged
0:05:09 the company into the world.
0:05:10 Yeah.
0:05:12 So obviously that was
0:05:12 the right decision.
0:05:13 It paid off.
0:05:14 Zooming out,
0:05:15 Ben, you’ve worked with
0:05:17 and you know all the great CEOs
0:05:18 of our own and worked with them.
0:05:19 Where does Ali spike
0:05:20 and where are his superpowers
0:05:21 as a CEO,
0:05:22 as a boss
0:05:23 that have helped
0:05:24 contribute to the impact?
0:05:26 I mean, Ali’s really good.
0:05:27 So I always rate CEOs.
0:05:29 Okay, if I was running
0:05:29 that company,
0:05:30 would I do a better job
0:05:31 or a worse job?
0:05:32 And with Databricks,
0:05:34 I do a way fucking worse job.
0:05:36 So he’s good
0:05:37 on many, many dimensions.
0:05:38 So I’d say first of all,
0:05:40 he is a real technologist,
0:05:43 like not a pseudo technologist
0:05:44 like his competitors.
0:05:46 I’m sorry.
0:05:48 So he really knows the product.
0:05:50 He understands the product strategy
0:05:50 in detail.
0:05:52 He also ran engineering
0:05:53 before he was CEO.
0:05:54 You know, mostly what I worked
0:05:55 with him on the early days
0:05:56 was just, okay,
0:05:57 go to market in BD.
0:05:58 And he’s really good
0:05:59 at both of those.
0:06:01 That’s where we had to catch up.
0:06:02 You know, stuff like
0:06:03 had an amazing go to market.
0:06:05 And then we needed a big deal
0:06:07 with kind of big partners.
0:06:10 I mean, I got him
0:06:11 like a little BD tutor,
0:06:12 John O’Farrell,
0:06:13 who did a nice job,
0:06:15 came in and kind of taught Ollie
0:06:17 about how you structure a deal,
0:06:17 how you do things.
0:06:19 But he learned everything so fast.
0:06:21 And then probably the thing
0:06:23 that he does that I wish
0:06:25 I could get all our CEOs to do
0:06:27 is he doesn’t hesitate.
0:06:28 He trusts his eye.
0:06:29 Like he’ll see something
0:06:30 and he doesn’t know if it’s right.
0:06:32 And so if you look
0:06:34 at the strategy changes
0:06:35 Databricks has had,
0:06:37 one big one was building
0:06:38 a data warehouse.
0:06:40 Like that is a pretty big swing
0:06:42 and a seemingly like
0:06:44 quixotic and sane idea
0:06:45 given where they were.
0:06:47 He’s paranoid enough
0:06:47 that he knew
0:06:48 that could be an issue
0:06:51 and then he trusted himself
0:06:54 enough to go get deep enough
0:06:55 to decide whether to do it or not
0:06:56 as opposed to, you know,
0:06:57 ignore it.
0:06:59 These guys are trying to kill me.
0:07:00 I don’t want to see it,
0:07:02 which is what a lot of CEOs do.
0:07:03 And so that kind of thing.
0:07:04 But there’s a lot of elements
0:07:05 to that job.
0:07:06 It’s a very complicated job.
0:07:09 Ali, talk more about the journey
0:07:10 about evolving from
0:07:11 an academic,
0:07:12 a technologist
0:07:14 to someone commercial.
0:07:15 It’s a journey
0:07:16 our CEOs go through.
0:07:17 Talk about what it was like for you
0:07:18 and in the context
0:07:19 of what others can learn from it.
0:07:21 So we were in academia,
0:07:22 so we were scientists.
0:07:23 And then, you know,
0:07:24 I led engineering and product.
0:07:24 So you got to learn
0:07:26 how to build a product
0:07:27 and get product market fit.
0:07:29 And then I became CEO afterwards.
0:07:29 Each of these
0:07:31 has different sort of challenges.
0:07:33 I think that in all of them,
0:07:34 the thing that is in common
0:07:35 is that you really have to understand
0:07:37 and be extremely good
0:07:39 at the task at hand.
0:07:40 And so number one,
0:07:42 admit that you don’t actually know
0:07:42 everything about the job.
0:07:43 So A,
0:07:45 first step of anonymous alcoholics.
0:07:48 Admit you have a problem.
0:07:48 Okay.
0:07:50 And then number two,
0:07:51 be a student
0:07:52 and learn everything you can about it.
0:07:53 Right?
0:07:53 Go all the way down
0:07:54 to the details
0:07:56 and try to learn from the best.
0:07:57 And then work your butt off.
0:07:58 You’re nothing.
0:07:59 You’re zero, right?
0:07:59 You know nothing
0:08:01 about writing reliable software.
0:08:02 I would say that was the same thing.
0:08:03 I tried to learn.
0:08:04 I tried to network
0:08:05 with the best heads of engineering,
0:08:06 best heads of products.
0:08:07 I tried to read every book.
0:08:09 I got as much as I could
0:08:10 out of Ben and Mark.
0:08:11 I read all of their blogs,
0:08:11 all of their books,
0:08:12 everybody else’s.
0:08:14 But then you do search.
0:08:15 Search firms are really great
0:08:16 at getting you to meet
0:08:17 who’s the number one product manager
0:08:18 by reputation
0:08:19 in the market right now.
0:08:20 Can you get 30 minutes
0:08:21 with that person?
0:08:21 Just sit down.
0:08:22 They’re not going to join you
0:08:23 because your company
0:08:24 is too crappy and too small.
0:08:25 Can you get 30 minutes with them?
0:08:26 Can you get a dinner with them?
0:08:27 Can you get a breakfast with them?
0:08:28 And then just ask them
0:08:29 lots and lots and lots of dumb questions
0:08:30 and they’ll tell you.
0:08:31 They’ll happily just tell you like,
0:08:32 hey, here’s how I do it.
0:08:33 The other guys are wrong.
0:08:34 And then they’ll give you a playbook
0:08:35 and you can go compare it.
0:08:36 You can go to the next person
0:08:36 and say, hey, this is the playbook
0:08:37 I heard from them.
0:08:37 It’s like, no, no, no,
0:08:38 that’s totally wrong.
0:08:39 You don’t do it that way.
0:08:40 Here’s how I do it.
0:08:41 And then very soon you learn enough.
0:08:43 And if you really have grit
0:08:44 and you work hard,
0:08:45 you’re going to be able
0:08:46 to do great things.
0:08:48 That’s about you yourself.
0:08:51 But also if you hire a great team,
0:08:51 because as a leader,
0:08:53 you alone can’t do much.
0:08:54 So can you hire
0:08:55 the best people out there?
0:08:56 So that’s also part of that.
0:08:58 Do you know what great looks like?
0:08:58 Have you interviewed
0:08:59 all the best people?
0:09:00 And then can you now sell them
0:09:01 and get the best people
0:09:02 to come work for you?
0:09:05 of excellent people,
0:09:07 then they will uplift you.
0:09:08 This is a managerial leverage
0:09:09 that I learned from Ben,
0:09:10 which is from
0:09:11 High Output Management
0:09:12 by Andy Grove.
0:09:13 But it’s basically,
0:09:14 are they so great
0:09:16 that you’re learning from them?
0:09:17 It’s like,
0:09:18 Ali was a great head of engineering,
0:09:19 but actually he was
0:09:20 a great head of engineering
0:09:21 because the people
0:09:21 that worked underneath me
0:09:23 were doing amazing things
0:09:24 and I was just standing
0:09:25 on their shoulders.
0:09:26 So you just have to do that
0:09:27 and you have to instill
0:09:28 everybody else to do that
0:09:29 recursively so that you end up
0:09:31 with a just amazing killer team
0:09:32 and you got to continue doing that.
0:09:33 Now you have to, okay,
0:09:34 for engineering,
0:09:35 it wasn’t actually that hard
0:09:36 because I had written
0:09:37 a lot of software,
0:09:38 but now you’re CEO.
0:09:39 So now you have to do that
0:09:40 for marketing.
0:09:41 You have to do that for sales
0:09:42 where they were super helpful
0:09:43 because they had done it
0:09:44 with Loud Cloud and Opsware.
0:09:45 So they knew how to build
0:09:45 a B2B machine,
0:09:47 how the game was played.
0:09:48 But you have to do it again,
0:09:49 but now you’re doing it
0:09:50 in a field where you’re
0:09:51 really clueless.
0:09:51 And also,
0:09:52 probably all of your instincts
0:09:53 are wrong
0:09:54 and your intuition
0:09:55 is completely wrong.
0:09:57 So can you be clairvoyant
0:09:57 and see the truth
0:09:59 or do you want to lie to yourself?
0:10:00 And that’s where I think
0:10:01 a lot of founders
0:10:01 make mistakes.
0:10:02 like they’ll do well
0:10:04 in their own archetype.
0:10:05 When they have to step outside
0:10:06 of their own archetype,
0:10:07 they make a mistake.
0:10:08 They hire people
0:10:09 that are like their own archetype
0:10:10 in other roles
0:10:11 where that could be lethal.
0:10:13 By the way,
0:10:15 that’s how we started Databricks
0:10:17 was I think everybody
0:10:19 was a PhD in computer science
0:10:20 who was running anything,
0:10:21 including sales.
0:10:24 That’s probably
0:10:25 the number one mistake
0:10:27 is you just go,
0:10:28 okay, well,
0:10:29 like I’m an engineer,
0:10:30 so I want a sales guy
0:10:32 who can talk to me
0:10:33 and understands engineering.
0:10:34 Well, that’s not a really
0:10:36 good criteria for sales.
0:10:37 And I remember
0:10:39 a really early conversation
0:10:40 Ali and I had
0:10:41 was Ali called me up
0:10:42 and he says,
0:10:43 I’m a little worried
0:10:44 about the sales organization.
0:10:45 And I said,
0:10:46 well, what’s going on with them?
0:10:47 He said,
0:10:47 well, they’re just
0:10:48 inventing products.
0:10:50 And I was like,
0:10:51 well, what do you mean?
0:10:52 Well, they’re selling
0:10:53 stuff we don’t have.
0:10:54 And I was like,
0:10:54 what did they sell
0:10:55 that you don’t have?
0:10:55 He said, well,
0:10:56 they sold this
0:10:58 training professional services
0:10:59 package for $200,000.
0:11:00 And I said,
0:11:01 well, why’d they do that?
0:11:03 And he said,
0:11:04 well, the sales guy said
0:11:06 the customer had $450,000
0:11:07 and the product
0:11:08 only cost $250,000,
0:11:09 so he needed to get
0:11:10 the rest of the money.
0:11:12 And I said,
0:11:14 that’s a really good sales guy.
0:11:15 I think you told me
0:11:16 that’s called price to budget.
0:11:21 So those kinds of things,
0:11:22 he just learned faster
0:11:24 than most other CEOs
0:11:26 who are in that position.
0:11:27 And then he’s taken
0:11:28 a lot of stuff
0:11:29 that I know how to do
0:11:30 and he’s done it much better.
0:11:32 So one thing
0:11:33 that I’m good at
0:11:34 is rather than
0:11:35 telling somebody
0:11:37 that’s stupid
0:11:38 and hurting their feelings
0:11:38 or so forth,
0:11:39 I’ll ask them
0:11:40 a really fucked up question.
0:11:42 Actually, I did it
0:11:42 in a board meeting.
0:11:43 I said, well,
0:11:44 could you help me
0:11:45 with the math on this?
0:11:45 Because I don’t understand
0:11:46 the math.
0:11:47 Actually, it was worse.
0:11:48 He said, help me with the,
0:11:49 I’m just trying to understand
0:11:50 basic math.
0:11:51 You have all these numbers
0:11:52 on the slide.
0:11:53 And if you said that
0:11:54 your conversion ratio
0:11:55 is 5%,
0:11:56 but I can’t divide
0:11:57 any of those two numbers
0:11:58 to get 5%.
0:11:59 And then the person
0:11:59 freaked out.
0:12:01 Don’t freak out.
0:12:02 Just tell me
0:12:02 which of the two numbers
0:12:03 do I divide
0:12:04 to get 5%?
0:12:05 Because I’ve divided
0:12:05 all of them
0:12:06 and none of them
0:12:07 is fired.
0:12:08 Am I going to be fired?
0:12:10 So he does a much better
0:12:11 version of that
0:12:13 which is if somebody’s
0:12:14 really, really screwed
0:12:15 something up
0:12:15 or messing,
0:12:16 he’ll go,
0:12:17 how do you think
0:12:18 it’s going?
0:12:19 And I was like,
0:12:21 since he told me that
0:12:21 I was like,
0:12:21 oh yeah,
0:12:22 that’s a better way
0:12:22 to do it.
0:12:23 That’s even better.
0:12:25 So yeah,
0:12:27 he’s a very good student.
0:12:27 Can I refund that?
0:12:28 You know,
0:12:28 there’s this book
0:12:29 called Radical Candor
0:12:30 and I think people
0:12:30 take it too far
0:12:31 and they misunderstand it
0:12:31 and so on.
0:12:32 But I think the essence
0:12:33 of that book
0:12:34 is that if feedback,
0:12:35 if,
0:12:36 are you criticizing me?
0:12:37 Are you saying I’m stupid?
0:12:38 I can’t do the vision?
0:12:39 Because my point is not
0:12:40 about the 5%.
0:12:40 I was trying to make
0:12:41 a different point
0:12:42 and now you’re just,
0:12:43 this is a cheap shot
0:12:44 and now I’m like hurt
0:12:44 and I think,
0:12:44 by the way,
0:12:45 I think you’re wrong.
0:12:45 It’s not five.
0:12:46 I said six and a half.
0:12:48 So are you criticizing me
0:12:48 or is it like,
0:12:49 no, no, no,
0:12:50 I’m here to help you.
0:12:51 I can not help you
0:12:53 but if you beg me for help,
0:12:53 maybe I’ll help you.
0:12:55 So which of the two modes?
0:12:56 So if you can get people
0:12:57 into the mode of,
0:12:57 oh wow,
0:12:58 I’m being helped.
0:12:59 They’re helping me
0:13:00 and I’m going to get further ahead
0:13:01 in my career
0:13:02 and I’ll be more successful.
0:13:03 Please,
0:13:03 no, no, no,
0:13:04 please don’t leave.
0:13:05 Come back and tell me more
0:13:06 because I’m taking notes here.
0:13:07 So if you can flip to that
0:13:09 and I think a lot of feedback
0:13:10 can be recast
0:13:11 into,
0:13:12 I’m just here to help you
0:13:13 but feel free to completely
0:13:14 ignore this advice
0:13:15 but if you want to be
0:13:15 really successful
0:13:16 and if you want to get that job
0:13:17 or if you want to get that project
0:13:18 next time,
0:13:19 if you did it this way
0:13:20 then you probably would have
0:13:21 had a higher probability
0:13:21 of getting that
0:13:22 but I don’t care.
0:13:23 You do whatever you want
0:13:24 and then people are much
0:13:25 more receptive.
0:13:25 They’re like,
0:13:26 no, no, no, please,
0:13:26 I want to know more.
0:13:27 Yeah, yeah, yeah.
0:13:28 Well, and then
0:13:29 just the frequency
0:13:31 that I think helps a lot too
0:13:33 where if I see you
0:13:34 once a year
0:13:35 at your review
0:13:36 and I tell you
0:13:37 what’s wrong with you,
0:13:38 you’re going to be offended
0:13:38 no matter what it is.
0:13:40 No matter how wrong it is,
0:13:41 no matter how correct I am,
0:13:43 it’s going to be offensive.
0:13:45 But if every day
0:13:46 if I see you doing
0:13:47 something I don’t like,
0:13:47 I go,
0:13:48 no, don’t do it that way,
0:13:49 do it this way,
0:13:50 then you get desensitized to it.
0:13:52 And so
0:13:53 I think the mistake
0:13:53 a lot of
0:13:55 particular engineers make
0:13:57 is they just don’t
0:13:58 say what they think
0:13:59 when they think it
0:14:00 because they’re afraid
0:14:01 of hurting someone’s feelings.
0:14:02 But that’s how you save
0:14:03 their feelings
0:14:04 because they’re used to you.
0:14:05 You’re always doing that
0:14:06 and you’re doing it
0:14:06 with everybody.
0:14:07 They see it.
0:14:07 They’re like,
0:14:07 oh yeah,
0:14:07 fuck,
0:14:08 Ben’s an asshole.
0:14:10 He’s like always doing this.
0:14:11 But that’s how he is
0:14:12 and that’s how we work
0:14:12 and it’s no problem
0:14:13 as opposed to
0:14:15 the hammer.
0:14:16 And you try and put it
0:14:17 in a shit sandwich.
0:14:18 Oh, you do this really well
0:14:19 but this is all fucked up
0:14:19 and this is good.
0:14:21 And people are like,
0:14:21 well,
0:14:22 so like
0:14:24 now in my written review
0:14:25 you’re telling me
0:14:26 that for the first time
0:14:27 this is all fucked up.
0:14:27 Fuck you.
0:14:29 And this is very common
0:14:30 and you can see this
0:14:30 in the industry
0:14:31 that the extreme version
0:14:32 of it is like
0:14:32 they get fired
0:14:33 and then head of HR
0:14:34 talks to them
0:14:34 and they’re like,
0:14:35 you know,
0:14:36 do you see this coming?
0:14:36 It was obvious, right?
0:14:37 You knew it.
0:14:38 No, I had no idea.
0:14:39 Like, wait,
0:14:40 you didn’t get any feedback
0:14:40 on this?
0:14:41 No.
0:14:42 I only got thumbs up
0:14:42 all along
0:14:43 for like all year
0:14:44 so I’m in shock.
0:14:46 This is super common, right?
0:14:48 So maybe on the topic
0:14:49 of managing talent,
0:14:50 you have this incredibly
0:14:51 high intensity culture
0:14:52 at Databricks
0:14:53 and there was this thread
0:14:54 recently in our CEO thread
0:14:56 where they asked everyone
0:14:57 but you had a great response
0:14:58 on,
0:14:58 hey,
0:14:59 we have 50 people,
0:15:00 how do we scale?
0:15:02 We have this culture
0:15:02 of 996, right?
0:15:03 You work 99,
0:15:04 six days a week.
0:15:06 How have you scaled
0:15:07 that intensity
0:15:09 well into 10,000 employees?
0:15:11 I think start with,
0:15:11 you know,
0:15:12 setting the tone at the top.
0:15:13 If you’re the hardest
0:15:14 working person,
0:15:15 you know,
0:15:16 it kind of everything
0:15:16 will take care of itself
0:15:17 from there on.
0:15:20 If you’re not working hard,
0:15:20 it’s very hard.
0:15:21 I mean,
0:15:21 if you have,
0:15:21 you know,
0:15:22 it’s a double standard.
0:15:22 I mean,
0:15:23 Ben has a whole book
0:15:23 about that,
0:15:24 which is,
0:15:24 you know,
0:15:25 it’s basically,
0:15:26 you know,
0:15:27 what you do is who you are
0:15:28 is the whole title
0:15:29 of the book, right?
0:15:29 So it’s like,
0:15:30 if you are working
0:15:30 extremely,
0:15:31 extremely hard,
0:15:32 the rest of the organization
0:15:33 is also as well.
0:15:34 You know,
0:15:35 are you calling people
0:15:35 at 9 p.m.,
0:15:36 10 p.m.?
0:15:37 Are you working weekends?
0:15:38 Do they expect you?
0:15:39 Not that you expect them
0:15:39 and you’re going to be angry
0:15:40 and yell at them
0:15:41 if they’re not dropping
0:15:41 everything for you.
0:15:42 Not that.
0:15:43 But the fact that they just know
0:15:44 that Ali’s working 24-7
0:15:45 and he’s working
0:15:45 seven days a week
0:15:46 and, you know,
0:15:47 he’s working at 11 p.m.
0:15:47 or 2 a.m.
0:15:48 or whatever it is.
0:15:49 I think that gets
0:15:50 a lot of it done.
0:15:51 The second thing
0:15:52 I would say is
0:15:54 you can vet for this
0:15:54 when you hire people.
0:15:55 You know,
0:15:56 and the easiest way
0:15:57 to vet for it
0:15:58 because, you know,
0:16:01 you got to be careful
0:16:01 with it
0:16:02 because the people
0:16:03 who say they’re going
0:16:03 to work the hardest
0:16:04 are not the ones
0:16:05 who work the hardest.
0:16:06 It’s the opposite.
0:16:08 Yeah, 100% true, right?
0:16:09 So, the best way
0:16:09 to vet for this
0:16:10 is to do backdoors.
0:16:11 You know, ask people.
0:16:12 People don’t,
0:16:13 if I ask someone,
0:16:13 you know,
0:16:15 hey, how was Sarah?
0:16:16 Like, did you like,
0:16:16 was she great?
0:16:17 They’re always going to say,
0:16:18 yeah, she was great.
0:16:18 Right.
0:16:19 But they’re going to be
0:16:19 much more honest
0:16:20 if you ask them like,
0:16:22 hey, how much does she
0:16:22 like grind the midnight oil?
0:16:23 Is she like,
0:16:25 and they’ll tell you right away.
0:16:25 It’s like, oh my God,
0:16:26 she works like crazy.
0:16:28 It’s like, you know,
0:16:28 she’s, I mean,
0:16:29 I think she has a good balance.
0:16:30 You know,
0:16:31 you can like suss that up
0:16:31 very easily.
0:16:32 From backdoors, references,
0:16:33 you can,
0:16:34 people will remember the people
0:16:35 and they’ll just offer it up
0:16:35 and say, oh,
0:16:37 that person was like nuts.
0:16:38 They were like working 24-7.
0:16:40 So, I think that way
0:16:40 you can get people
0:16:41 that are hardworking.
0:16:41 By the way,
0:16:42 I don’t want to overemphasize.
0:16:43 I don’t think everything
0:16:44 is just work harder.
0:16:44 You know,
0:16:45 you have to also work smarter.
0:16:48 And I think that you want
0:16:49 to make sure that
0:16:50 it’s sustainable.
0:16:52 I can work insanely hard.
0:16:52 I’m motivated.
0:16:55 Everybody has a different threshold
0:16:56 for how hard they can work.
0:16:58 I don’t think you want
0:16:58 a culture where people
0:16:59 are burning out.
0:17:00 I think you really
0:17:01 should avoid that.
0:17:01 In fact,
0:17:01 you know,
0:17:02 at Databricks,
0:17:03 I’m very often going in
0:17:03 and saying,
0:17:03 hey,
0:17:04 this team,
0:17:04 like,
0:17:05 you know,
0:17:06 your scores are really bad
0:17:06 on work-life balance.
0:17:07 Like,
0:17:07 what are you doing about it?
0:17:08 Or you guys should take
0:17:10 several days off
0:17:10 or you should do some offsides
0:17:11 or do something.
0:17:11 Like,
0:17:12 we actually go in
0:17:13 and if we see that
0:17:14 there’s some groups.
0:17:14 And other groups
0:17:15 at Databricks,
0:17:16 you know,
0:17:17 their work-life balance scores
0:17:18 are like 100%.
0:17:18 They’re like,
0:17:18 you know,
0:17:19 slacking off.
0:17:21 You know,
0:17:22 so then it’s kind of the opposite.
0:17:24 But I do think
0:17:24 that you can kind of
0:17:25 up for that.
0:17:28 And I think that also
0:17:29 setting the expectation.
0:17:30 You know,
0:17:31 I would say,
0:17:31 you know,
0:17:33 one of my competitors,
0:17:34 Frank Slootman,
0:17:34 wrote a book
0:17:35 called Amp It Up.
0:17:36 It’s a great book
0:17:37 on how you get,
0:17:38 you know,
0:17:39 execution into a company.
0:17:40 Like,
0:17:41 getting a high-performance
0:17:41 culture read.
0:17:43 Everybody’s always trying
0:17:43 to excel
0:17:44 and do better
0:17:45 and better.
0:17:46 Sort of that kind
0:17:46 of culture into a company.
0:17:47 And that’s a good book
0:17:49 if you want to just study.
0:17:49 And he’s doing it
0:17:50 at scale
0:17:50 at bigger companies.
0:17:51 So I think
0:17:52 that’s highly recommended
0:17:52 reading as well.
0:17:53 Yeah,
0:17:53 and I think,
0:17:55 you know,
0:17:56 a lot of it
0:17:58 at his scale
0:17:59 ends up being
0:18:00 things like
0:18:02 organizational design
0:18:03 and do,
0:18:05 are people feeling
0:18:05 like they’re having
0:18:06 an impact when they’re,
0:18:06 like,
0:18:08 if people are feeling
0:18:09 like they’re having
0:18:09 an impact
0:18:10 and they’re good,
0:18:11 then they’ll work
0:18:11 very hard.
0:18:13 But if you’re
0:18:14 in some kind
0:18:14 of weird
0:18:16 three-legged race
0:18:17 that the CEO
0:18:18 has constructed
0:18:19 where everybody’s
0:18:19 got dependencies
0:18:20 on everybody else,
0:18:23 it just doesn’t matter.
0:18:24 You know,
0:18:24 like,
0:18:24 you’ll just have
0:18:25 a lot of people
0:18:25 who go,
0:18:25 like,
0:18:27 I know if I work hard,
0:18:27 it’s not going
0:18:28 to make a difference.
0:18:28 So,
0:18:29 like,
0:18:29 why would I do that?
0:18:30 And,
0:18:30 like,
0:18:31 you can’t
0:18:32 overcome that
0:18:33 with rah-rah
0:18:34 and,
0:18:34 you know,
0:18:36 lead by example
0:18:36 or anything else.
0:18:36 Like,
0:18:37 that’s just
0:18:37 fundamental
0:18:38 to how it is.
0:18:39 And you see,
0:18:40 in,
0:18:40 like,
0:18:41 in any company
0:18:41 of any scale,
0:18:43 you know,
0:18:45 even at our scale,
0:18:46 like,
0:18:46 there are some groups
0:18:49 who really
0:18:51 can have impact
0:18:51 and work
0:18:52 extremely hard
0:18:53 and then groups
0:18:54 who have lesser impact
0:18:55 will work less hard
0:18:56 and you just see that.
0:18:56 You know,
0:18:57 people who are motivated
0:18:58 and they feel excited
0:18:58 about work
0:18:59 and they don’t see
0:19:00 the impact that they’re having,
0:19:00 they’re going to work
0:19:01 way,
0:19:01 way,
0:19:01 way harder.
0:19:03 Versus if you’re demoralized
0:19:04 and you feel like
0:19:05 it’s not going well,
0:19:06 I’m not having impact,
0:19:06 I don’t have any autonomy,
0:19:08 then,
0:19:09 you know,
0:19:10 you’re not going to,
0:19:11 you just don’t want to even,
0:19:12 you’re like kind of depressed
0:19:12 sitting down working.
0:19:14 I do think there’s one thing here
0:19:15 where leaders can really help,
0:19:17 which is to make your team
0:19:18 feel like they’re winning
0:19:19 and that they’re doing
0:19:19 a great job.
0:19:20 Like,
0:19:21 you can ask more from people,
0:19:22 but if I feel like,
0:19:24 this thing is wrong
0:19:25 and I’m putting in
0:19:26 all these hours
0:19:27 and it’s stupid,
0:19:28 like there’s like,
0:19:29 what’s the point of this?
0:19:30 Then people don’t want to work.
0:19:30 So,
0:19:31 so I think it’s like
0:19:32 feeling like we’re winning,
0:19:33 like we’re the winning team,
0:19:34 we’re winning,
0:19:34 like,
0:19:34 you know,
0:19:35 and wow,
0:19:36 they’re expecting more from me
0:19:37 and,
0:19:37 you know,
0:19:38 so then I think you can get,
0:19:38 you need that motivation
0:19:39 in people.
0:19:39 Yeah,
0:19:40 yeah,
0:19:40 which is why,
0:19:41 by the way,
0:19:42 the hard job is
0:19:44 when you aren’t winning
0:19:46 to get the output,
0:19:47 like particularly
0:19:50 in Silicon Valley
0:19:50 because,
0:19:51 you know,
0:19:53 you’re battling attrition,
0:19:53 this and that
0:19:54 and to get things
0:19:55 on the right track,
0:19:58 that takes a whole different
0:20:00 kind of level of technique
0:20:01 and storytelling
0:20:03 and show you
0:20:04 how you could be winning
0:20:05 and all that kind of stuff.
0:20:06 That gets very,
0:20:06 very complicated.
0:20:07 We’ve both done that,
0:20:07 right?
0:20:07 Yeah,
0:20:08 yeah.
0:20:09 There’s been phases
0:20:10 in our company’s lives
0:20:11 where we weren’t winning.
0:20:12 Yeah,
0:20:12 yeah.
0:20:13 I mean,
0:20:13 especially,
0:20:13 you know,
0:20:14 the story you had
0:20:15 in Hard Thing About Hard Things,
0:20:16 which is probably
0:20:17 the best business book
0:20:18 I’ve read,
0:20:18 which I read,
0:20:19 by the way,
0:20:20 before starting Databricks
0:20:21 and it influenced us a lot.
0:20:23 you know,
0:20:24 those are super important.
0:20:25 Yeah,
0:20:27 that’s the difficulty.
0:20:29 That’s such an important point
0:20:31 because even if you’re winning,
0:20:32 people got to feel like that.
0:20:34 But if you’re not winning,
0:20:34 getting them to feel
0:20:35 like you’re winning is…
0:20:37 We have a path to winning.
0:20:37 Yeah,
0:20:38 we have a path to winning.
0:20:39 You know,
0:20:40 and it’s like rock solid.
0:20:41 It’s going to work.
0:20:43 but it would demand sacrifice
0:20:44 from all of us,
0:20:44 you know.
0:20:46 And there is no feeling
0:20:47 as good
0:20:49 as when you’re not winning
0:20:51 and then you get it to winning.
0:20:51 Yeah.
0:20:51 Like,
0:20:52 that’s the best feeling.
0:20:54 You can’t replicate that.
0:20:55 Once you’re super successful,
0:20:56 you never can quite
0:20:57 get that feeling again.
0:20:57 Yeah,
0:20:58 that’s true.
0:20:59 But you also never feel
0:21:01 that horrible pain again.
0:21:01 Well,
0:21:02 it’s easier to be the underdog
0:21:03 in some ways,
0:21:03 right?
0:21:04 You know,
0:21:04 you have nothing to lose.
0:21:06 in some ways.
0:21:07 In most ways,
0:21:07 not.
0:21:10 Well,
0:21:11 I want to explore this
0:21:12 leading from the top
0:21:12 because that was kind of
0:21:13 the first thing you started with.
0:21:15 We actually hired
0:21:16 an ex-Databricks employee
0:21:18 to A16Z.
0:21:19 So we have some inside scoop
0:21:20 on your leadership style.
0:21:21 And one of the things
0:21:23 he said was
0:21:23 you have this,
0:21:24 and Ben sort of touched
0:21:25 on this too,
0:21:26 but you have this amazing ability
0:21:28 to be strategic,
0:21:29 help your team focus,
0:21:31 but you’re also very in the weeds.
0:21:31 Like,
0:21:32 you’re giving product feedback,
0:21:33 you respond to emails
0:21:34 super quickly,
0:21:38 and product launch emails,
0:21:39 no matter how small they are,
0:21:40 you’ll respond congrats,
0:21:40 which,
0:21:41 you know,
0:21:42 he found hugely motivating.
0:21:44 How do you do all that?
0:21:44 Like,
0:21:45 and where do you fly high?
0:21:46 Where do you fly low?
0:21:47 Yeah.
0:21:47 By the way,
0:21:49 I respond even to progress reports
0:21:50 on all those products
0:21:51 and I follow them in detail.
0:21:51 Every one of them.
0:21:53 I try to respond to every product.
0:21:53 Insane.
0:21:54 Respond.
0:21:54 But,
0:21:56 look,
0:21:57 I think this is,
0:21:59 if you’re just going to fly high
0:22:00 and give high level
0:22:01 like inspirational speeches
0:22:01 and then,
0:22:02 you know,
0:22:02 we’ll trust them,
0:22:03 we’ll delegate to people,
0:22:04 it’s not going to work.
0:22:05 So,
0:22:06 my way is,
0:22:06 you know,
0:22:07 you got to get in the weeds,
0:22:08 you got to understand,
0:22:09 this is back to what I said
0:22:09 at the very beginning,
0:22:10 like,
0:22:11 how do you become great
0:22:11 at the head of engineering?
0:22:12 How do you hire a great
0:22:13 head of marketing?
0:22:14 The only way you can do that
0:22:16 is by being really excellent at it.
0:22:16 So,
0:22:17 you need to study the game
0:22:18 and become the best.
0:22:19 So,
0:22:22 I try to stay tuned
0:22:23 to all of these things.
0:22:24 There’s this quote,
0:22:25 if you do everything,
0:22:26 you will win.
0:22:27 Yeah.
0:22:27 And then the question is,
0:22:28 have you done everything?
0:22:29 Exactly.
0:22:29 Exactly.
0:22:30 Exactly.
0:22:30 So,
0:22:30 yeah.
0:22:30 So,
0:22:31 you know,
0:22:31 you just,
0:22:33 it takes a lot of effort.
0:22:34 You know,
0:22:34 you need to learn
0:22:35 all your keyboard shortcuts.
0:22:37 But,
0:22:38 but I think that’s,
0:22:39 you know,
0:22:40 people feel motivated that,
0:22:40 hey,
0:22:41 I have like direct relationship.
0:22:41 You know,
0:22:42 we used to say,
0:22:43 we used to have one of the culture principles
0:22:44 used to be,
0:22:44 hey,
0:22:45 be a co-founder.
0:22:47 And we don’t want to have any employees
0:22:48 at Databricks.
0:22:49 We just want co-founders.
0:22:50 So,
0:22:51 and the key point was like,
0:22:51 hey,
0:22:52 you’re kind of the owner of this company.
0:22:54 You’re not just a renter.
0:22:54 Come here.
0:22:55 And yeah,
0:22:56 we can talk about it.
0:22:56 And you can,
0:22:57 you can suggest an idea.
0:22:58 You might have just joined
0:22:59 and you’re straight out of school.
0:23:00 You might have a great idea
0:23:01 for a product.
0:23:02 Tell me about it.
0:23:02 You know,
0:23:04 I’m happy to push it.
0:23:06 And so it’s making people feel
0:23:07 like they have an impact
0:23:08 and they’re inspired
0:23:09 back to Ben’s point.
0:23:10 Then it’s going to be
0:23:11 much more exciting for them,
0:23:12 right?
0:23:13 Then following some bureaucracy.
0:23:15 So I don’t follow the bureaucracy,
0:23:16 basically.
0:23:18 I go talk to anyone I like.
0:23:18 You know,
0:23:19 I try to go to the person
0:23:20 that is actually the closest
0:23:21 to the work that’s being done
0:23:22 at any given time.
0:23:24 But there are some tricks
0:23:25 and rules around how you do that
0:23:25 without breaking
0:23:26 the whole organization.
0:23:27 So you can’t just
0:23:29 willingly talk to anyone.
0:23:30 But yeah,
0:23:31 that’s part of it.
0:23:31 Yeah.
0:23:34 Listening and giving direction
0:23:35 is very different on that.
0:23:36 If you give direction,
0:23:39 you can cause a lot of chaos.
0:23:42 But if you go talk to people,
0:23:43 you listen to understand
0:23:43 the problem
0:23:45 and then send it back
0:23:46 down the chain of command.
0:23:47 It tends to work
0:23:48 very, very well.
0:23:49 But look,
0:23:49 generally,
0:23:51 if you’re a CEO
0:23:54 and you don’t fly
0:23:54 low and fast,
0:23:58 it’s going to be a mess
0:24:01 because you never get the truth
0:24:02 because the truth
0:24:04 never makes it to you
0:24:05 like through your people.
0:24:07 Like if I go talk
0:24:08 to Ali’s executive staff
0:24:09 about what’s going on
0:24:10 in their organization
0:24:11 or anybody’s,
0:24:12 first of all,
0:24:13 they’re going to spin it.
0:24:13 Second of all,
0:24:14 they don’t actually know.
0:24:18 You need to help them
0:24:19 debug their organizations
0:24:21 because they’ve got
0:24:22 a million things going on.
0:24:23 They’re also
0:24:25 kind of going to the problem,
0:24:26 going to the bottleneck,
0:24:27 trying to figure out
0:24:28 what’s happening.
0:24:30 And so,
0:24:32 it’s just a very
0:24:33 unreliable source
0:24:34 of information.
0:24:35 All the knowledge
0:24:36 in a company
0:24:38 is with the individual
0:24:39 contributors
0:24:40 who are doing the work
0:24:41 and the customers
0:24:42 that like there’s no,
0:24:45 there’s no knowledge
0:24:45 with the people
0:24:46 who are talking to you
0:24:47 as CEO
0:24:48 who are on your staff.
0:24:49 That’s not,
0:24:51 that’s not the way
0:24:52 information moves.
0:24:53 And so you got to be,
0:24:55 now he’s like
0:24:58 super fast
0:25:00 and which enables
0:25:01 him to go super low.
0:25:02 But,
0:25:02 you know,
0:25:04 in any given time,
0:25:06 the way to think about it
0:25:07 as a CEO is
0:25:08 it’s not like
0:25:09 you’re spending
0:25:10 the exact amount
0:25:11 of attention
0:25:12 to HR
0:25:13 as you are
0:25:13 to,
0:25:14 you know,
0:25:16 the key engineering project
0:25:16 as you are
0:25:17 to the,
0:25:18 you know,
0:25:20 the key kind of sales
0:25:21 competitive deals
0:25:21 and that,
0:25:22 you have to,
0:25:25 you don’t address
0:25:25 everything evenly.
0:25:27 You can never do that.
0:25:28 It’s just a bad idea.
0:25:28 Now,
0:25:29 you’ll probably
0:25:30 get to everything
0:25:31 eventually,
0:25:32 but you’re not
0:25:33 spending the same
0:25:34 amount of time
0:25:35 on every single
0:25:35 department.
0:25:37 The org chart
0:25:37 is not the way
0:25:38 that company works.
0:25:39 It’s just a
0:25:40 communication architecture.
0:25:41 Yeah,
0:25:42 I think the best way
0:25:42 I would say it is like,
0:25:43 it’s kind of like a T
0:25:45 and one of you broad
0:25:45 and then you have
0:25:46 the leg that goes down
0:25:47 and goes really,
0:25:47 really deep.
0:25:48 But you want to do
0:25:49 that anchoring
0:25:51 and the key thing
0:25:52 is to have a really
0:25:52 good priority order
0:25:54 of what’s most important
0:25:55 and kind of drop
0:25:55 everything else.
0:25:55 Like,
0:25:56 you know,
0:25:56 you drop that T
0:25:57 and go really,
0:25:58 really low.
0:25:58 like,
0:25:59 it might be HR.
0:26:00 I might be deep diving
0:26:01 all the way down to HR,
0:26:02 looking at our HR handbook,
0:26:03 our policy,
0:26:03 everything.
0:26:04 Like,
0:26:04 who is this person?
0:26:05 What happened?
0:26:06 Why is this happening
0:26:06 in that group?
0:26:07 What’s going on in that group?
0:26:08 What’s the culture in that group?
0:26:09 What happened here?
0:26:09 Like,
0:26:10 you might want to do that.
0:26:11 It might be existential
0:26:11 for a company,
0:26:12 as we’ve seen.
0:26:13 Some companies went under
0:26:14 because of HR problems,
0:26:15 right?
0:26:16 or ethical issues
0:26:17 that were going on.
0:26:17 So,
0:26:18 I think having a really
0:26:19 good priority order
0:26:20 is really important.
0:26:21 I think some execs,
0:26:21 like,
0:26:22 they just want to have
0:26:23 a perfect ducks in a row.
0:26:24 I have my weekly one-on-ones,
0:26:25 I have my weekly staff meeting,
0:26:26 I have my weekly this,
0:26:27 and then I do this,
0:26:28 and then we follow the rules,
0:26:28 and we do all of this,
0:26:29 and then that’s like,
0:26:30 just like the top part
0:26:31 of the T.
0:26:32 And then there’s nothing
0:26:33 that goes deep.
0:26:35 And that’s the issue,
0:26:35 I think.
0:26:35 Yeah,
0:26:36 over-systematizing
0:26:38 or making it,
0:26:38 like,
0:26:38 symmetrical.
0:26:42 You don’t have to have
0:26:43 one-on-ones
0:26:45 with the same frequency
0:26:46 of all your staff,
0:26:47 or like,
0:26:48 some of them,
0:26:49 you know,
0:26:49 like,
0:26:50 you can meet very seldom.
0:26:50 Like,
0:26:52 everything is different.
0:26:53 Every part of the company
0:26:53 is different.
0:26:54 You may need to meet
0:26:55 with somebody every day.
0:26:56 Yeah.
0:26:57 And then other people,
0:26:59 you know,
0:27:00 you can meet once a quarter
0:27:01 for now
0:27:02 because it’s just
0:27:03 not that serious.
0:27:06 And you can’t
0:27:08 get caught up
0:27:09 in making everything
0:27:10 fair and symmetric.
0:27:10 Particularly,
0:27:11 like,
0:27:12 your staff,
0:27:12 they’ve got to be able
0:27:13 to deal.
0:27:14 And this is actually
0:27:15 the biggest conversation
0:27:17 that I had
0:27:19 with Ali early on.
0:27:19 It’s like,
0:27:22 if they can’t do it,
0:27:23 they can’t do it.
0:27:25 That’s it.
0:27:26 It’s a wrap.
0:27:27 Yeah,
0:27:27 yeah,
0:27:27 yeah.
0:27:28 Don’t try and fix them.
0:27:29 They can’t be fixed.
0:27:30 It’s not going to happen.
0:27:31 And,
0:27:32 yeah,
0:27:33 it’s a sad lesson,
0:27:34 but an important lesson.
0:27:36 I actually,
0:27:37 I want to turn the
0:27:38 conversation to
0:27:38 an area that
0:27:39 Ben was saying
0:27:40 you had to catch up on,
0:27:40 at least in the beginning,
0:27:41 which is the
0:27:43 BD deal-making stuff,
0:27:44 which is interesting to me
0:27:45 just because I think of you
0:27:46 as like a consummate
0:27:47 deal-maker now.
0:27:48 I feel like you’re
0:27:49 playing chess,
0:27:49 everyone else is
0:27:50 playing checkers.
0:27:52 I want to go back
0:27:53 to 2017
0:27:54 with maybe one of the
0:27:55 first game-changing
0:27:56 deals that you guys did,
0:27:57 and that was the
0:27:58 deal with Microsoft.
0:27:59 Can you guys talk
0:28:00 a little bit more
0:28:01 about how that deal
0:28:02 came about,
0:28:03 anything you’d do
0:28:03 differently?
0:28:04 And by the way,
0:28:06 founders still to this day
0:28:07 ask us about it
0:28:07 because it’s sort of
0:28:08 a model for how
0:28:09 they’d like to do deals.
0:28:11 Maybe I should start
0:28:12 by saying that,
0:28:12 you know,
0:28:13 we had tried to get
0:28:13 close to Microsoft
0:28:14 for a long while.
0:28:15 I think Ben had told us
0:28:15 you need to,
0:28:16 that’s an important partner
0:28:18 because they have
0:28:19 the biggest distribution channel.
0:28:20 You know,
0:28:21 today they have
0:28:22 60,000 sellers.
0:28:23 If you can unlock
0:28:24 that in any small way,
0:28:25 it’s going to be
0:28:25 a game changer for you.
0:28:26 And I had been
0:28:27 CEO for a year,
0:28:28 so I’d been trying
0:28:30 hard to get in there.
0:28:30 And many,
0:28:31 many people offered me,
0:28:31 you know,
0:28:32 hey,
0:28:33 so here’s,
0:28:34 I actually know Satya,
0:28:34 so I’m going to get
0:28:35 you introduced.
0:28:36 And I got multiple
0:28:36 introductions to Satya.
0:28:37 He just like,
0:28:39 either never responded
0:28:40 or just CC’d his EA
0:28:41 and it went into
0:28:42 the EA loop,
0:28:42 you know,
0:28:43 like we’re still
0:28:44 trying to find time.
0:28:45 He’s been so busy
0:28:46 this last six months,
0:28:46 you know.
0:28:50 but then he had a meeting
0:28:50 with Ben
0:28:52 and I think he was here
0:28:52 actually
0:28:54 at A16Z
0:28:55 and
0:28:58 they actually just talked
0:28:59 and I was not actually
0:28:59 in the loop.
0:29:00 And then he called me up
0:29:00 and said,
0:29:01 hey,
0:29:02 I talked to Satya
0:29:03 and I think this is,
0:29:03 he’s excited,
0:29:04 he wants to do this.
0:29:06 And I saved the email.
0:29:08 So Ben introduced me
0:29:08 to Satya
0:29:09 and this was
0:29:11 I think 3 or 4 a.m.
0:29:12 I was like in New York
0:29:14 and the email went to Satya
0:29:15 and then Satya added
0:29:16 like four or five people
0:29:16 to the email thread
0:29:17 and then they added
0:29:18 four or five people.
0:29:19 So like within an hour
0:29:21 I had like 25 emails
0:29:21 in my inbox.
0:29:23 And suddenly all these people
0:29:23 that were not responding
0:29:24 to my emails
0:29:25 from Microsoft
0:29:25 right after Satya
0:29:26 CC’d them
0:29:27 and CC’d the next person
0:29:27 they were all like,
0:29:27 hey,
0:29:29 I’m clearing my calendar,
0:29:30 would love to meet you,
0:29:30 do you have any time
0:29:31 in the next two days
0:29:31 or three days or,
0:29:32 you know.
0:29:33 But really kind of
0:29:35 the original pitch
0:29:36 of what’s the give and get
0:29:38 was Ben and Satya
0:29:39 at A16Z
0:29:40 and they kind of figured it out
0:29:41 and I was not actually even there.
0:29:42 So we had some luck
0:29:44 and then Ali
0:29:47 did a couple of things
0:29:48 that were,
0:29:50 or quite a few things
0:29:51 that were very,
0:29:52 very effective.
0:29:54 So the luck was
0:29:56 at the time,
0:29:58 deal with big companies,
0:29:59 there’s always a timing element.
0:30:01 And there was a company
0:30:02 called Hortonworks
0:30:04 that had a deal
0:30:05 with Microsoft
0:30:06 to provide
0:30:07 some similar
0:30:08 kind of functionality
0:30:10 and they were
0:30:10 basically
0:30:13 putting a gun
0:30:14 to Microsoft’s head
0:30:14 saying like,
0:30:15 you pay us more money
0:30:16 or we’re going to
0:30:17 pull our product.
0:30:18 And they were on-prem
0:30:19 and they were in cloud.
0:30:20 So it was like
0:30:21 a big mismatch also.
0:30:21 Yeah,
0:30:22 so it was a real,
0:30:23 so Microsoft was like
0:30:24 super pissed at them
0:30:26 and wanted to stick it to them.
0:30:28 and so that was,
0:30:28 you know,
0:30:30 so you have Satya
0:30:31 going like,
0:30:32 I think this company’s
0:30:32 interesting
0:30:35 and then this
0:30:36 ground level thing
0:30:37 going like,
0:30:38 we want to fuck
0:30:38 these guys.
0:30:40 And that,
0:30:42 that kind of opened
0:30:43 enough of a door
0:30:45 to get it going.
0:30:46 but there were,
0:30:48 so like one of the
0:30:49 most important things
0:30:50 in the deal
0:30:51 was,
0:30:52 which,
0:30:53 you know,
0:30:54 and John O’Farrell
0:30:55 really emphasized this
0:30:56 for,
0:30:57 for both of us
0:30:57 was,
0:30:58 look,
0:31:00 you got to get them
0:31:01 to put enough,
0:31:03 they’re such a big company
0:31:04 that they’re going to
0:31:05 lose interest
0:31:06 many times.
0:31:07 So if you don’t
0:31:09 have them write you
0:31:10 such a big check
0:31:12 that somebody in there
0:31:13 is going to get fired
0:31:14 if it doesn’t go well,
0:31:14 you’re,
0:31:15 it doesn’t matter
0:31:16 if you get the deal,
0:31:16 you’re going to
0:31:17 lose the deal.
0:31:18 And so,
0:31:20 what we did
0:31:21 is we’re like,
0:31:23 and the technique
0:31:24 that we had was,
0:31:24 okay,
0:31:26 give us a forecast.
0:31:26 Like,
0:31:28 we’re a little company,
0:31:29 we can’t afford
0:31:30 to do this deal,
0:31:31 you know,
0:31:31 we can only afford
0:31:32 to have one partner,
0:31:33 so give us a forecast
0:31:34 of what you’ll do.
0:31:34 Yeah,
0:31:35 because our engineers
0:31:35 are busy.
0:31:36 So like,
0:31:36 they’re going to do
0:31:37 this integration,
0:31:38 that wipes out
0:31:39 12 months of our roadmap.
0:31:40 We don’t have anything else.
0:31:40 You guys have like
0:31:41 many thousands
0:31:42 of engineers.
0:31:42 So this is,
0:31:43 we only have one
0:31:43 of these.
0:31:44 Yeah,
0:31:44 so we,
0:31:45 you know,
0:31:46 whoever can sell them,
0:31:47 we think you can sell
0:31:47 the most,
0:31:47 but we don’t know,
0:31:48 like,
0:31:48 what’s your first,
0:31:48 so,
0:31:49 you know,
0:31:49 like,
0:31:50 kind of challenged
0:31:52 their kind of manhood
0:31:52 a little bit.
0:31:54 And so they come out
0:31:55 with this big-ass forecast
0:31:55 and we’re like,
0:31:56 okay,
0:31:56 great,
0:31:56 just give us
0:31:57 a little portion
0:31:57 of that.
0:31:58 It was a huge deal.
0:31:59 It was a lot of money.
0:32:02 And then Ali said,
0:32:03 look,
0:32:04 you know,
0:32:05 when we got all the way
0:32:06 down to the deal,
0:32:07 he was like,
0:32:08 if I don’t get this number,
0:32:10 Ben’s going to fire me.
0:32:12 And so can you help me out?
0:32:15 It was a very interesting dynamic.
0:32:15 Good cop, bad cop.
0:32:17 It was a very interesting dynamic.
0:32:17 So,
0:32:18 you know,
0:32:19 John O’Farrell had to strategize
0:32:20 with us and told us that,
0:32:20 you know,
0:32:23 they have to do a big pre-commit
0:32:24 because then they have skin
0:32:25 in the game.
0:32:25 Otherwise,
0:32:26 they’re just going to forget.
0:32:27 They’ll do like the PR,
0:32:28 but then they’ll forget about you.
0:32:29 But then when we were trying
0:32:30 to get that from Microsoft,
0:32:31 I remember I was talking
0:32:32 to Takeshi Nomoto,
0:32:32 who is,
0:32:32 you know,
0:32:33 one of the main brains
0:32:34 at Microsoft,
0:32:35 like one of the key strategists there.
0:32:37 And his thing was,
0:32:38 I don’t want to give you
0:32:38 a big commit
0:32:39 because you’re such
0:32:40 a small company.
0:32:41 I’m worried you take this money
0:32:42 and you get drunk off of it
0:32:42 and you’re not going
0:32:43 to do anything afterwards.
0:32:45 And so I had to really
0:32:46 convince him that,
0:32:46 no,
0:32:47 I’m extremely hungry.
0:32:47 Like,
0:32:48 there’s no way,
0:32:49 like I will continue
0:32:50 to have crazy appetites.
0:32:51 Don’t worry about it.
0:32:52 So both sides
0:32:53 were kind of worried
0:32:53 about different things.
0:32:54 But yeah,
0:32:56 the give and get was important
0:32:57 that you said in the beginning,
0:32:57 which was,
0:32:59 they had a gap
0:32:59 in the product portfolio,
0:33:00 right?
0:33:01 They were competing with AWS.
0:33:02 They had a gap at the time
0:33:05 and we had a great product.
0:33:07 They have an amazing
0:33:08 distribution channel.
0:33:09 So like,
0:33:10 in these BD deals,
0:33:13 there always has to be
0:33:13 a give and get
0:33:13 that actually
0:33:14 is kind of commensurate.
0:33:15 And this is why
0:33:16 most of these deals
0:33:17 fall apart
0:33:17 and they don’t work.
0:33:19 There has to be something
0:33:20 that you as a small player
0:33:20 can give
0:33:21 that they don’t have.
0:33:22 and usually you don’t
0:33:23 have anything to give them.
0:33:25 Usually I find
0:33:25 all these small companies
0:33:26 show up and they come,
0:33:26 for instance,
0:33:27 to Databricks now and say,
0:33:28 oh, we’d love for you
0:33:28 to partner with us.
0:33:29 But what am I getting
0:33:30 out of it, right?
0:33:31 You don’t report to me.
0:33:32 I don’t report to you.
0:33:34 So the moment
0:33:35 we’ve closed the deal,
0:33:36 if it’s not good for me,
0:33:37 neither of us
0:33:37 will just do
0:33:38 our side of the bargain.
0:33:40 So there has to be something
0:33:41 in the deal dynamics,
0:33:42 in the construct,
0:33:43 that inherently
0:33:45 is extremely beneficial
0:33:46 both sides.
0:33:47 There has to be a trade
0:33:48 that makes sense.
0:33:49 Microsoft really
0:33:49 wanted that product.
0:33:51 We really wanted
0:33:52 their distribution channel.
0:33:53 So that made it
0:33:53 the perfect marriage
0:33:54 at that time.
0:33:55 You know,
0:33:56 if you don’t have
0:33:56 that giving yet,
0:33:57 it’s not going to work.
0:33:58 And then the other thing
0:33:59 that I think a lot of
0:34:00 entrepreneurs understand
0:34:03 is any big deal
0:34:03 of that size,
0:34:04 you lose at least
0:34:05 three times
0:34:06 before you win it.
0:34:07 and we lost that deal.
0:34:09 Ten times.
0:34:09 Ten times.
0:34:10 And like,
0:34:11 including like
0:34:13 the day before
0:34:13 we were supposed
0:34:14 to launch it.
0:34:15 You know,
0:34:16 the antibodies
0:34:18 came out of the company
0:34:20 and Ali had to
0:34:21 fly up to Redmond
0:34:21 and sit there.
0:34:23 There was one engineer
0:34:23 that just said,
0:34:24 not doing this.
0:34:25 This is not going to go.
0:34:26 We don’t want…
0:34:26 He just…
0:34:27 He was…
0:34:28 They actually put a guy
0:34:29 in place at Microsoft
0:34:31 who was actually super…
0:34:31 He had a great reputation,
0:34:33 but he was a builder.
0:34:35 So he just had
0:34:36 huge problems with this.
0:34:37 It’s like,
0:34:38 this is not a product I built.
0:34:39 Why would we…
0:34:40 Why would I make this successful?
0:34:41 So yeah,
0:34:41 there’s like…
0:34:43 Usually there’s like many times.
0:34:43 So like,
0:34:44 if you don’t have grit,
0:34:45 those deals will die.
0:34:47 Because this deal died
0:34:47 multiple times,
0:34:48 as Ben said.
0:34:49 It was like completely over.
0:34:50 Like it was completely blocked
0:34:51 by some exec
0:34:51 that said,
0:34:52 absolutely not.
0:34:53 I’m blocking it.
0:34:53 It’s veto.
0:34:54 It’s over.
0:34:55 And no one wanted to overrule him.
0:34:56 So you have to go in there
0:34:57 and work.
0:34:58 And the only way we did it,
0:34:59 like they call it the nerd bird.
0:35:01 I would take the,
0:35:01 you know,
0:35:02 SF,
0:35:03 Seattle,
0:35:04 flight up there.
0:35:05 I was up there so much.
0:35:06 I knew all the buildings,
0:35:07 all the rooms,
0:35:07 everything.
0:35:09 So you just have to spend time
0:35:09 on the ground
0:35:10 and talk to as many people
0:35:10 as possible
0:35:11 and sort of influence
0:35:12 that organization from within.
0:35:13 I will say,
0:35:14 look,
0:35:14 you know,
0:35:16 with all the difficulty
0:35:17 of the deal
0:35:18 and,
0:35:19 you know,
0:35:21 and Microsoft being Microsoft,
0:35:22 they’ve been
0:35:24 as good a partner
0:35:25 as,
0:35:26 not only we’ve had
0:35:27 a Databricks,
0:35:28 but an entire portfolio.
0:35:29 I mean,
0:35:29 they’ve really,
0:35:31 you know,
0:35:31 lived up
0:35:32 and delivered
0:35:33 what they said
0:35:34 they would do,
0:35:34 which is,
0:35:36 I think you have to give
0:35:38 Satya a huge credit
0:35:38 because like
0:35:39 in the whole
0:35:40 Gates and Ballmer era,
0:35:42 they were never
0:35:42 that good a partner
0:35:43 to anybody
0:35:44 and he’s really
0:35:44 turned that around
0:35:45 and,
0:35:46 you know,
0:35:47 they’ve been fantastic
0:35:48 with us.
0:35:49 this was around the time
0:35:50 where Satya had taken over
0:35:50 and,
0:35:51 you know,
0:35:52 he was giving to everyone
0:35:53 at Microsoft the book
0:35:54 Growth Mindset
0:35:54 or Mindset,
0:35:55 which is about
0:35:56 Growth Mindset.
0:35:58 So there was this kind of
0:35:58 aura in the air
0:35:59 that,
0:35:59 you know,
0:36:00 we should try.
0:36:00 Like,
0:36:01 let’s try to make things happen.
0:36:02 Let’s have a growth mindset here.
0:36:03 Let’s see.
0:36:04 Is there a way we can partner?
0:36:05 So this would have been impossible
0:36:06 five years earlier.
0:36:06 So,
0:36:07 it’s kudos to Satya
0:36:08 and they put us on the map
0:36:10 and he’s been a great partner
0:36:10 ever since,
0:36:11 you know,
0:36:11 whenever there’s been issues,
0:36:13 they always resolve it.
0:36:14 So,
0:36:15 you know,
0:36:16 we are very thankful.
0:36:17 wouldn’t be where we are
0:36:18 without them.
0:36:18 Yeah,
0:36:20 just amazing.
0:36:20 Amazing,
0:36:21 really.
0:36:22 I want to open up the conversation
0:36:24 to deal making more broadly
0:36:25 now that you’re not
0:36:27 a small company anymore
0:36:27 and you’re a big company
0:36:28 making acquisitions,
0:36:29 you know,
0:36:30 Tabular,
0:36:30 Neon,
0:36:31 Mosaic,
0:36:32 just to name a few.
0:36:33 What is your sort of
0:36:34 your approach
0:36:34 in terms of
0:36:36 when to build
0:36:37 versus when to buy
0:36:37 slash how do you think about
0:36:39 sort of acquisitions
0:36:39 more broadly?
0:36:39 Yeah,
0:36:40 I mean,
0:36:41 what we try to not do,
0:36:42 so let’s start with a simple thing,
0:36:44 is a lot of companies,
0:36:45 especially at scale,
0:36:46 they’ll buy revenue
0:36:47 so they’ll look at a company,
0:36:48 they’ll say,
0:36:48 hey,
0:36:49 this company is this size,
0:36:51 we’ll just buy that company,
0:36:53 we’ll put more salespeople on it,
0:36:54 then we can accelerate the revenue,
0:36:55 we’re buying that revenue,
0:36:57 and that’s how they’re doing it.
0:36:59 We’re not doing that.
0:37:00 You know,
0:37:01 what we’re really doing is,
0:37:01 number one,
0:37:02 we spend a lot of time
0:37:03 with the team and the founders,
0:37:04 so we’re trying to see,
0:37:04 hey,
0:37:05 can we build together?
0:37:05 Like,
0:37:06 you come here
0:37:07 and you build together.
0:37:08 That’s very different
0:37:09 from that buying revenue model.
0:37:10 The buying revenue model,
0:37:11 oftentimes you part ways
0:37:12 with the CEO
0:37:13 from day one.
0:37:14 you can see it,
0:37:15 the big companies,
0:37:16 they literally have a plan.
0:37:17 Like,
0:37:18 I have some execs
0:37:19 that come from these big companies
0:37:19 to say,
0:37:19 hey,
0:37:20 our plan usually
0:37:21 is to part ways with the CEO.
0:37:21 Like,
0:37:22 you make a deal
0:37:23 and the CEO can leave.
0:37:24 And then,
0:37:25 but also the key people
0:37:26 in those companies
0:37:27 quickly leave,
0:37:27 all of them.
0:37:27 like,
0:37:28 the top management
0:37:28 and then,
0:37:29 you know,
0:37:29 you keep promoting
0:37:30 the people from below
0:37:31 that couldn’t get promoted before
0:37:32 and then eventually
0:37:33 you bring in your own people
0:37:34 to take over the company
0:37:35 and then the company is dead.
0:37:37 There’s nothing left of it
0:37:38 and there’s no integration
0:37:39 between that asset
0:37:39 that you bought
0:37:40 and the platform
0:37:41 that you have.
0:37:42 So,
0:37:43 to avoid all of those,
0:37:44 can you get people
0:37:45 that really feel like
0:37:46 they’re your co-founders?
0:37:47 So,
0:37:47 we spent just
0:37:48 an enormous amount of time
0:37:49 with who we’re buying.
0:37:49 Like,
0:37:50 the company we’re buying,
0:37:51 who are the founders?
0:37:52 How do they work?
0:37:53 Are we culturally the same?
0:37:54 Spend time with them.
0:37:54 Do we get along?
0:37:55 Do we see the world
0:37:55 the same way?
0:37:56 You know,
0:37:57 are we going to click?
0:37:58 Are we going to do this together?
0:37:59 Are we going to be able
0:38:00 to build in the next five years?
0:38:00 So,
0:38:02 that’s where we spend,
0:38:02 number one.
0:38:03 Number two,
0:38:04 we spend a lot of time
0:38:06 on the product.
0:38:06 You know,
0:38:07 what’s the product experience?
0:38:08 How would we integrate this?
0:38:09 What would it look like?
0:38:09 How much can we,
0:38:11 can we rewrite most of it?
0:38:11 Can we not rewrite it?
0:38:12 What’s the,
0:38:13 what programming,
0:38:14 I always ask this
0:38:14 and people are like,
0:38:14 wait,
0:38:15 that’s such a dumb question.
0:38:16 I ask,
0:38:16 hey,
0:38:17 what language did you write it in?
0:38:18 Why do you want to do that?
0:38:19 What does that matter?
0:38:19 No,
0:38:26 so the product is something
0:38:27 we spend a huge amount of time
0:38:28 and talking to customers,
0:38:29 understanding what the,
0:38:30 what the,
0:38:32 what the excitement around
0:38:33 that product looks like
0:38:33 and how the integration
0:38:34 would look like.
0:38:35 The last thing we do
0:38:36 is we’ll look at the financials.
0:38:36 You know,
0:38:38 what’s the revenue multiple
0:38:39 and,
0:38:39 you know,
0:38:40 how much can we grow it
0:38:41 and what’s the three-year plan,
0:38:42 five-year plan and so on.
0:38:43 And I feel like
0:38:44 big companies,
0:38:45 cooperative departments
0:38:45 do it exactly
0:38:46 in the reverse order of this.
0:38:47 They start with,
0:38:48 hey,
0:38:48 the revenue is this
0:38:49 but we could accelerate it
0:38:50 and the multiple is so low
0:38:51 and like,
0:38:51 you know,
0:38:52 in this,
0:38:52 in my Excel sheet here,
0:38:53 this makes perfect sense.
0:38:55 you know,
0:38:55 and then second,
0:38:56 they go to like,
0:38:56 hey,
0:38:57 is this a good product?
0:38:58 And then lastly,
0:38:58 like,
0:38:58 hey,
0:38:59 how do we convince
0:38:59 these knuckleheads?
0:39:00 I mean,
0:39:00 we probably don’t want
0:39:01 to have them here
0:39:02 but we got to pay them off somehow.
0:39:03 So,
0:39:04 so I think,
0:39:04 you know,
0:39:05 thinking about it that way,
0:39:06 you get more longevity
0:39:06 out of it.
0:39:07 Yeah,
0:39:09 and this is,
0:39:11 this really comes,
0:39:12 it sounds like he’s talking
0:39:16 like a product guy
0:39:18 but this is really the thing
0:39:19 that people get wrong
0:39:19 on the go-to-market
0:39:21 because what happens is
0:39:22 if you’ve got
0:39:24 multiple product architectures,
0:39:27 that’s going to mean
0:39:28 multiple SE forces,
0:39:30 multiple post-sales things
0:39:34 and your entire sales efficiency
0:39:35 is going to go through the floor.
0:39:37 And because,
0:39:38 you know,
0:39:40 they have a keen eye on that,
0:39:41 everything they buy
0:39:42 ends up looking like
0:39:42 a Databricks product,
0:39:43 you know,
0:39:43 like,
0:39:45 and that work is going in,
0:39:46 they’re not just selling
0:39:47 some shit to get some money
0:39:48 to,
0:39:48 you know,
0:39:50 go on a corp dev thing.
0:39:51 and I would say
0:39:52 so many,
0:39:54 like,
0:39:54 when you bring in
0:39:55 a professional CEO,
0:39:57 this is what they screw up
0:39:59 because they don’t understand
0:40:00 that,
0:40:00 yeah,
0:40:01 engineering goes,
0:40:01 yeah,
0:40:01 yeah,
0:40:02 we can take it on,
0:40:03 there’s another set of engineers,
0:40:04 we don’t care if they work on that,
0:40:05 blah, blah, blah.
0:40:07 And engineering gets less efficient too
0:40:08 but it wrecks the field
0:40:10 and that’s,
0:40:12 and then the customers hate it
0:40:13 because it’s like,
0:40:13 yeah,
0:40:14 like,
0:40:14 okay,
0:40:15 I’ve got to learn another
0:40:16 access control model,
0:40:17 I’ve got to do this,
0:40:17 I’ve got to,
0:40:17 you know,
0:40:19 these are not things
0:40:20 anybody wants to be part of.
0:40:21 Yeah,
0:40:22 100%.
0:40:22 Yeah,
0:40:24 it’s the go-to-market side
0:40:25 that you’re worried about,
0:40:25 that,
0:40:25 you know,
0:40:26 that experience that those
0:40:27 customers will have.
0:40:28 You know,
0:40:29 they’re going to come back
0:40:30 immediately and say,
0:40:30 hey,
0:40:31 we were already upset
0:40:32 about these things
0:40:32 before the acquisition,
0:40:34 maybe you can fix them now.
0:40:34 It’s like,
0:40:34 no,
0:40:35 actually,
0:40:35 you know,
0:40:36 several of those people
0:40:37 actually quit
0:40:38 and now we’re going
0:40:38 to just work on integration
0:40:39 and that thing now
0:40:40 got pushed out
0:40:40 another two years.
0:40:42 So you don’t want
0:40:42 to be in that situation.
0:40:43 So there’s a lot
0:40:43 of companies that do that
0:40:44 and by the way,
0:40:45 what they’re doing works
0:40:46 revenue-wise.
0:40:47 They are getting the revenue,
0:40:48 they are getting the,
0:40:49 the stock swap works.
0:40:50 Like,
0:40:50 you know,
0:40:50 if they’re multiple
0:40:51 of the companies,
0:40:52 it’s a creative deal
0:40:53 temporarily.
0:40:54 Yeah,
0:40:54 it works
0:40:55 and then the first year
0:40:55 you get the bump
0:40:56 in the revenue
0:40:57 and you get a second year
0:40:58 boost in revenue growth
0:40:58 as well.
0:41:00 So financial engineering
0:41:01 actually works great
0:41:01 for those companies.
0:41:02 It’s just long-term
0:41:03 ends up being like,
0:41:04 you know,
0:41:05 a bag of crap
0:41:06 that doesn’t work together.
0:41:07 And it affects the brand.
0:41:08 You know,
0:41:08 like one of the things
0:41:09 is,
0:41:11 one of the reasons
0:41:13 Databricks is so powerful
0:41:14 is all their customers
0:41:15 want to buy
0:41:16 all their products
0:41:17 because they’re like,
0:41:18 we know that’s
0:41:20 the best software
0:41:20 we buy
0:41:22 and as soon as
0:41:23 you start
0:41:24 chipping away at that
0:41:25 with these financial
0:41:26 strategies,
0:41:27 like you can’t
0:41:27 get it back
0:41:28 because the reputation
0:41:31 is every customer’s
0:41:31 experience.
0:41:32 There’s,
0:41:33 there’s no marketing
0:41:33 through that.
0:41:35 It’s the best software
0:41:36 because it was written
0:41:38 by the engineers
0:41:39 and built by those
0:41:39 that were the best.
0:41:40 Yeah.
0:41:41 Including the acquisitions
0:41:41 that we bought.
0:41:42 They were phenomenal people
0:41:43 that came in
0:41:43 and they continued
0:41:44 and since we gelled
0:41:45 they continued building it.
0:41:47 So that’s why it’s great.
0:41:47 It’s like the,
0:41:47 you know,
0:41:49 so we pay a lot of attention.
0:41:50 That’s like back to the,
0:41:50 you know,
0:41:50 who are you getting
0:41:51 into your company?
0:41:52 Yeah.
0:41:53 Yeah.
0:41:55 That’s the other thing,
0:41:55 right?
0:41:57 Like you can buy
0:41:58 something that’s
0:41:59 got a lot of sales
0:42:00 where you’re
0:42:01 downgrading
0:42:02 your whole
0:42:04 like company.
0:42:05 Ross Perot actually
0:42:05 wrote about,
0:42:07 in Citizen Perot,
0:42:09 his biggest fear,
0:42:10 which definitely
0:42:10 came true,
0:42:11 was he built
0:42:12 this elite thing
0:42:13 at EDS
0:42:14 and then
0:42:15 they would actually
0:42:17 acquire IT departments
0:42:17 and they’re like,
0:42:18 he was like,
0:42:20 they’re going to
0:42:21 absorb us,
0:42:22 not vice versa
0:42:23 and that does happen.
0:42:25 there is one
0:42:25 really good company
0:42:26 that,
0:42:27 well,
0:42:28 one really successful
0:42:28 company that
0:42:29 we never acquired
0:42:31 and I always vetoed it
0:42:32 whenever it came up
0:42:32 because I just think
0:42:33 that the quality
0:42:34 of their employee base
0:42:34 is not great
0:42:35 and I didn’t want it
0:42:36 to dilute Databricks.
0:42:37 Otherwise,
0:42:38 from every other angle,
0:42:40 that deal always made sense
0:42:41 and I always vetoed it
0:42:42 because I felt that,
0:42:43 you know,
0:42:43 it’s just,
0:42:45 they’re all going to quit
0:42:46 or be super unhappy
0:42:47 or let’s just not do it.
0:42:47 Yeah.
0:42:48 It’s also why
0:42:49 like merger of equals
0:42:49 are,
0:42:51 because the cultures
0:42:52 aren’t equal,
0:42:53 the people aren’t equal.
0:42:55 And what made you feel
0:42:55 that way?
0:42:56 You just spent time
0:42:56 with them
0:42:57 and they just didn’t
0:42:58 exude to Databricks culture?
0:42:59 Well,
0:42:59 I mean,
0:43:00 it’s,
0:43:00 look,
0:43:01 it’s like everything else.
0:43:02 Like,
0:43:02 it’s like when we were
0:43:03 grading students
0:43:04 at the university,
0:43:04 it’s like,
0:43:04 okay,
0:43:04 the rock stars
0:43:05 are super easy
0:43:05 to find out
0:43:06 so they’re like there
0:43:07 and then the people
0:43:07 that are really,
0:43:07 really bad,
0:43:08 that’s like,
0:43:08 it’s not hard
0:43:09 and then there are
0:43:09 people in the middle
0:43:11 that’s in the gray zone.
0:43:12 This was a company
0:43:12 that was,
0:43:13 you know,
0:43:14 I feel like the talent
0:43:15 is not phenomenal
0:43:15 and you don’t need
0:43:16 to be a genius
0:43:16 to know that
0:43:17 and then there’s
0:43:18 some startups
0:43:18 you know immediately.
0:43:19 Like,
0:43:19 you know,
0:43:19 okay,
0:43:20 these guys are
0:43:20 Olympiad winners
0:43:21 and they’re like phenomenal
0:43:22 and they’re like executing
0:43:23 like crazy
0:43:23 and they have a track record.
0:43:25 So,
0:43:25 I don’t think those
0:43:26 are that hard
0:43:27 and we try to hire these
0:43:28 and this is the one
0:43:29 that I vetoed.
0:43:29 The hard part is
0:43:30 what do you do
0:43:30 with the ones in the middle?
0:43:31 That’s always where
0:43:32 you spend all of your energy
0:43:33 trying to suss out.
0:43:33 Like,
0:43:33 you know,
0:43:33 okay,
0:43:34 they’re not stellar stellar
0:43:35 but maybe they are,
0:43:36 maybe they just didn’t have,
0:43:37 maybe they didn’t have
0:43:37 the go-to-market,
0:43:38 they didn’t have the funding,
0:43:39 they didn’t have the support
0:43:40 that they needed
0:43:40 and so on.
0:43:41 Maybe they could
0:43:42 if we give them a chance
0:43:43 or maybe they’re just mediocre
0:43:45 and that’s where you spend
0:43:45 a lot of your time
0:43:46 but you got to spend time
0:43:46 with them.
0:43:47 You have to interview
0:43:47 all the people.
0:43:48 You know,
0:43:49 you have to have your people
0:43:50 interview all the people.
0:43:50 Can’t just be,
0:43:52 this can’t be an Excel sheet exercise.
0:43:53 Yeah.
0:43:54 And Silicon Valley
0:43:55 has a lot of lopsided companies
0:43:56 so,
0:43:56 you know,
0:43:57 you’ll have
0:43:59 a great engineering team
0:43:59 and a bad company
0:44:01 because like,
0:44:01 you know,
0:44:02 bad leadership,
0:44:03 bad go-to-market.
0:44:04 You also can have
0:44:05 like,
0:44:07 guys who can sell anything
0:44:08 with a ridiculously,
0:44:09 like,
0:44:10 poor engineering team
0:44:11 and they can just sell it
0:44:12 and,
0:44:13 you know,
0:44:13 you got to be
0:44:14 very,
0:44:15 very careful about that.
0:44:15 Actually,
0:44:16 our,
0:44:17 you know,
0:44:19 our CRO at Databricks
0:44:20 is,
0:44:20 you know,
0:44:21 he came from a company
0:44:21 that,
0:44:22 you know,
0:44:23 he’d sell anything.
0:44:27 He was selling SFTP,
0:44:27 secure FTP,
0:44:28 which is free.
0:44:30 And he was selling it
0:44:31 for a lot.
0:44:31 That’s how you know he’s good.
0:44:32 He was selling it for a lot.
0:44:33 He was making a lot of money.
0:44:33 He was saying,
0:44:33 you know,
0:44:35 the electronic medical health records,
0:44:35 you know,
0:44:36 how important are they?
0:44:37 If they got dropped,
0:44:37 you know,
0:44:39 how much of a risk
0:44:39 is it to your business?
0:44:39 Well,
0:44:40 this is secure FTP.
0:44:43 You need it to be secure.
0:44:46 Me on somebody grabbing that file.
0:44:47 He’s good.
0:44:48 Yeah,
0:44:49 the only thing I’d add too
0:44:50 is this strategy
0:44:51 is probably making you
0:44:52 more attractive
0:44:52 to the people
0:44:53 you want to acquire too.
0:44:54 They don’t want to sell
0:44:55 if they’re going to get
0:44:55 fired right away.
0:44:55 Yeah,
0:44:56 for sure.
0:44:56 very competitive.
0:44:57 Yeah,
0:44:57 100%.
0:44:58 Yeah.
0:44:58 I mean,
0:44:58 you know,
0:44:59 there’s also a reputation,
0:44:59 right?
0:45:00 People know,
0:45:00 like they’ll look back
0:45:01 and say,
0:45:01 okay,
0:45:01 well,
0:45:01 what happened
0:45:02 to your previous acquisitions?
0:45:03 You know,
0:45:04 was there a huge fight
0:45:04 and everybody’s quitting
0:45:05 left and right?
0:45:05 Or,
0:45:05 you know,
0:45:06 did they work out?
0:45:06 You know,
0:45:07 how are you taking care
0:45:07 of those people?
0:45:08 You know,
0:45:09 what roles do they have?
0:45:10 Do they have influential roles
0:45:11 in your company?
0:45:12 You know,
0:45:13 that’s also important.
0:45:13 So,
0:45:14 you’re setting a precedent.
0:45:15 You’re setting a precedent
0:45:15 in many,
0:45:16 many ways with acquisitions,
0:45:17 M&A,
0:45:17 you know,
0:45:18 deal dynamics,
0:45:27 is a precedent
0:45:27 for the next deal.
0:45:28 Yep.
0:45:29 Totally.
0:45:29 Yeah.
0:45:31 Maybe actually just to turn,
0:45:32 so we’re talking about
0:45:33 Databricks as an acquirer.
0:45:34 If we go back in time again
0:45:36 to maybe a moment
0:45:37 where you thought about selling
0:45:38 and maybe that,
0:45:38 you know,
0:45:40 you didn’t actually
0:45:41 seriously consider that.
0:45:42 But,
0:45:44 I wanted to actually
0:45:45 just sort of quote
0:45:46 this infamous email
0:45:48 sort of circulating our firm
0:45:50 that Ben sent to Ali.
0:45:50 Yeah,
0:45:51 Ali brought it,
0:45:52 I had forgotten about it.
0:45:52 He brought it up
0:45:53 at a board dinner.
0:45:53 Okay, go.
0:45:54 And I was like,
0:45:54 oh shit,
0:45:55 I said that.
0:45:57 And actually,
0:45:59 this was not pertaining
0:46:00 to selling the company,
0:46:00 but it was,
0:46:00 I think,
0:46:01 selling a candidate,
0:46:02 right?
0:46:04 And you talk about,
0:46:05 hey Ben,
0:46:06 can you sell this candidate
0:46:07 on the fact that
0:46:08 it will be worth $10 billion?
0:46:09 Maybe $25.
0:46:09 Yeah, because the candidate
0:46:10 was worried about
0:46:11 that company getting sold.
0:46:12 Yeah, he wanted to have
0:46:13 a double trigger
0:46:14 because the company,
0:46:15 if Databricks sells,
0:46:18 and they fire me
0:46:18 as a salesperson,
0:46:20 what equity am I going to get?
0:46:21 So give me double trigger
0:46:21 so I’m protected.
0:46:22 If we get bought
0:46:23 and I get fired,
0:46:25 I invest all my equity immediately.
0:46:26 Yep, exactly.
0:46:27 And so,
0:46:27 you know,
0:46:29 in response to this,
0:46:30 Ben,
0:46:31 and I’m going to
0:46:32 paraphrase this a little bit,
0:46:34 but he writes back to you,
0:46:35 and I’m like thinking
0:46:36 about Ben’s tone in this,
0:46:36 you are severely
0:46:38 underselling the opportunity.
0:46:39 We are Oracle in the cloud,
0:46:40 and we will be worth
0:46:42 10x what Oracle is.
0:46:44 But what was your reaction
0:46:45 when you saw that?
0:46:46 And did that give you
0:46:47 more fortitude
0:46:48 to not sell the company?
0:46:50 Yeah, Ben’s crazy.
0:46:52 I think the first thought
0:46:54 was exactly Ben’s crazy.
0:46:56 But no,
0:46:58 I think both Ben and Mark
0:47:01 always kind of pushed us
0:47:01 to think bigger.
0:47:03 I remember we did the pitch
0:47:04 at A16Z
0:47:06 for I think our Series D,
0:47:07 which would have been
0:47:08 around 2017 or so.
0:47:11 And the question was asked,
0:47:12 you know,
0:47:13 what’s your biggest bottleneck?
0:47:13 And I said,
0:47:14 biggest bottleneck is hiring.
0:47:15 I said,
0:47:15 okay,
0:47:15 well,
0:47:17 who are you losing to?
0:47:18 And I said,
0:47:18 well,
0:47:18 it’s Google,
0:47:19 you know,
0:47:20 it’s the Fangs.
0:47:22 And the response
0:47:24 I got back was,
0:47:24 well,
0:47:25 you need to just add
0:47:26 the Databricks to Fang.
0:47:26 It needs to be FangDB.
0:47:28 With a straight face.
0:47:30 And it was like,
0:47:30 and I was like,
0:47:32 and my reaction was I laughed.
0:47:32 I literally said,
0:47:33 oh, yeah, yeah.
0:47:34 I mean,
0:47:34 this is not serious.
0:47:34 Like,
0:47:35 it’s like,
0:47:35 yeah,
0:47:36 that’s the bottleneck.
0:47:36 It’s like,
0:47:36 no,
0:47:37 I’m serious.
0:47:38 You need to add Databricks
0:47:39 to Fang,
0:47:39 you know?
0:47:42 And then there was like a pause
0:47:42 and there was like this,
0:47:43 and I think it’s doable.
0:47:46 And so then actually
0:47:46 I went back
0:47:49 and I was like,
0:47:50 am I the crazy one
0:47:51 or are they the crazy one?
0:47:52 Who’s the crazy one here?
0:47:52 Like,
0:47:53 who’s nuts here?
0:47:54 And,
0:47:54 you know,
0:47:56 and that kind of pushed us
0:47:57 to think about
0:47:57 how do we change
0:47:58 our comp philosophy?
0:47:59 How do we,
0:48:00 if we wanted to go
0:48:01 and get the best of the best
0:48:02 out of Google,
0:48:03 what does it require?
0:48:04 And we developed a new model
0:48:05 where we’re like,
0:48:05 hey,
0:48:06 actually the way to think about it
0:48:07 is your market cap
0:48:08 divided by number of employees.
0:48:09 That’s how much money
0:48:09 you can give away
0:48:10 in terms of dilution.
0:48:11 And actually we did
0:48:12 calculate the number
0:48:12 at that time.
0:48:13 And we’re like,
0:48:13 wait,
0:48:14 we’re actually richer
0:48:15 than Google
0:48:15 in terms of,
0:48:16 you know,
0:48:16 how much dilution
0:48:17 we can afford per engineer
0:48:18 because they were,
0:48:18 at that time,
0:48:19 this was like,
0:48:19 you know,
0:48:20 before the Twitter downsizing
0:48:21 so all the companies
0:48:22 were oversized.
0:48:22 So,
0:48:24 so we did the calculation
0:48:25 and it turned out that
0:48:26 we actually can probably
0:48:28 pay P95th percentile.
0:48:28 We did the math
0:48:30 on P95th percentile
0:48:30 for engineering
0:48:31 and it was like,
0:48:31 yeah,
0:48:31 this actually
0:48:32 the math works out.
0:48:33 We moved all the compands
0:48:34 and we told them
0:48:34 and we told the employees
0:48:35 about it.
0:48:35 It’s like,
0:48:35 hey,
0:48:36 we’re paying your P95th,
0:48:36 you know,
0:48:38 and we can afford it.
0:48:40 And so that came out
0:48:40 of that simple,
0:48:41 you know,
0:48:41 so these simple,
0:48:42 you know,
0:48:46 and so on.
0:48:47 They’re silly
0:48:48 and they’re kind of crazy
0:48:49 but they do push you
0:48:50 and you go back
0:48:50 and think about,
0:48:50 hey,
0:48:52 what is the fundamental reason
0:48:53 from first principles
0:48:54 that we couldn’t do
0:48:54 something like that?
0:48:55 Why couldn’t we be a trillion?
0:48:57 What’s the bottleneck
0:48:58 from being a trillion
0:49:00 or being part of FANG?
0:49:01 And then you think about it
0:49:02 and you start zooming in
0:49:02 on like,
0:49:03 can we unblock that?
0:49:05 So it has helped us
0:49:06 and it’s been a driving force
0:49:07 even though it’s,
0:49:07 you know,
0:49:08 it’s a little annoying,
0:49:09 you know,
0:49:10 you know,
0:49:10 it’s like,
0:49:11 you know,
0:49:11 hey,
0:49:12 you know,
0:49:12 hey,
0:49:12 mom,
0:49:13 dad,
0:49:13 I got the A plus.
0:49:14 It’s like,
0:49:14 yeah,
0:49:15 but we ranked,
0:49:16 I was number two
0:49:16 in the class.
0:49:17 So it was someone
0:49:17 better than you?
0:49:19 Yeah.
0:49:20 Yeah,
0:49:20 for what it’s worth,
0:49:21 when I joined the firm
0:49:22 in 2019,
0:49:24 the series F of Databricks
0:49:25 was the first deal
0:49:25 I worked on
0:49:26 and I think
0:49:27 the valuation
0:49:27 was six billion.
0:49:29 Ben says to us,
0:49:29 oh,
0:49:29 well,
0:49:30 it’s going to be
0:49:30 a hundred billion
0:49:31 dollar company.
0:49:32 And we’re like,
0:49:32 yeah,
0:49:33 yeah,
0:49:33 sure,
0:49:33 Ben,
0:49:34 lo and behold.
0:49:35 They’re doing all this work.
0:49:35 I’m like,
0:49:36 what are you doing?
0:49:37 Like six billion
0:49:38 to seven billion
0:49:39 doesn’t matter.
0:49:40 I was right.
0:49:41 It was right.
0:49:41 Yeah,
0:49:42 that’s when you have proven
0:49:43 to be right.
0:49:43 Yeah.
0:49:44 Still have ways
0:49:45 to go for two trillion.
0:49:46 Well,
0:49:49 the thing that you
0:49:50 almost never get
0:49:51 and Ali and I
0:49:52 had this conversation
0:49:52 the one time
0:49:54 we did have
0:49:55 a real acquisition
0:49:56 offer on the company
0:49:58 is you just
0:49:58 don’t get
0:50:00 this good
0:50:01 a market opportunity
0:50:02 with this good
0:50:03 an entrepreneur.
0:50:04 Like that’s the rarest
0:50:05 of rare things.
0:50:06 Like we see
0:50:07 great entrepreneurs
0:50:08 but their market
0:50:09 opportunity is limited
0:50:10 and then we see,
0:50:11 you know,
0:50:12 companies that have
0:50:14 a great market opportunity
0:50:15 but the entrepreneur
0:50:16 is not big enough
0:50:17 to fulfill that.
0:50:19 but this was
0:50:20 a case where
0:50:20 we have both.
0:50:22 Yeah,
0:50:22 I remember actually
0:50:23 the conversation
0:50:24 that kind of flipped me.
0:50:26 The acquisition offer
0:50:27 was on the table.
0:50:28 It was six times bigger
0:50:29 than the valuation
0:50:30 we had at the time
0:50:32 and I had done
0:50:32 the mistake
0:50:33 of telling my co-founders
0:50:34 and
0:50:35 they were like,
0:50:36 let’s go.
0:50:37 They were like,
0:50:37 we’re done.
0:50:38 So everyone’s like,
0:50:39 stop the work,
0:50:40 stop working,
0:50:41 take your hands
0:50:41 off the keyboard,
0:50:42 nobody work anymore,
0:50:43 we’re done here,
0:50:43 right?
0:50:45 And let’s count my money.
0:50:45 You’re like,
0:50:45 you know,
0:50:46 how much money do I have?
0:50:47 You know,
0:50:47 what would you buy
0:50:48 for that amount of money?
0:50:49 You know,
0:50:49 so they were like
0:50:50 completely like
0:50:50 not doing anything
0:50:51 and there was just
0:50:52 this crazy gossip
0:50:53 going around
0:50:53 and then it was like
0:50:54 they had told
0:50:54 some of the exec
0:50:55 and then they were
0:50:55 calling each other
0:50:56 every day like,
0:50:56 hey,
0:50:57 what do Ali think?
0:50:57 Like,
0:50:57 you know,
0:50:59 he looked in a bad mood
0:50:59 today,
0:50:59 do you think he’s
0:51:00 going to say no?
0:51:00 No,
0:51:00 it’s like,
0:51:01 what did he say?
0:51:02 Like he said this thing,
0:51:02 he said this one time.
0:51:03 So there was just
0:51:04 a lot of politicking
0:51:04 going around
0:51:05 and nobody was doing
0:51:14 and,
0:51:16 you know,
0:51:17 he says that he drops
0:51:18 the F-bombs
0:51:19 and he pisses people off
0:51:20 and so on
0:51:20 and they don’t take
0:51:20 the feedback
0:51:21 but actually he did
0:51:22 exactly the radical
0:51:23 candor thing with me
0:51:24 which is he said,
0:51:24 hey,
0:51:25 you can do whatever you want.
0:51:27 I’ll support you
0:51:27 either case
0:51:29 and actually if you sell
0:51:29 for this number
0:51:30 it’s really great for me,
0:51:31 me being done.
0:51:32 Like,
0:51:33 we make a lot of money
0:51:33 at A6 and Z,
0:51:34 I’ll pay the investors
0:51:35 back many times over
0:51:36 so honestly
0:51:37 if it’s for me personally
0:51:38 that’s probably
0:51:38 the better option.
0:51:41 But I’m just thinking back,
0:51:41 I was CEO
0:51:43 of Cloud Cloud Opsware
0:51:44 and,
0:51:44 you know,
0:51:45 just the cards
0:51:46 that were given
0:51:47 that company
0:51:48 wasn’t the company
0:51:48 you have.
0:51:50 And when I look back,
0:51:51 how often do you
0:51:52 in life get a chance
0:51:53 to even have a company
0:51:54 like Cloud Cloud
0:51:54 or Opsware
0:51:56 let alone a Databricks
0:51:57 and it’s just such
0:51:59 a freaking big market.
0:52:00 You can sell,
0:52:00 you’re going to make
0:52:01 a lot of money
0:52:02 and you’ll be
0:52:03 super successful in life
0:52:04 but,
0:52:04 you know,
0:52:05 if you’re like me,
0:52:06 you’re going to look back
0:52:07 the rest of your life
0:52:08 thinking,
0:52:09 you know,
0:52:10 I missed that one shot.
0:52:11 That was the one thing.
0:52:11 I should have taken it
0:52:12 all the way
0:52:13 and now I’ll never know
0:52:13 how far I could have
0:52:14 taken it.
0:52:14 What could have been.
0:52:15 So do you want to live
0:52:15 with that
0:52:16 or do you want
0:52:17 to just have the money?
0:52:17 You know,
0:52:18 I’ll support whatever
0:52:18 you want to do.
0:52:20 I really couldn’t care less.
0:52:21 I really couldn’t care less.
0:52:22 And I was like,
0:52:22 okay,
0:52:22 thanks,
0:52:23 hang up.
0:52:24 We’re never doing this.
0:52:25 We’re done.
0:52:26 This is not happening.
0:52:27 What a pep talk.
0:52:29 So that’s how we did it.
0:52:30 So it was excellent.
0:52:31 I think I also said,
0:52:33 I guarantee you,
0:52:34 you’ll never have an idea
0:52:35 this good again
0:52:36 as long as you live.
0:52:39 This is the best idea
0:52:39 you’re ever going to have.
0:52:40 Yeah,
0:52:40 yeah,
0:52:41 yeah.
0:52:42 Well,
0:52:43 an idea that also takes off
0:52:43 and works,
0:52:44 right?
0:52:44 Yeah.
0:52:46 So I want to tie
0:52:47 one thing that you said
0:52:47 in all of that
0:52:49 is you were company building
0:52:50 but then also
0:52:52 just sort of the calculus
0:52:53 that founders
0:52:54 but also your employees
0:52:54 are making
0:52:55 and that’s around comp.
0:52:57 So in the early days
0:52:57 you could afford
0:52:59 to pay 95th percentile,
0:52:59 right?
0:53:01 Today,
0:53:04 there’s crazy AI talent wars
0:53:04 going on.
0:53:05 We’ve talked about this
0:53:06 a bunch this summer
0:53:08 and we know
0:53:09 that you can bring
0:53:09 the best talent
0:53:10 in the house,
0:53:11 right,
0:53:11 to Databricks.
0:53:13 How do you keep them
0:53:14 with all of this
0:53:15 craziness going on?
0:53:16 Because now 95th percentile,
0:53:17 I don’t even know
0:53:17 what that means.
0:53:18 Is that like you pay
0:53:19 a billion dollars?
0:53:20 Yeah,
0:53:20 exactly,
0:53:20 exactly.
0:53:21 Yeah,
0:53:21 the joke is
0:53:22 which company says
0:53:23 we’re P50 yet?
0:53:25 We pay P50 yet.
0:53:25 Like who does that?
0:53:26 There’s no company
0:53:26 that does that.
0:53:28 So how does this action?
0:53:28 Yeah,
0:53:29 the 75th percentile
0:53:31 is the single biggest lie
0:53:32 in Silicon Valley.
0:53:33 Probably.
0:53:34 It’s like a complete fabrication.
0:53:35 Probably,
0:53:35 probably.
0:53:35 So,
0:53:36 but,
0:53:38 you know,
0:53:38 I think that
0:53:41 it is a crazy time
0:53:41 with AI
0:53:42 and I do think,
0:53:43 I feel bad
0:53:43 I did actually
0:53:44 an exit interview
0:53:45 with someone this morning.
0:53:46 I feel bad for the kids
0:53:47 right now
0:53:47 because there’s like
0:53:48 too much pressure on them.
0:53:49 Like they feel like
0:53:49 oh,
0:53:50 they have to start companies
0:53:50 and they have to
0:53:51 and I’ve never actually
0:53:52 had anything like this
0:53:52 because every year
0:53:53 I talk to the interns
0:53:55 and,
0:53:56 you know,
0:53:56 I get questions about
0:53:57 how do we build
0:53:57 our own company,
0:53:57 you know,
0:53:58 how do we succeed
0:53:58 at Databricks
0:53:59 and so on.
0:54:00 The last two years
0:54:01 have just been crazy.
0:54:03 All the kids are like,
0:54:04 when should I become a CEO?
0:54:05 When should I start
0:54:06 my own company?
0:54:07 What’s a good valuation?
0:54:08 Am I missing out?
0:54:08 If I do like
0:54:09 an internship here
0:54:10 for three months
0:54:10 at Databricks,
0:54:12 will I have wasted
0:54:13 my opportunity in life
0:54:13 and this is like
0:54:14 the time for AGI
0:54:15 and I could have like
0:54:16 been one of the guys
0:54:17 that’s super intelligence
0:54:17 and like,
0:54:18 you know,
0:54:19 how would you time that?
0:54:19 How was it for you?
0:54:20 How old were you?
0:54:21 When you were 22,
0:54:21 what did you do?
0:54:22 And,
0:54:22 you know,
0:54:23 and so,
0:54:25 I do think it’s
0:54:26 kind of crazy times
0:54:27 and I do think
0:54:28 it’s also exaggerated.
0:54:29 Like,
0:54:29 you know,
0:54:29 I don’t think anyone’s
0:54:31 getting $100 million offers.
0:54:31 You know,
0:54:32 I mean,
0:54:32 yes,
0:54:33 there’s like one,
0:54:34 you know,
0:54:34 character that AI
0:54:35 and so on,
0:54:35 but I don’t think
0:54:36 it’s actually true
0:54:37 and it’s also
0:54:39 in the interest of CEOs,
0:54:40 you should know,
0:54:41 to say that,
0:54:41 hey,
0:54:42 you know,
0:54:44 people try to poach
0:54:45 people from Databricks
0:54:45 for a billion dollars
0:54:46 and they said no.
0:54:47 It’s in our interest
0:54:47 to say that,
0:54:47 right?
0:54:48 Because that kind of
0:54:48 sets the bar
0:54:49 at the billion
0:54:50 and then any employee
0:54:51 that gets an offer
0:54:51 for half of that
0:54:52 is going to feel
0:54:52 really insulted.
0:54:53 It’s like,
0:54:53 when did I get
0:54:54 a billion dollar offer?
0:54:56 I heard like on the news
0:54:56 that the other people
0:54:57 are getting a billion.
0:54:58 So,
0:54:59 I do think that the most-
0:55:00 By the way,
0:55:00 Sam used that
0:55:01 in reverse on Meta.
0:55:02 He’s like,
0:55:02 oh yeah,
0:55:06 into the next guy
0:55:07 who got the $50 million.
0:55:08 Now they have to pay $100
0:55:08 at least,
0:55:09 right?
0:55:10 That’s like the smart move.
0:55:11 But I would say that,
0:55:11 look,
0:55:13 you know,
0:55:14 and not all startups
0:55:15 have the valuation of theirs.
0:55:16 We’re worth $100 billion
0:55:17 and,
0:55:17 you know,
0:55:18 with 10,000 employees,
0:55:19 we actually can’t afford
0:55:21 to actually pay significant
0:55:22 and we do pay significant
0:55:23 for the right talent.
0:55:24 But,
0:55:24 you know,
0:55:25 what did you do
0:55:25 when you’re smaller?
0:55:26 Like,
0:55:26 we were smaller
0:55:27 at some point.
0:55:27 Well,
0:55:28 then it’s talk about
0:55:29 how big you are going to get
0:55:30 and what the opportunity is
0:55:31 and what you could do together
0:55:32 and what it would work
0:55:33 together.
0:55:34 But I think
0:55:36 most people earlier
0:55:36 in their career,
0:55:37 they really want to learn
0:55:38 and they want to really feel
0:55:39 that they can have impact.
0:55:40 So,
0:55:42 if you can really bring them in
0:55:43 and you can sort of mentor them,
0:55:44 you can stay close to them
0:55:44 and as a CEO,
0:55:46 you have huge power.
0:55:46 Like,
0:55:47 if you could just spend
0:55:47 two minutes with,
0:55:48 you know,
0:55:49 a kid out of school,
0:55:51 it’s immense to them.
0:55:51 And you say,
0:55:51 hey,
0:55:52 you know,
0:55:53 I’ll even mentor you.
0:55:53 I’ll help you.
0:55:53 Like,
0:55:54 what do you want to do
0:55:54 in five years?
0:55:55 I’m thinking about
0:55:56 starting my own company
0:55:57 actually in six months,
0:55:58 you know,
0:55:58 but I’ll work at Databricks
0:55:59 for 10 years,
0:56:00 but in six months
0:56:00 I would love to be a CEO.
0:56:01 So then you can say,
0:56:01 hey,
0:56:02 I can coach you to,
0:56:03 you know,
0:56:04 I know how the fundraising,
0:56:05 I know the early days
0:56:06 and so on.
0:56:06 And you can actually mentor
0:56:07 a lot of them
0:56:07 and that’s actually
0:56:08 worked a lot to them as well.
0:56:10 But in general,
0:56:11 like help them be successful
0:56:12 and help them build their careers.
0:56:13 And also,
0:56:13 if you’ve done it before,
0:56:14 like we have,
0:56:16 you can kind of calm them down
0:56:17 a little bit
0:56:18 and say,
0:56:18 hey,
0:56:19 you have like a few decades,
0:56:20 you know,
0:56:20 don’t worry about it.
0:56:21 it’s like,
0:56:21 it’s not,
0:56:22 you know,
0:56:24 it’s the FOMO
0:56:25 and the pressure,
0:56:25 you know,
0:56:27 has to be kind of reduced.
0:56:28 And I think that’s also calming
0:56:29 that they feel good about it.
0:56:29 Yeah.
0:56:30 Yeah.
0:56:32 I always say
0:56:33 the best cure
0:56:35 for starting your own company fever
0:56:37 is to start your own company
0:56:38 and that’ll teach you.
0:56:40 It’s not that easy.
0:56:41 By the way,
0:56:41 they come back,
0:56:42 like after start companies,
0:56:43 oftentimes they come back
0:56:44 to Databricks
0:56:45 and they’re much more thankful
0:56:45 and they understand.
0:56:46 And actually,
0:56:47 I didn’t mention this earlier
0:56:49 when you asked about acquisitions.
0:56:50 My favorite acquisition,
0:56:51 because I said
0:56:52 you start with the people,
0:56:52 right?
0:56:53 And then the product.
0:56:55 With the people,
0:56:56 I love to hire people
0:56:57 who have seen great
0:56:57 at a big company,
0:56:58 like,
0:56:59 or I don’t know if it’s great,
0:57:01 but they’ve seen process scale
0:57:02 big company.
0:57:03 They’ve been at a Google,
0:57:03 they’ve been at Amazon,
0:57:05 they understand the processes
0:57:06 so they understand
0:57:07 how to navigate a bureaucracy
0:57:08 and work with it
0:57:09 and they’re not going to just
0:57:10 be inundated by it.
0:57:11 But then they’ve gone on
0:57:12 and done their own startup
0:57:15 and that’s really,
0:57:15 really hard,
0:57:16 right?
0:57:16 It’s like,
0:57:17 it’s like extremely hard
0:57:19 trying to do everything yourself
0:57:20 and you don’t have any help
0:57:20 and you know,
0:57:21 you’re trying to do this
0:57:22 in this crazy market
0:57:23 and you’re trying to compete
0:57:24 with $100 million offers
0:57:25 when you have like nothing.
0:57:27 So that takes a certain amount
0:57:28 of grit
0:57:29 and it’s really humbling.
0:57:31 So I love the people
0:57:32 that have done both of those.
0:57:33 They end up being actually
0:57:33 the perfect employees
0:57:34 at Databricks
0:57:34 because they come in
0:57:35 and they’re really thankful.
0:57:36 They’re like,
0:57:36 hey,
0:57:37 what these guys have done
0:57:37 at Databricks
0:57:37 is actually really,
0:57:38 really hard.
0:57:39 I tried it
0:57:40 and I’m really good.
0:57:40 I was like one of the best
0:57:41 at Google or somewhere
0:57:42 and then I did my own startup
0:57:44 and we absolutely failed.
0:57:45 And so,
0:57:45 hey,
0:57:46 show some respect here.
0:57:46 Like,
0:57:46 you know,
0:57:47 these guys know
0:57:47 what they’re talking about.
0:57:49 So those are great employees
0:57:49 actually.
0:57:49 But,
0:57:50 you know,
0:57:52 I think keep a great relationship
0:57:53 with people who leave your company
0:57:54 because they can boomerang
0:57:54 back in a couple years.
0:57:55 Yeah,
0:57:57 and it’s very hard
0:57:58 to make these things work.
0:58:02 And it also requires
0:58:02 a lot of luck.
0:58:03 I mean,
0:58:04 I think one of the things
0:58:06 people don’t realize
0:58:08 is a lot of things
0:58:09 have to go right
0:58:10 that should never go right.
0:58:11 And a lot of things
0:58:12 will go wrong,
0:58:14 but like if you can grab
0:58:15 your lucky moments,
0:58:16 that’s a rare,
0:58:17 that’s a rare thing.
0:58:18 Yeah,
0:58:19 one way to prove that
0:58:19 is if Databricks
0:58:21 started in 2013,
0:58:23 if we had started in 2012,
0:58:24 you know,
0:58:24 that rocky year,
0:58:25 that difficult year,
0:58:26 2015,
0:58:26 would have then happened
0:58:27 in 2014,
0:58:28 right,
0:58:29 to start the funds
0:58:30 or we don’t have the revenue.
0:58:31 But we were a cloud,
0:58:32 AI,
0:58:33 open source company.
0:58:34 Those things
0:58:36 didn’t take off in 2014.
0:58:37 So,
0:58:38 you know,
0:58:38 even if like,
0:58:39 if we had to do the CEO change
0:58:40 and all of that
0:58:41 and I had become CEO
0:58:42 a year earlier,
0:58:43 it’s just,
0:58:44 we were too early in the market.
0:58:45 The cloud hadn’t taken off.
0:58:46 AI was not,
0:58:46 nobody,
0:58:48 that was not even a phrase.
0:58:48 AI meant robotics.
0:58:50 people use machine learning
0:58:50 as a phrase.
0:58:52 And so,
0:58:53 company would have failed.
0:58:53 We wouldn’t have had
0:58:54 enough momentum.
0:58:55 There’s not enough cloud,
0:58:55 you know,
0:58:56 TAM there to be had.
0:58:57 If we started the company
0:58:58 in 2014,
0:58:59 a year later instead,
0:59:00 so a year later
0:59:01 than we actually did,
0:59:03 then we would have had
0:59:03 our difficult year
0:59:05 in 2016.
0:59:06 But by 2016,
0:59:07 the cloud was starting to happen.
0:59:08 AI was starting to happen,
0:59:09 you know.
0:59:10 So we would have done
0:59:12 the fixes in 2017
0:59:12 and it would have been
0:59:14 too late to the party
0:59:15 and probably the hyperscalers
0:59:16 would have taken it away.
0:59:17 You know,
0:59:17 our competitors
0:59:18 were taking it away
0:59:19 and we just wouldn’t have
0:59:21 get enough momentum
0:59:22 to be able to succeed.
0:59:24 And that’s timing
0:59:24 of when we started.
0:59:26 So how did we clock it
0:59:26 so well?
0:59:28 We had to wait for Matei
0:59:29 to finish his PhD thesis.
0:59:29 That’s it.
0:59:32 That was the whole thing.
0:59:33 So there’s a lot
0:59:33 of randomness
0:59:35 and you got to get lucky.
0:59:37 And it was so on the edge
0:59:38 as it was
0:59:40 that on the Series C,
0:59:42 Jan had a handshake
0:59:44 with Redpoint
0:59:44 with Redpoint
0:59:46 and Redpoint
0:59:47 just stopped
0:59:48 returning his calls
0:59:48 to the point
0:59:50 where the Series C
0:59:51 was led by us
0:59:52 who also led
0:59:53 the Series A
0:59:55 and NEA
0:59:56 who led the Series B
0:59:57 like we co-led
0:59:58 the Series C
0:59:59 because nobody else
1:00:00 would do it.
1:00:01 It was that
1:00:03 close to going under.
1:00:05 And most companies
1:00:05 went even,
1:00:06 right,
1:00:07 that would be it.
1:00:07 Yeah,
1:00:08 it was very close
1:00:08 because we couldn’t
1:00:09 get funding for anyone.
1:00:10 It’s just funding
1:00:10 freezed up
1:00:11 and nobody wanted
1:00:11 to invest anymore.
1:00:13 So it was really
1:00:13 a lifeline
1:00:14 from A, 6, and Z.
1:00:15 Yeah,
1:00:15 we were just,
1:00:15 you know,
1:00:16 burning a lot of cash.
1:00:17 We weren’t generating
1:00:18 much revenue
1:00:19 other than Spark Summit.
1:00:20 We had a lot of downloads.
1:00:21 We had a lot of downloads.
1:00:22 A lot of downloads.
1:00:23 And recurring
1:00:24 conference revenue.
1:00:24 Yeah,
1:00:26 and recurring conference.
1:00:27 How confident
1:00:28 were you in that time
1:00:29 when things were out there?
1:00:29 Oh, I mean,
1:00:30 I seriously consider
1:00:31 taking the professor job
1:00:31 at Berkeley
1:00:33 because I,
1:00:33 you know,
1:00:35 I seriously thought
1:00:35 this was going to be
1:00:36 very, very hard
1:00:36 to pull off.
1:00:37 You know,
1:00:38 it’s like,
1:00:38 you know,
1:00:39 I think the sentiment
1:00:40 at Databricks was,
1:00:41 or at least my sentiment
1:00:42 was,
1:00:43 look,
1:00:44 you win some things
1:00:44 and you lose
1:00:45 some things in life.
1:00:46 We created
1:00:47 Apache Spark
1:00:48 and we made it
1:00:48 a worldwide sensation.
1:00:49 Everybody’s downloading.
1:00:50 The downloads
1:00:50 are through the roof.
1:00:51 We have this
1:00:52 greatest conference.
1:00:52 Like, you know,
1:00:53 thousands of people
1:00:54 come to our conference.
1:00:54 It’s awesome.
1:00:55 Let’s go back.
1:00:56 Let’s do it again.
1:00:57 Let’s publish another paper
1:00:58 and do those kind of things.
1:00:59 We’re just not business, guys.
1:01:00 We just don’t understand business.
1:01:01 You know,
1:01:01 that’s okay.
1:01:02 You know,
1:01:03 we don’t want to be business guys.
1:01:04 So,
1:01:05 so that’s kind of
1:01:05 how I felt about it,
1:01:06 right?
1:01:07 But,
1:01:08 what I knew
1:01:08 was that,
1:01:09 hey,
1:01:10 you know.
1:01:11 By the way,
1:01:12 Matei went back
1:01:13 and became a professor.
1:01:14 Like,
1:01:15 all this stuff happened.
1:01:16 Yeah.
1:01:17 Jan went back,
1:01:17 you know.
1:01:18 And,
1:01:19 but,
1:01:20 you know,
1:01:20 in 2015,
1:01:22 we knew that
1:01:23 we tried everything.
1:01:23 And,
1:01:24 by the way,
1:01:25 PLG is something
1:01:25 that we had tried
1:01:26 very hard
1:01:27 and it didn’t work for us.
1:01:28 Like,
1:01:28 and actually,
1:01:29 one of our biggest failures
1:01:30 was PLG at Databricks.
1:01:32 Everybody kept telling us,
1:01:32 PLG,
1:01:32 PLG,
1:01:33 PLG.
1:01:33 And we’re like,
1:01:34 okay,
1:01:34 product-led growth.
1:01:35 And Amazon did it.
1:01:35 And you just swipe
1:01:36 your credit cards.
1:01:37 We don’t need salespeople.
1:01:38 But so,
1:01:39 in 2015,
1:01:40 I had to kind of like,
1:01:40 Except cranny.
1:01:41 Exactly.
1:01:42 Yes,
1:01:42 that is true.
1:01:43 Kudos to Mark.
1:01:44 So,
1:01:44 that year,
1:01:45 we had like formed
1:01:46 some hypothesis that,
1:01:47 you know,
1:01:48 we have nothing to lose.
1:01:50 What if we just pivoted
1:01:50 these things?
1:01:51 What if we went all in
1:01:53 into B2B enterprise sales?
1:01:53 You know,
1:01:54 because certainly it’s not,
1:01:55 PLG is not working.
1:01:56 You know,
1:01:57 at 3 million ARR,
1:01:57 that’s not going to
1:01:58 take you anywhere.
1:01:59 And,
1:02:00 you know,
1:02:01 and they’re just taking
1:02:01 our open source software.
1:02:01 So,
1:02:02 we have to have
1:02:03 proprietary code around it.
1:02:04 You know,
1:02:04 and yeah,
1:02:05 the execs team
1:02:07 are all PhDs.
1:02:07 You know,
1:02:08 so what if we bring in
1:02:08 someone that doesn’t
1:02:09 have a PhD
1:02:10 and see how it goes?
1:02:12 I never forget
1:02:13 Arsalan going,
1:02:14 we made the number.
1:02:15 I was like,
1:02:17 you made a ridiculous number.
1:02:19 You made the number.
1:02:19 Like,
1:02:20 if you keep making that number,
1:02:21 you’re going to go bankrupt.
1:02:23 You didn’t make the number.
1:02:24 You made a number
1:02:24 that you set
1:02:25 that was way too low.
1:02:28 Ben was very nice
1:02:28 and complimentary
1:02:29 in our board meetings
1:02:30 that year,
1:02:30 2015.
1:02:33 We were in a bit of trouble.
1:02:34 Let’s say it was
1:02:35 very truth-seeking.
1:02:36 But,
1:02:36 yes,
1:02:37 we had nothing to lose.
1:02:38 So,
1:02:38 we didn’t know
1:02:39 that we were going to succeed,
1:02:40 but we had nothing to lose
1:02:41 to make those big changes
1:02:42 and we made them in 2016.
1:02:43 And it turned out
1:02:44 those were the bottlenecks,
1:02:45 you know,
1:02:46 giving away your software
1:02:46 for free.
1:02:48 Not having execs
1:02:49 that have seen the movie before,
1:02:50 like Ron who came in.
1:02:51 And,
1:02:52 you know,
1:02:54 just the PLG motion
1:02:56 is not going to cut it.
1:02:56 So,
1:02:57 maybe we should just try.
1:02:58 We weren’t certain
1:02:59 that B2B would work,
1:03:00 but we knew that
1:03:01 PLG is not working for sure.
1:03:02 Yeah,
1:03:02 well,
1:03:03 another,
1:03:03 like,
1:03:03 I mean,
1:03:03 you know,
1:03:05 we got Ron.
1:03:09 I mean,
1:03:10 like,
1:03:11 the fact that
1:03:11 the first sales guy
1:03:12 we hired
1:03:13 was
1:03:14 a
1:03:15 sales
1:03:16 savant
1:03:17 like a genius
1:03:18 shock.
1:03:19 I mean,
1:03:19 like,
1:03:20 that never happens.
1:03:20 And,
1:03:22 and he was a guy
1:03:22 we didn’t know.
1:03:23 Like,
1:03:24 we,
1:03:24 like,
1:03:25 our talent team,
1:03:25 like,
1:03:26 found him
1:03:27 from some company
1:03:28 we never heard of.
1:03:29 Yeah,
1:03:29 a French company,
1:03:30 Aksuay.
1:03:31 And the only,
1:03:32 really,
1:03:33 the only reason
1:03:34 we hired him
1:03:34 was because
1:03:35 he was the only guy
1:03:36 Cranny ever liked.
1:03:38 Like,
1:03:39 in all the sales guys
1:03:39 he ever interviewed,
1:03:40 he was like,
1:03:40 this is the guy.
1:03:41 Wow.
1:03:42 And,
1:03:42 yeah,
1:03:44 he is a new generation
1:03:45 Mark Cranny T2.
1:03:46 But we just,
1:03:46 like,
1:03:47 stumbled into him.
1:03:48 Yeah,
1:03:48 yeah,
1:03:48 like,
1:03:49 unbelievable.
1:03:49 And,
1:03:51 without Ron,
1:03:54 very hard to see
1:03:54 this company
1:03:55 getting to where it got to.
1:03:57 so there’s some luck
1:03:57 involved in us
1:03:58 even finding him.
1:03:59 But then,
1:04:00 he was phenomenal.
1:04:00 And also,
1:04:01 kudos to Jan,
1:04:02 who actually led
1:04:03 the search in 2015.
1:04:04 Yeah,
1:04:04 yeah.
1:04:04 So,
1:04:05 you know,
1:04:06 but,
1:04:07 yeah,
1:04:08 Ron was game-changing
1:04:08 for us.
1:04:09 but Ron was a very
1:04:11 uncomfortable hire
1:04:12 because he did not
1:04:13 have a PhD.
1:04:15 And,
1:04:16 he did have an
1:04:16 engineering degree.
1:04:17 Yeah,
1:04:17 he does have an
1:04:18 engineering degree
1:04:18 from Stanford
1:04:19 that helped a little bit.
1:04:20 But,
1:04:21 but he’s a sales,
1:04:22 you know,
1:04:23 true and true sales guy.
1:04:23 Like,
1:04:24 he’s not a,
1:04:25 you know,
1:04:26 he’s not one of these,
1:04:27 he’s a classic salesperson,
1:04:28 like,
1:04:29 who grew up in sales,
1:04:30 even though he has
1:04:30 an engineering degree.
1:04:31 So,
1:04:32 the comfortable thing
1:04:32 would have been
1:04:33 to pick someone,
1:04:34 and we had some candidates
1:04:35 in the mix,
1:04:37 who were just super technical,
1:04:38 using the product,
1:04:39 giving us feedback,
1:04:40 but they probably would not.
1:04:41 And that would have been
1:04:42 much more comfortable for us.
1:04:42 Yeah.
1:04:43 Yeah,
1:04:44 Ron was uncomfortable.
1:04:45 He was a very uncomfortable hire,
1:04:47 and he made it very uncomfortable
1:04:48 for us for many years,
1:04:49 and he still does.
1:04:52 But that’s a lot of the key
1:04:52 to the company.
1:04:53 Yeah.
1:04:54 Is it forces,
1:04:58 it forces a customer focus
1:04:59 that would be impossible
1:04:59 to have
1:05:00 without
1:05:01 somebody
1:05:03 that
1:05:04 smart
1:05:05 and crafty
1:05:07 about getting his way.
1:05:08 I mean,
1:05:08 just,
1:05:08 like,
1:05:09 unbelievable.
1:05:09 Yeah.
1:05:11 If you can keep also
1:05:11 the original team together,
1:05:12 that’s important.
1:05:12 We’re,
1:05:13 you know,
1:05:14 seven co-founders still.
1:05:15 Many of the co-founders,
1:05:15 like,
1:05:15 you know,
1:05:16 you said data warehousing
1:05:17 was a big push for us.
1:05:18 My co-founder,
1:05:18 Reynolds,
1:05:19 was really the one
1:05:20 that kind of pushed this.
1:05:20 Yeah,
1:05:20 like,
1:05:21 well,
1:05:23 the contribution level
1:05:25 from a large number
1:05:26 of co-founders
1:05:27 is unique in the industry.
1:05:28 I mean,
1:05:29 you’ve got Patrick,
1:05:30 you have Reynolds,
1:05:31 you have Matei,
1:05:32 you have Arsalan.
1:05:33 I mean,
1:05:33 like,
1:05:34 it’s crazy
1:05:35 how much
1:05:37 the original team
1:05:38 contributes.
1:05:39 Yeah,
1:05:40 so the PhDs all contribute.
1:05:40 Like,
1:05:41 Arsalan really made
1:05:42 the go-to-market work,
1:05:42 you know,
1:05:43 and he really made
1:05:44 sort of Ron work
1:05:45 with the rest of the company
1:05:46 that was super critical.
1:05:47 Matei continued doing
1:05:47 lots of innovations
1:05:48 over the years.
1:05:49 Patrick led all of engineering
1:05:50 in big chunks of it
1:05:51 and,
1:05:51 you know,
1:05:52 so on.
1:05:52 And we’ve had other people
1:05:53 who’ve been lucky
1:05:54 to get such folks.
1:05:55 So hiring is critical
1:05:56 and keeping the original talent,
1:05:57 I think.
1:05:58 Those were some of the things.
1:05:59 Usually founders,
1:06:02 usually only one
1:06:02 of the co-founders
1:06:04 contributes long-term.
1:06:05 And so to have,
1:06:07 you know,
1:06:08 to have that going
1:06:09 and Jan’s still on the board
1:06:10 and Scott’s still on the board,
1:06:10 I mean,
1:06:11 like,
1:06:12 it’s very unusual.
1:06:14 We have a lot more
1:06:15 we can get into,
1:06:16 but we’re out of time,
1:06:16 so we’ll leave it
1:06:17 for future episodes
1:06:18 of Boss Talk.
1:06:19 This is a great episode.
1:06:20 All right.
1:06:20 That was fun.
1:06:21 Thanks so much.
1:06:22 Thank you so much, guys.
1:06:25 Thanks for listening
1:06:26 to this episode
1:06:27 of the A16Z podcast.
1:06:29 If you liked this episode,
1:06:30 be sure to like,
1:06:31 comment,
1:06:31 subscribe,
1:06:32 leave us a rating
1:06:33 or review,
1:06:34 and share it
1:06:34 with your friends
1:06:35 and family.
1:06:36 For more episodes,
1:06:37 go to YouTube,
1:06:38 Apple Podcasts,
1:06:39 and Spotify.
1:06:40 Follow us on X
1:06:41 at A16Z
1:06:42 and subscribe
1:06:43 to our Substack
1:06:45 at a16z.substack.com.
1:06:46 Thanks again for listening
1:06:47 and I’ll see you
1:06:48 in the next episode.
1:06:50 As a reminder,
1:06:51 the content here
1:06:51 is for informational
1:06:52 purposes only,
1:06:53 should not be taken
1:06:54 as legal business,
1:06:55 tax,
1:06:56 or investment advice,
1:06:57 or be used to evaluate
1:06:58 any investment or security
1:06:59 and is not directed
1:07:00 at any investors
1:07:01 or potential investors
1:07:03 in any A16Z fund.
1:07:04 Please note that
1:07:05 A16Z and its affiliates
1:07:06 may also maintain
1:07:07 investments in the companies
1:07:08 discussed in this podcast.
1:07:09 For more details,
1:07:10 including a link
1:07:11 to our investments,
1:07:12 please see
1:07:13 a16z.com
1:07:14 forward slash
1:07:15 disclosures.