AI, Data, Culture: In Conversation with Ben Horowitz, GP, a16z

Photo Credit: Jack Cohen.  Bottom left picture:  Amr Awadallah (VP Developer Relations, Google Cloud), Matt Turck (FirstMark), Ben Horowitz (a16z), Michael James ( Founder and Chief Software Architect at Cerebras Systems)

Ben Horowitz resoundingly falls in the category of “needing no introduction”: a highly successful entrepreneur who navigated a perilous situation with his business (Loudcloud, which became Opsware) to a $1.65B acquisition by HP, he’s also the founder of premier Silicon Valley venture capital firm Andreessen Horowitz (aka “a16z”), and the best selling author of two books: “The Hard Thing About Hard Things” and the newly-released “What You Do Is Who You Are”.

It was a special treat to host Ben for a fireside chat at the most recent most recent edition of Data Driven NYC – a great evening that included two other terrific speakers: Amr Adwallah, now VP of Developer Relations at Google Cloud, and previously co-founder and CTO at Cloudera (NYSE: CLDR) and Michael James, co-founder of AI chip Cerebras.

We spent a good hour with Ben and covered a bunch of topics, loosely organized in two parts, first AI and data, and then culture an his new book.

Below are two videos covering each part, as well as a FULL TRANSCRIPT for anyone who prefers to read.

Part I: Ben on AI & Data (investments, opportunities, and how a technical founder can become a great CEO)

Part II: Ben on culture and his new book, “What you do is who you are”

The FULL TRANSCRIPT of both parts is below, lightly edited for clarity and brievity (many thanks to FirstMark’s Karissa Domondon for her fantastic help):


Matt: You guys at a16z are very active in the AI and data space. I was just reviewing on the trusted Crunchbase (Ben: Crunchbase is a little inaccurate but they are very good at SEO! Matt: yes that’s half of the battle!]… Just in the last five or six months, between new investments and follow-ons, it’s a long list: Labelbox, Cresta, ActionIQ (which is a company that we have the honor of sharing with you guys, a great company here in New York), Imply (which spoke at this event), Neural Magic, Databricks (which you’re on the board of), Instabase (spoke here as well), Sisu, Preset, Fivetran, Shield AI…  So a long list and those are just investments that were announced in the last 4-5 months. Maybe to start at that high level, walk us through your investment thesis around AI and why you’re so excited about the space? 

Ben: We kind of think of AI as a very powerful new tool on how to work with computers.  A good analogy would be if you think of conventional programming like deductive reasoning and then AI is like inductive reasoning.  So if you think about doing math and you only had deductive logic, there’s a whole class of problems you could never solve and that’s definitely the case with AI.  And it’s on the order of importance as deductive math as important as conventional programming and so of course the first things that are getting solved with AI end up being things that you couldn’t solve with conventional programming, or very easily, or at all.  But over time what we’re finding is that some things are just easier to do in AI than conventional programming so it’s a very very big space and there’s a whole giant set of new problems that can be solved and then a whole (as you can see in our investments), a rather robust tool chain that needs to be developed so that you can work in AI much more easily and effectively than you can today, which is still kind of early days.  

 The first things that are getting solved with AI end up being things that you couldn’t solve with conventional programming, or very easily, or at all.  But over time what we’re finding is that some things are just easier to do in AI than conventional programming, so it’s a very, very big space.

I heard your partner Marc Andreessen talk about this topic and he was contrasting the question of “is AI a feature?” with “is AI a new architecture?”.  You guys presumably fall in the “AI as a new architecture” camp? 

Yes as I said it’s sort of a whole new problem solving technique.  It’s in the same domain as programming, but it takes a different, probably/approximately correct approach, as opposed to a kind of step by step approach.  It just turns out that there are many things that we can do now that were very very difficult to do before, and we had the amazing breakthroughs in image recognition starting around 2012 and we’re hitting a new point with natural language processing now that’s very similar.  There’s no way to get the kind of results we’re getting with conventional programming, so it’s a big, big space. That doesn’t mean that there won’t be AI features that will be, but it’s certainly bigger than that. 

Any thoughts on where the specific opportunities are? Are you more investors in the tool chain as you mentioned or data/AI driven applications, or does it not matter? 

Anything with high impact, I think, we’re very interested in.  I think there’s a lot of very interesting areas in [AI] infrastructure now, because the things built with the original tools are just not going to be as good as the things built with the next-generation tools and so building the next generation tools is quite exciting.  Probably one of the best companies that I’m involved in is a company called Databricks, which is a sort of foundational component of AI. And things get much much easier when you have tools like that. And of course there’s workflow tools and labelling tools and scaling tools and on and on and on. 

I think there’s a lot of very interesting areas in [AI] infrastructure now, because the things built with the original tools are just not going to be as good as the things built with the next-generation tools and so building the next generation tools is quite exciting. 

Speaking of Databricks, maybe walk us through the journey of the company.  You know this is a room full of entrepreneurs starting new companies. We had Ion at this event years ago, but maybe walk us through – some of the people are technical people here, some less technical – through what it is, what the investment thesis was and how it’s evolved? 

It’s quite an amazing story.  So originally I got a call from Scott Shenker, a professor at Berkeley, and he said “Ben we have a guy here who’s the best distributed systems guy we’ve seen in academia in the last decade”.  That was Matei who is kind of CTO there and so I was like okay, I’m interested, that was a good pitch. And they had built this thing called Spark and the simple way to think of Spark is – you know Hadoop was always a good idea but a really horrible piece of software for anyone who tried to use it, it was hard to program it, it was hard to deploy, it was just a pain in the ass – and Spark was like the good version built by somebody who was, like, a better architect.  And that was Matei. That was what it was to begin with and they just have the open source Spark and they were figuring out what they wanted to do.  

Ion’s idea was okay we’ll make this a SaaS offering and we won’t even mess with the kind of open distribution, we won’t do anything on-prem.  Which in 2013 was kind of a big deal in that it was thought then that well your data is on-premise and so you need your big data tool to be on-premise and that’s just the way it’s going to go.  But by going to the cloud early it made them much more resilient to things like Amazon offering Spark in the cloud and that kind of thing. Because the Databricks solution, I’ll just give you one example – with Databricks, compute and storage were completely independent of one another so you can have lots of compute going against a little bit of storage and vice versa.  You could have the storage on-premise and the compute on the cloud, whereas Amazon’s solution was literally one-to-one. So even if Amazon gave away Spark in the cloud free it was going to cost you more money than Databricks because you were going to overpay for storage or overpay for compute. Just that little thing was a big deal in getting them off the ground.  

[Ali Ghodsi became CEO of Databricks in January 2016] Ali turned out to be probably the best CEO in the entire Andreessen Horowitz portfolio.  This guy is phenomenal. I’ve never seen anything quite like him. He’s built an amazing sales force. He cut a deal with Microsoft, which Microsoft came and told me is the best deal they’ve ever given a third party in the history of Microsoft. That good. And he’s just this PhD in Computer Science. 

As a tangent, why did he become such a great CEO from this technical background? 

He is a natural leader and just very good at confronting all the issues all the time. [Edited]. When he hired the head of sales we brought him a very strong, capable head of sales – he kind of had that moment, he was learning everything that they were doing in sales and he called me up one day, and said “Ben, I’m a little worried about Ron and his team”, and I said why are you worried about them?  He says, “well they’re making products that we don’t have”, and I was like what do you mean? He said, “well, they just sold this customer a training package and a consulting package that we don’t have”, and I said, why do they do that? And he said “well Ron said our product cost $250,000 and the customer had $450,000 and he needed to get the rest of the money”. I’m like, you know, that’s good. But he would come up that learning, that’s enterprise software pricing – “how much does it cost? Well how much have you got?”.  

He learned everything very fast.  His uncle was a founder of OPEC and when the Ayatollah Khomeini moved in and the shah got ousted, his family was basically targeted for execution and so he became a refugee and moved to Sweden.  That whole upbringing, growing up – it’s funny because I’m friends with Quincy Jones’s son, QDIII, who also grew up in a pretty bad neighborhood in Sweden and they grew up in the same neighborhood so it’s just a weird connection but I knew exactly what that was and he came through that and got a PhD in Computer Science and became a great CEO.  I think a lot of what makes you a great CEO is you understand people. Because he had such an unusual life he ended up meeting a lot of different kinds of people in his journey. He knows what motivates people, what moves them, what you can do culturally. Very interesting guy. 

Interesting – to this PhD in Computer Science and AI topic, a lot of the companies that [a16z] hasinvested in have those very tight connections to academia.  I’m thinking of Ion again at Anyscale and the CEO of Sisu at Stanford and Sebastian who’s also at Stanford and a co-founder of Vectra.  Do you think that’s really important, that’s a big differentiating factor to have that DNA of pure research, for AI? 

It depends on the type of AI company that you’re building, but I think that if it’s not an AI application or AI ingredient, because the field is progressing so fast it’s pretty hard to build a winning company if you’re not deep in the research.  If you’re not fully up to speed in everything that’s going on, and there’s a lot going on in AI right now, like if you don’t know the latest in reinforcement learning and GANs and all that, it’s pretty hard to know where the breakthrough is going to come from.  So if you’re building that kind of company, you either need to be a PhD from Stanford or University of Toronto, somewhere that does it, or you have to have the equivalent. You don’t necessarily need to be in academia to be a breakthrough researcher but that tends to be where a lot of them come from.  

Again in your list of investments, a lot of companies are based on open source.  So far Databricks, Ray for Anyscale and so on and so forth, do you think that, to the extent that there is a way, that’s the way of building a successful enterprise software company these days? 

Well I think that if the company has developers as the primary user then that’s pretty hard to do if you’re not open source.  And look, it’s just a matter of fact, for a developer product to work you need developer momentum and I don’t know how in 2020 you get developer momentum for a new architecture or a new API without it being open source.  It just would get rejected very quickly. Now that does mean you have to have a business model. SaaS turns out to probably be the best business model for monetizing open source and it doesn’t mean you can’t ever do anything proprietary but it’s pretty hard to get developer momentum on a piece of proprietary software.  I can’t think of one that’s done it lately. 


Matt: Let me start with saying that I’m a huge fan of the book, which I’ve read probably a couple of times at this stage. To start with a level set, what is culture, and how is it different to values and virtues? 

Ben: I always like to start with where the ancient samurai started,  which is: culture is not a set of beliefs, it’s a set of actions. 

It’s why I called the book What You Do Is Who You Are, because what you believe is not who you are, what’s in your heart is not who you are, what you tweet is not who you are, the values you put on your wall is not who you are.  It’s what you do every day, it’s how you treat each other, it’s how you show up. Do you show up on time? Do you do a deal for the price or the partnership? Do you return a call that day or the next day or never? Are you responsive to your colleagues or not, or only if somebody’s looking?  All those things define the culture of the company and those are the ‘every single day’ behaviors.

The reason why I wrote the book is if you look at the management literature it’s all about the mission statement, the OKRs, the KPIs, the objectives. But none of that’s in any of that. The culture is not captured in the goals.  The culture is the daily process of how you accomplish the goals and these behaviors. It is tricky to manage with conventional techniques. It’s funny because we talk to companies and they’re like “yeah we had an offsite and we got all these values” and I’m like sure you’re gonna have a culture of hypocrisy because I’m sure you’re not gonna follow, and well how are you going to implement them? And they’ll say “well we’ll put in people’s annual performance review whether they follow the values”.  So you don’t even know if they got the phone call. How do you know if they returned the phone call, or when they returned the phone call. That’s not how you move culture, through an annual performance review. This is a daily thing. People have to face it daily. You have to interact with it. So it’s a very different set of techniques than KPIs and OKRs and mission statements and that kind of thing. Not that those aren’t good, they’re good you should have them in your company. I’m not saying that.  I’m just saying culture is a different thing.  

How do you know if you have a culture that’s failing? And related to the question, is Uber or maybe WeWork, are those cultures that are examples of cultures that failed, or cultural failures? 

Well, let me say this – there’s not a company in the world that’s got total cultural cohesion where everybody behaves the way you’d like them to at all times.  That just doesn’t exist. Nobody gets close to that. You can get to good, good enough, better, not having a fatal flaw. I think Uber had a very compelling culture under Travis with a very bad issue and I can get into that.  How do you know it’s not working? I think the little thing that tells you is when people’s behavior surprises you. Like that person’s running around taking credit for that person’s work, or lying all time time, you’re going – “what’s going on?”   

There’s actually an example I have in the book and I want to be careful about this because when I talk about it with not enough context people don’t know what I’m talking about. One of the things I go through in the book, I go through leaders and people who define culture in different contexts.  And one of the contexts is by a friend of mine Shaka Senghor who built a culture in prison as a leader of a gang called the Melanics. The reason I used that example – and some people are like you’re comparing business and prison, what are you doing you’re an asshole. I’m comparing leadership and leadership.  And I’m comparing organizational dynamics and organizational dynamics. These are super insightful leaders so that’s where I took that from. He told me a story that really helped me understand how culture works a lot. He went to prison for murder when he was 19 years old. His first day in, he and six other guys come out of quarantine and the very first thing that happens is while they’re in the rec area, a prisoner walks up to another prisoner and stabs him in the neck.   Then he pulls out the shank, throws it in the trash and walks to the chow hall and has a sandwich. Shaka’s telling me this story: “so we’re looking at each other like where are we at, like this is different”. And he said, “I had to ask myself, could I do that?” The way I wrote the book, he and I are friends so we’re in the backyard and I’m like what do you mean you had to ask yourself ‘could you do that’, you were in for murder you did that didn’t you? And he’s like, “what happened is I’ve been on the streets since I was 14 years old.  I had been shot 18 months earlier, I had PTSD from being shot, I was always paranoid. I’m in a drug deal and these guys come up and there’s supposed to be one guy in the car, there’s two guys in the car. The guy in the back of the car was very belligerent, he was supposed to stay in the car. He gets out of the car, I had a gun in my pocket, I reacted and I shot him. That’s what I did”. He said “this guy spent weeks fashioning a 2 liter bottle into a weapon and decided, and am I gonna stab this guy in the stomach and wound him or am I going to stab him in the neck and kill him?  He stabs him in the neck and kills him and keeps walking and has a sandwich.” He said, “there was no way I could do that, but I had to ask myself could I do that because that’s what it would take to survive here”. I could see immediately, that set his culture, that’s how violence in prison gets set. With that cultural thing, looking at the person that’s succeeding and seeing their behavior.

When you look at your company, you can’t say people are going to behave according to the values on the wall, they’re gonna go “oh that person that just got promoted lobbied the CEO for six months, took credit for her work, did those things”. That’s your culture, that’s how it gets set, because when I’m coming in that’s what I have to do to succeed in your company.  So when you think about why is your company failing you have to ask yourself what is it like for a new employee here, what have they experienced, what’s their experience of the environment. It’s a complex dynamic and one of the reasons I wrote the book that way I did is you kind of have to understand holistically how does a culture move and then you can recognize it and make corrections and learn to program it, that kind of thing.  

That’s a fascinating story.  For the entrepreneurs in the room, you start a company – originally, how do you set the culture? You quote Patrick Collison, one of the founders of Stripe, saying that the first 20 people are the people that set the culture. Is that how that works? 

Yeah I think he’s not quite right about that.  He grew Stripe so slowly, that comes into play.  I do think it’s very influential because people try to adopt the working habits of people that are there.  But as you grow fast and as you bring on leaders from different cultures and companies, they’re going to start to modify it and things are going to start to change and evolve. You have to do things that are stronger than just “okay I hired 20 people who I think I like and I think are going to behave the way I want them to and people are going to follow that.”  That’s not actually strong enough.

I’ll give you an example of something we did at Andreessen Horowitz. I had the problem going in, and every VC says this, of I wanted us to respect entrepreneurs, I want us to respect how hard it was to build a company and have the whole firm feel like that – okay we know what you’re doing, we know what it’s about. Every VC says this.  How many VCs show up on time to entrepreneurs? How many VCs actually get back to you if they’re not investing in the deal and tell you why? It’s very disrespectful to go present all your work to somebody and have them ghost you when you’re trying to find out what’s going on, right? So how do you get from, “we respect entrepreneurs”, to that kind of behavior? The thing is, you’re fighting as a VC this dynamic that’s basically “I have the money, you as an entrepreneur want the money, you have to come see me to get the money and I have to decide.  So that makes me feel like a big person and you’re the little person.” So that’s an attitude a lot of these VCs walk around with, it’s very pervasive in the business.

I was determined, how do we set the culture so that doesn’t happen? One of the first rules I put in place is if you’re late to a meeting with an entrepreneur, that’s $10 a minute fine. So you have to go to the bathroom, you owe me $50 right now. You had a really important phone call to make, great you owe me $100. The thing about that idea is that we have lots of meetings with entrepreneurs and every time somebody would pay me money, they’re like Ben why am I paying you to work here, I had to go to the bathroom and you’re charging me? I took three meetings in a row.   I’m like, look you have to plan when you’re going to the bathroom because you have to understand how hard these guys are working and how much stress they’re under, you have to show up on time. They’re putting the time in, okay we’re putting the time in. Don’t tell me you can’t do it, because if you were getting married like I know you won’t have to go to the bathroom when you were supposed to be at the altar, you have to plan that shit. It’s that mechanic where they have to ask themselves why, why am I planning when to go to the bathroom because it’s not something that you would ordinarily do. Why am I planning that? Because building a company is really hard and I need to know that and we need to show that and that’s what we’re going to be about.  I call that type of rule a shocking rule which is a rule that you have that everybody wants to know why we have that rule. I wasn’t expecting that rule, I wasn’t expecting to pay to go to a meeting, but that’s why we have it and if people have to ask themselves that then that drives home the message and enforces the culture in that dimension. 

I call that type of rule a shocking rule which is a rule that you have that everybody wants to know why we have that rule. I wasn’t expecting that rule, but that’s why we have it and if people have to ask themselves that, then that drives home the message and enforces the culture in that dimension. 

How do you think about the trade-offs in enforcing something like this, either at Andreessen Horowitz or companies that you work with? For example, the high performer who turns out not to be a cultural fit, is that somebody that you fire because of culture, or do you keep them?

I think you have to ask people to play in the culture.  It’s not going to be across everything. If you start and you say okay we have, I call them virtues not values, because I’m concerned with the behavior not the ideas.  Not everybody’s going to have every virtue you want them to have in life. People are people, they’re messed up. So you have to say, okay what are the things that I gots to have and I have zero tolerance for anybody not doing.  I fine Andreessen if he’s late for a meeting so I have no tolerance on that one because it’s such a core value of ours. But there are other things where people, they’re just coming from such a different place and they’re such a high performer and it’s not a core value that’s critical to the strategic position for your business, maybe that’s not the most important thing. 

I think companies end up with having subcultures as a result of that. I’ll give you a really interesting example of that. Do you remember the Amazon article in The New York Times where they basically said Amazon’s like the worst place to work, it’s like hell on earth. All of us in tech were going, wait a minute Amazon is known for the best culture in the world. They do all these things, so meticulous and they even come to the meetings with their little three page memo and they all have to read it before the meeting and then everybody’s got that culture of you really have to think through your ideas and all that kind of thing. So how can they have the worst culture when we all think they have the best culture.  It really kind of messed me up and at first I thought okay you know fake news, Donald Trump, whatever they’re just doing a hit piece on Amazon. But I know that writer, in fact he did a piece on me. He’s one of the most earnest writers, David Streitfeld, this guy he’s the kind of journalist where even if you don’t like journalism you have to have respect for what he does so I knew he didn’t make it up. Writing the book I talked to a bunch of Amazon executives about it and they said: it wasn’t entirely right in that most people like working for Amazon, tech workers love working for Amazon, they love it, it’s great, the culture is great there’s no complaints on that end. But what we didn’t realize until that article came out is we hired so many people from retail, from Nordstrom, from Macy’s from Neiman Marcus and they were like, “we’re gonna work 12 hour days, we got to prepare like this for a meeting, like the fuck? I didn’t go into retails for this, this isn’t the life I signed up for”.  You know, I’m in tears man. So they actually started to create a subculture, a different part of the culture, for people coming from that environment because it just wasn’t going to work. You know they’re not paying them that kind of money to do that kind of thing and they’re not having those kinds of careers. They’re interested in other things in life.

That’s a component in every company in a way. At Slack you know, they started out as bottoms up SaaS and they had to put in a sales force. I remember I spent a lot of time talking to Stewart, like we have to pay commissions, this is sales, it’s a prize fight, no prize no fight! I was just getting like it’s going to be a different culture. But when I went to talk to the sales team, you know we’ve got the original people in and one of the first things I said them was like, look I don’t care that the general culture is at Slack you can’t come to work at 10am in your pajamas, you got to have a clean shirt on, you have to show up at 8, get on phone, you got to call the customers, you gotta do that thing.  You’re going to have different elements here as a salesperson. It’s not going to be uniform all the time.  

Something that’s not in the book, maybe you’ve experienced this. What have you seen in terms of companies that scale globally and actually have literally different cultures from different countries, how do you include those? 

I think what it comes down too, when you think about the culture it’s like what do you need to have that’s universal? To me the things that have to be universal are the things that support the business strategy.  For us, I don’t care where you work at Andreessen Horowitz, if you don’t respect entrepreneurs you are defeating our whole brand, our whole message, our whole positioning in the market. I gotta have that. But there aren’t many things like that that I have to have.  It’s a handful. Whereas I think there’s a lot of unstated parts of a culture that you’re never going to state and some of those will be different. If you’re in Spain you’re going to take big lunches, you can’t stop them they love it, siesta the whole thing it’s crazy, drink at lunch.  I had a hard time adjusting to that culture when I spent time there. That’s not anybody’s business strategy to not have a big lunch, that’s just kind of how you do things in your place or wherever. Some places everybody eats lunch at their desk and that’s the culture. But it’s not an important part of the culture, that’s just how it went so you don’t have to try enforce that internationally.  But there are going to be things that define who you are as a company and what you’re like to do business with. Like for example if you have a thing, like we are responsive to each other if somebody asked me for something in the company I’m going to say yes or no I’m gonna get back to them I’m not going to just like leave them in my email inbox and they’re like a lower rank than me so I’m not going to respond because we’re hierarchical like that.  If it’s part of your culture I don’t care if you’re in France, you have to still do that. It’s a subset of the totality of what the culture is. 

How do you create an inclusive culture, particularly with respect to gender, minorities, etc? 

I learned a lot.  I think that in general in tech people do it, they’re almost destructive with how they try to do inclusion.  Let me explain why. I learned a lot, this is something I spent a lot of time on and I learned a lot on. The breakthrough for me was a conversation I had with a friend of mine Steve Stoute, who’s been in the paper a little bit for the Knicks for those of who are Knicks fans, he’s doing work with the Knicks and causing trouble there.  He’s a kind of boisterous, charismatic personality. He calls me up one day, he goes, “Ben, I used to be President of Sony Urban Music,” and I already knew that you know, I’ve known Steve for years but I knew he was saying that to me for a reason, he wants to tell me something else. I go, yeah Steve, I know you were President of Sony Urban Music, and he’s like “yeah but it wasn’t Sony Urban Music it was Sony Black Music but they wouldn’t let us call it Sony Black Music because that would be racist so we had to call it Sony Urban Music.” And I said well that’s silly, and he said “no, no that wasn’t what was silly.  What was silly was because they called it Urban Music they didn’t let me market in rural areas. They only marketed in cities. Like there’s no black people living in rural areas.” I said, that’s stupid. He said, “you’re not even listening to me, Ben. I was President of Sony Urban Music, I had Michael Jackson.  What, white people don’t like Michael Jackson? That wasn’t black music, it was music.  What are they talking about, it was music. They put me in this category and it didn’t make sense at all”. 

It struck me, I was like oh shit, that’s what we do in Silicon Valley, urban talent. It’s not urban talent, it’s not black talent, women talent, Hispanic talent, it’s talent.  If you don’t know that, it’s because you can’t see the talent just like they couldn’t hear the music. If you look at the epilogue on Sony Urban Music, once streaming came out, guess what the urban music department, the black section of Tower Records went away, the radio station segregation went away.  All the stuff that they had built, the whole infrastructure to market black music went away. Look at Spotify, 92 of the top 100 songs are “black music”. It wasn’t black music, it was just music. And because some executives couldn’t hear it, they cut off most of their market. And that’s what we’re doing in the talent market.

So I’m going to take it all the way down.  I’m early in the firm when we have this conversation.  We have what a lot of you who have companies have. You have a woman running a department, guess what you have a lot of women working there.  You have a white man running that department, guess what you have a lot of white men. You have an Aisan person running that guess what? Why is that? Well I know what I’m good at, I value it highly and I can test for it in an interview.  I can see me so that’s what I’m hiring. Me, more of me, because that’s what I can see. I had Frank Chen and he had all Asian people working for him, I had Margit Wennmachers running marketing, she had all women working for her. So I’m like, I’m cracking the code on this thing.  So I go to Margit and I’m like, Margit what is in your hiring profile, in your job criteria where no men can get a job working for you? She looked at me and she said `helpfulness’.  I was like, oh boy, I don’t know a lot of helpful men.  But let me tell you how stupid I was and how blind to talent I was when she told me that. We’re a venture capital firm.  We’re a service firm, we’re in the business of helping entrepreneurs, so that is our business – helpfulness. Anticipating a need before you ask for it, making sure you get what you need.  No other job profile in that company outside of her organization had helpfulness – she was the only one that had it. So here’s this criteria that’s material to our business that we don’t even know was a criteria.  That’s how dumb we were. We corrected and just doing that, we’re 180 people, we’re 52% women, we have something like seven general partners, something like that. But we don’t have a Head of Diversity, we don’t have a program, we just have criteria that’s wide where we can see talent that we ourselves might not have.  

I’ll give you a second example on this because you mentioned the Cultural Leadership Fund.  We had a position early in the firm where we wanted it to be, if you were with a venture capital and we liked your company we wanted to be able to do more for you that those guys.  We wanted a person who would go around and meet CEOs from firms we aren’t invested in and connect them to the firm and we could help them find talent and all these other things. The person hiring for that position was an investment banker.  Everybody who I saw in the interview process was an investment banker. All investment bankers. So I go to him, and I’m like, hey what is the job profile for this thing? He said, well they have to be great at follow- up, really smart, detail oriented, you know, all the things investment bankers said and I was like okay great, but isn’t the main criteria for this job is to be able to build a relationship with somebody who has no business reason to do so, like a great relationship person.  And he said, yeah. Well I said, who likes investment bankers? He goes, “oh wow, like that’s crazy, so what do you want me to do?” I said, well think about the kind of talent pools we can draw from, what culture do you know of that’s very relationship oriented. Where the relationship is where it starts, that’s the center of the activity, that’s the jump off point. He goes, well I have no idea. I said, well I’m not gonna tell you how to do your job but in my experience if you grow in an an African-American family your parents will teach you how to talk to people when you’re young, that’s how much focus there is on relationships there, so that might be a good talent pool to look at.  And he says, “well do you want me to hire a black person?” I was like, “no I didn’t say that, I want you to hire on the right criteria.   I’m saying that you may be able to find somebody in that talent pool that could come into venture capital that could do that.  

And he says, “well do you want me to hire a black person?”I was like, “no I didn’t say that, I want you to hire on the right criteria.   I’m saying that you may be able to find somebody in that talent pool [African Americans] that could come into venture capital that could do that [be great at building relationships].

We ended up hiring a guy who was previously a sound engineer for Jermaine Dupree.  The reason I bring that up, in Atlanta, is you know, people go “there’s not enough pipeline”.  Well there’s not a pipeline if you don’t know what the fuck you’re looking for, but for that there’s tonnes of pipelines.  So we hire him, his name’s Chris Lyons and after we hire him – and again this is a talent blind thing, I find out he used to work at The Cheesecake Factory.  So The Cheesecake Factory measures every waiter by the percentage of tips against bill. He’s a number one waiter nationwide. He’s like the best instant relationship person in the world.  And today if anybody knows, he’s the best relationship person in venture capital, no question. He’s the best. And why is that important? Because we hired him on that criteria. [Edited] He’s running the Cultural Leadership Fund now.  But it was only because we knew why we hired him.  

If you hire someone based on race and gender through the urban talent department instead of through the normal talent department, through the fucking side door instead of through front door because you can’t see the talent then you’re never gonna be good at inclusion.  You have to start with understanding people and understanding people who have talent that you don’t have. Do you know how hard it is to know what somebody wants before they even say anything? That’s a real talent and skill. That’s a real fucking tangible thing that’s real and you have to be able to recognize it if you’re gonna get it and get that competitive advantage.  You can’t just go, “I’m hiring a woman”, which is what people in Silicon Valley do and it’s why they get it all screwed up and then you go into these companies and you look at attrition and employees sat and so forth, and it segments by race and gender. That’s the metric. Employee satisfaction, promotion, like is that equal? Not coming in the door. Coming in the door you can fake that by doing the dumbest thing in the world by creating the urban talent department.  Sorry I got too excited about it. I’m sorry I went off like that’s the longest answer to a question ever. That’s it, sorry. 

You can’t just go, “I’m hiring a woman”, which is what people in Silicon Valley do and it’s why they get it all screwed up and then you go into these companies and you look at attrition and employees sat and so forth, and it segments by race and gender. That’s the metric. Employee satisfaction, promotion, like is that equal? Not coming in the door.  

Maybe one last question from me because I want to open it up to people.  As a venture capital investor, how much do you both factor culture before you invest, how do you test for it, how do you get a sense for what kind of culture you’re investing in – and then as you become a board member, to what extent is it your role to be the helper or guardian of a specific type of culture.  

It’s funny, once I wrote the book people are like, whoa so you evaluate all companies based on culture and then people in my company are like, whoa are you sure you want to do that culturally? I’m not like a magic person, I can’t anticipate the culture of a company that comes to me with two entrepreneurs and like what they’re going to have.  There’s not much to evaluate when they start. I don’t think day one you try to set the culture of a company, I think you don’t know enough. You kind of have to learn, okay what works, what makes us a good place to work, what makes us a good company to do business with, what do we need in the culture that supports the strategy. Then you develop it over time.  So there’s not much to evaluate – when Databricks came to me, I invested the day they incorporated the company. They don’t have any culture, they don’t have any work space or anything. Having said that I think that if you have a founder who’s like telling you some pretty aggressive, not true things in the pitch that’s going to be a tricky founder to work for because it’s very rare to have a founder who lies only to investors and not to their employees and that kind of thing.  So that tends to be one that kind of goes everywhere that person goes so you have to be mindful. You have to invest in people who you think have the leadership skills to build a culture. I wouldn’t know to walk into a new company and evaluate what their culture is going to be. 

One of the key, interesting things in the book is that there’s not one culture that wins.  Different cultures, and what worked for Amazon didn’t work for Apple…

Right, so one of the things I point out in the book is that Amazon was famous for its frugality.  When the first employees came in your desk was a door two by fours holding it up, that kind of thing.  “We ain’t gonna buy a desk, we’re gonna get an old door and put it up because that’s how much we wanna save money”.  But that was very key to their strategy because the Amazon strategy if you remember his original chart was to be the low price leader in everything because that way as a consumer you wouldn’t have to price compare you just know you could get it at the best price on Amazon.  That meant that you really had to control cost if you were going to do that. That was integral to the overall strategy of being the everything store. Apple could say look frugality worked for Amazon, I’m gonna do it. But Apple doesn’t have that value, and nor should they, and nor can they, because their strategy is to have the best and most beautiful tools in the world for their customers and spare no expense. In fact Steve Jobs got fired originally for sparing no expense and that’s exactly what got him fired in the beginning.  Even their campus is like five billion dollars, the doorknobs cost like eight thousand dollars each. So that’s not the right value for them, Apple will probably always build more beautiful products than Amazon and Amazon’s products will always cost less and that’s fine you can have two different cultures. So I always say whatever your culture is should support your business strategy, it should be what you borrowed from Jeff Bezos ‘cos he built a different company than you’re building.  


Can you quickly comment on the quantification of culture and in the emerging space of HR Analytics? 

I think it all depends on the virtue or behavior that you’re trying to drive.  There are things you can measure, for example do a customer satisfaction survey of entrepreneurs we reject to measure whether we treated them respect, because that to me is the litmus test – are you treating people with respect? Would they recommend us even if we didn’t do the deal? Did we get back to them, did we give a clear reason why we weren’t investing, all that kind of thing.  So we have numbers around that. We have other things in the firm that are trickier to put a metric on, one of them, one virtue that we have is we support “not trash dreams”, because we believe every entrepreneurial idea is a dream, so there’s not a metric, there’s not a carrot on that one there’s a stick. Which is one of the things I tell everybody at new employee orientation – if you get on Twitter like some VCs and start shitting all over somebody’s company like “oh they’ll never make any money, that’s a dumb idea, I can’t believe they called it Bodega”, some of you remember that company, then you’re fired.  I don’t care how bad the idea is, if your idea is to spit on somebody’s dream to make yourself look like a smartypants then like you can’t work here. [Edited] Other people can do that, but you can’t do that and work at this firm. But that’s not really gonna be a metric. You do not do that under any circumstance, you might as well shoot somebody in the head because you’re gone. It’s not an objective is a thing I would say about culture, it’s a process. It’s the time you spend on your way to the objective, which by the way is most of your time. It’s the thing that people value most about their time at your company when they’re no longer there and it’s that kind of thing.  

We’ve been talking about corporate culture and culture in VC firms, etc.  I was wondering if you could comment on how culture develops in cities for example the tech scene in Silicon Valley versus New York, Denver, Austin.  How do we think about those, how do they develop, how do we create those types of different cultures as time goes on?  

Yes and it is the same kind of thing which is adoption of the behaviors that work or are cool or make you the friends that you want work which is the same thing in a company.  I think the big difference is a company’s got a CEO who has special powers, to fire people and reprimand them and so forth but in broader culture it doesn’t work that way. Actually one of the reasons we started the Cultural Leadership Fund is wanted a partnership with people who knew how to move culture in the general sense.  If you look at the LPs in that fund, Quincy Jones, P.Diddy, Kevin Durant, so forth, they’re all people who have genius ability to do that. I would just say it’s very different. I won’t presume like, I’m not a real expert on how to move culture in the broader society. Which is why we have the partnership. I think it is important, like if you’re trying to change consumer behavior or something like that, understanding how culture moves is critical.  It’s a good question, but it is the same thing but different. The section on the book on samurai was an interesting one because it was a broad society, it wasn’t a hierarchy but they developed a code that sort of dictated the culture for the society.  


Thank you:

Leave a Reply

Your email address will not be published. Required fields are marked *