Delian Asparouhov: [00:00:00] Hi everyone. My name is Delian and I'm a principal at Founders Fund; a venture capital firm based in San Francisco. This is Operators, where I interview non VC, non CEO, non founder operators that make the startup world go round. Today, I'm interviewing Keith Rabois, who is currently a general partner at Founders Fund.
Prior to joining Founders Fund, Keith was an early executive at several successful companies, including PayPal, LinkedIn, and as a COO of Square. He also spent six years at Khosla Ventures prior to joining Founders Fund. I hope you enjoy the show.
So Keith, you've been on plenty of podcasts over the course of the past five years. You know, ever since those early days, but I feel like most of them have been focused on kind of like startup formation or like, you know, talent, or overall corporate strategy and fundraising, but today I wanted to actually dive into sort of your operating history and go into the specifics some of like the companies you helped build up, and the projects that obviously you worked on there. And maybe, you know, to backtrack all the way to the beginning of your career, which I think most people don't realize, is that you actually started off as a lawyer, at a top tier law firm in New York. Can you maybe talk about what made you consider leaving, you know, the very sort of traditional, let's say law firm track all of a sudden, to go into the world of technology?
Keith Rabois: [00:01:22] So, yes, I started my career as a lawyer. It's what I really wanted to do since I was six years old. I wanted to be a litigator, and so I optimized my entire academic career really to maximize the chance of getting into a top tier law school. So I went to college. Studied political science, started all these organizations, ran all these organizations, optimized my GPA. And I was terrified of doing anything that would take me off the Harvard law school trajectory. So that's basically what I did, I went directly from undergrad to law school. Clerked for a well-known judge in the United States court of Appeals fifth circuit and then joined a canonical Wall Street law firm. If you have any stereotypes in the back of your mind about what a Wall Street law firm looks and acts like, that was it.
I had a full-time secretary. We didn't have email. I hand-wrote all my documents. I wore suit and tie in the office six days a week. Your casual day was Sunday and, pretty much practiced law for three and a half years. Somewhat ironically, I actually worked at the exact same law firm that Peter Thiel did. Peter was smart enough to quit after three months and four days; it took me three and a half years, which he still introduces as my greatest character flaw, that is it took me three years to figure out what he figured out in three months. It's probably still true of most of our conversations, but, basically I got the right answer and left the practice of high end litigation, to jump in quite cold turkey into startups. So I didn't do the bridge into like, a corp dev role. I didn't do the bridge into the legal department of a technology company. I went to full stop, you know, I'm the business person and I'm going to do this review for real. What led me to this was, fundamentally at the end of the day, a lot of my friends who I’d grown up with at college at Stanford had been involved with the first generation of internet companies, either as founders or senior executives, and they kept trying to persuade me that I'd actually enjoy this, that I needed to stop wasting my life, you know, writing briefs and stuff, and that I needed to try this and every year or so, I'd come out once or twice to visit Silicon Valley, and they were trying to convince me to join this. What they actually, at the time, correctly labeled as gold rush, and I was missing this opportunity. And then finally in early 2000, I was persuaded by a friend of mine to join his new startup and jumped in February of 2000. Now, diving back into history, the advice might've been good and pretty accurate about why I'd be successful, why I would be happy involved in technology companies, but the timing was miserable. So the market collapsed, March 28th of 2000, and permanently collapsed in June of 2000. So I sort of made this career transition at the worst possible time, like six weeks after making a career transition, the world exploded. So not necessarily ideal, but obviously with the benefit of 20 years of hindsight, it worked out fine. But they were mostly right that I would enjoy various parts of technology companies, team building, mentoring people. Law, even at the high end of law is mostly a solo or very small group activity. So a substantial team working on a substantial piece of litigation might involve anywhere from two to five lawyers.
And that would be like the massive amount of scale that you ever see in law. At least the law firm I worked at is a very generalized-everybody's a general practitioner and therefore there's parts of your job in anybody's job and career where you're more proficient and more interested in some activities and less proficient and less interested in others.
And the way we practice law, you didn't really have the ability to do more of the things you've enjoyed and less of the things you didn't want to do. You had to do it all, and that was a little frustrating. Obviously business revolves around specialization and domination of expertise and being the best in the world at some things and hire people to do the things you're not good at or don't want to do.
So that was extremely attractive to me. And then fundamentally there was some objection to the lack of short-term meritocracy in law. I think long-term law is a quite meritocratic profession, but we would progress as associates on a very lock-step model. So you have second year associate you're a third year associate, then a fourth year associate, then a fifth year associate. And there was basically no way to skip anywhere in line based on performance. And all of your compensation was basically homogenized to whatever associate group you were in with like the variation of maybe one to 2%. And that was incredibly frustrating to me and several of my talented colleagues.
So eventually I jumped cold turkey, moved out to Silicon Valley, and figured out somehow or another, despite the internet almost collapsing all around me in the nuclear winter, starting how to become successful both personally and collectively as part of an organization.
Delian Asparouhov: [00:06:25] Yeah, your first jump was into this company, voter.com that you spent about 10 months at, and then headed over to PayPal from, if I remember correctly, a call from Peter.
I remember when there was, this person writing a book on PayPal, they actually were able to dig up one of your earliest memos, which I believe was-it almost basically read like a brief, against Visa in order to keep PayPal functioning. Can you talk about just sort of like what some of the early work that you dove into was?
You talked about how you didn't try to like jump into the legal side or lobbying. You jumped straight into like business. What were parts of that? What were sort of like easier to pick up versus what kind of stuff was much harder to pick up?
Keith Rabois: [00:06:58] so initially I jumped into voter.com as general purpose business person doing business development, and then ultimately general managing pipelines, and financing for helping raise money, et cetera.
And general strategy. But when I moved to PayPal, one of the crises that Peter was confronting and where he was really unhappy with his executive team was around managing the existential threat with Visa and MasterCard.
So PayPal basically built a product interface on top of the Visa/MasterCard Card rails, quote on quote, which is like a term of our industry. And neither Visa or Mastercard liked it, and both at different points in time could try to sabotage our company, and Peter didn’t think that the current leadership of the company of PayPal was taking the threat seriously and being aggressive about it.
So he recruited me to start quarterbacking that strategy differently. So the memo you're alluding to I wrote actually literally on my first day on the job, and unfortunately reading it in retrospect, it's probably better than anything I could write today, which is embarrassing as hell, to reread. But yes, there's a forthcoming book about the PayPal history, which I'm quite optimistic about because as the benefits of 20 years hindsight, whereas, you know, the only book really written about PayPal was written right after our, you know, sort of sell to eBay. Insofar as it has merits, which it does, it also has a lot of disadvantages having been so close to the timeline. So this book I'm really optimistic about, I think it will be published early in 2021, but the author has done a phenomenal job of researching, got access to a lot of regional documentation and, including this memo, which basically devises and outlines a strategy for setting us up for success in confronting Visa and MasterCard.
And some of the elements of that were to set up a legal, either defensively or offensively, case against Visa and MasterCard, by clever tweaks in our product and marketing, as well as some other elements, but in any event I'm embarrassed that I can no longer write anything close to that.
And would probably, even if I try, it would take me two months.
Delian Asparouhov: [00:09:09] It seems like, you know, most of these startups that have to deal with like super highly regulated environments have to understand a variety of different aspects, both like the incumbents incentives, how to, you know, lobby governments and regulators. I feel like, I mean correct me if I'm wrong, but PayPal was one of the first tech startups to really have to have a substantial sort of public affairs campaign and lobbying efforts.
Can you talk about just like the early days of spinning that up? Like, you know, obviously you wrote that memo to outline the strategy and then what was it like executing the strategy?
Was it, you know, a team they had to build in house? Was it, you know, external sort of firms you'd worked with before? How did you end up sort of executing on that memo?
Keith Rabois: [00:09:43] Most important point is: PayPal had a lot of enemies. Visa and MasterCard really despised us.
eBay hated us, and then later the federal government post 9/11 started to really wanted to regulate financial transactions. And some of the burden of those regulations would have been very threatening to our business and offensive to our users. So we had multiple people and then there's state governments and banking issues.
So we just had lots of people that were opposed to our business strategy. And about a year after I joined, we started making more of an investment in lobbying, all across the board, both at the local level and the federal level. And we started small truthfully, basically I was doing a very small budget to hire effectively what was an outside agency sort of thing, and prove that we could get traction with that very small budget.
So by the time I left the company, we had 17 lobbyists working for us, and you know, we're spending a considerable amount of money—call it like—low digit millions per year—but I had to validate that what we were able to do would translate to business success. As you might know, Peter is pretty demanding.
David Sacks and Roelof were certainly people with a lot of scrutiny, Max Levchin needs to be persuaded on lots of things. So every step of the way, I had to show that the injection of this amount of time, energy, and money would produce business results. It wasn't like, 'hey, go off and just spend money and we're going to have this lobbying thing cause someone told us it would be valuable.'
It was go do acts. Show us the benefits of acts, and then we'll give you more. So over the next two years or so, wound up building out a very substantial team, both internally. I moved a couple of colleagues over to my team to help manage this, and then we recruited experts because we needed lobbying expertise at federal level, state level.
There's differences between lobbying the executive branch of the United States government and the Congress, even within the Congress. There's Democrats Republicans, there's the house, the Senate. So if you really want to be excellent at any initiative, you need the people who have the right skill set and background, and so we put together a pretty comprehensive strategy and team.
I was basically flying to Washington, on average, once every four to six weeks. I even forced Peter to come with me once in a while, which, you could imagine that conversation. Peter's favorite thing to do is to fly to Washington, and put on a suit and tie. It was very successful ultimately with Visa and MasterCard. Extremely successful with eBay, and surprisingly and somewhat counter-intuitively even against my own external assessment, against the US Treasury Department's regulations, we were very successful at deferring or undermining that.
Whereas I was somewhat nervous about taking on that project, originally. Which I literally took on the project right after 9/11 and was not internally optimistic about our likelihood of success, but it turned out to be quite…we tapped into some waves and gains that I didn't realize existed. We were able to really parlay that into massive success as well.
So yeah, we, we basically proved that a small startup could leverage the lobbying efforts more akin to what a Fortune 500 company would do.
Delian Asparouhov: [00:13:19] And can you describe, one of those, let's say of those three examples that you just listed off—the Treasury Department one, I think would be the most interesting. If you remember both, like, what were the potential regulations that were coming down the pipe that could have affected, PayPal? But then also what were some of these sort of like, you know, waves that you tapped into that were potentially counterintuitive and, how did it end up sort of like, you know, playing out or getting fixed for PayPal?
Keith Rabois: [00:13:39] The post 9/11 Treasury Department promulgated a lot of regulations designed to limit terrorist financing. So one of the issues that was allegedly a problem in causing 9/11 was terrorists were able to raise money, transfer money without a lot of scrutiny. So the government started cracking down on this.
It added a lot of incremental regulatory burden to banks and financial institutions opening accounts, moving money around, filing SARS, suspicious activity reports. And one of the specific pieces of the regulation that was most concerning for PayPal was the requirement that it was like social security number from every user.
And as drafted initially it would require having a social security number for not just the merchant side of PayPal transactions, but also the consumer side. You can argue about the merits of the merchant side collection and how much that would deter people-successful businesses from using PayPal, but it was almost surely the case that collecting on the buyer side would have been a disaster.
And so having to kind of undermine these regulations, and because this was so close in time to the 9/11 experience, I felt like my initial reaction was that the lobby against this would be very difficult because there was this sort of patriotic belief that all of these regulations were wise and smart and would be successful. Obviously with the distance of time, probably all those things are false, but it was very challenging to make that case in the short term, but we had to because we really believed that at a minimum it would reduce our transaction volume by 30%.
And that was like a best case kind of internal analysis. So obviously when you see, you know, 30% potential drop in your business, you're willing to allocate some energies, efforts, and money, and even some risk to try to solve that. And then that's what we did.
What I didn't realize at the time was there's a lot of people in different parts of the government, different constituencies on Capitol Hill that were somewhat, for their own reasons, opposed to these regulations. We were able to craft the kind of unusual alliance between basically extreme left-wing Democrats and extreme right wing Republicans, and put a lot of pressure on the Treasury to sort of either modify the rags or withdraw them. And were so successful, in the fact, that maybe the highlight of my lobbying career, highlight of my business career perhaps, but was the Undersecretary of Treasury just called me up one day and said, "What do we have to do to make this stop?"
And what he meant by that was I'm getting so deluged by congressman and senators’ staff and congressmen and senators calling me complaining, that it's driving me absolutely crazy. So, what do we need to promise you to make them go away? And, so, that's when I knew we'd be very successful. But I was not at all optimistic initially, and it turned out, in fact, the momentum that we created against these regulations was able to basically suspend their implementation for about seven years thereafter.
So we basically burn the Treasury so badly and they were so frustrated and so annoyed and so worried about more negative feedback, that they basically would refuse to basically test this issue for a very long time. By then lots of other things in the world had obviously shifted around them as well.
So that was massively successful. We're also able to avoid eBay killing us basically by acquiring them to dumb down their product initiatives by running them through lawyers. So part of the art with eBay was to basically remove any of the most edgy features that they could launch that might be frightening to us because they would have to show them and review them with their lawyers. Who'd be nervous and also slow them down because legal review, debate, you know, consensus decision making takes time, and we could still be agile.
We didn't have a consensus driven culture. We can be very decisive. So there's very much a time-based strategy. And then with Visa and MasterCard, it was over time allowing us to be successful enough, get enough scale, get enough data, get enough predicted algorithms in place that we didn't need to leverage their system in a way that alienated them.
So basically we were, again it’s similar, playing for time.
Delian Asparouhov: [00:18:22] And with eBay, was it an anti-competitive case or what was like the fundamental tenant that you like pushed on them?
Keith Rabois: [00:18:26] With eBay it was an antitrust case. And we were actually very close, extremely close at one point to actually filing the lawsuit against eBay, within hours to a day of actually filing it.
And we decided not to, for a variety of reasons. But we were generally threatening them all the time in a fairly sophisticated way. These were not like generic threats of like,'Oh, this is antitrust violations that we hear and see all the time. These were very specific, narrow, specific claims with very sophisticated lawyers, serving them.
So they would have been, you know, completely distracted as a management team for a year or two, having filed this litigation. That said looking in the mirror, this is definitely one of the reasons rationales behind us not proceeding, was it would have distracted our company and net-net we decided that we had a higher velocity of execution.
We had a higher pace of innovation. The distraction to us in some ways would actually be worse than the distractions to them.
Delian Asparouhov: [00:19:32] Right. I feel like you've always talked about both when you're considering funding startups or working at them, that you don't tend to like companies that are heavily dependent on external partners for success. Yet. PayPal seems like the epitome of this. Like, you know, platform dependency with eBay, you know, partner dependency with like Visa, MasterCard, you know, huge potential sort of like liability and risk there, I guess like, you know, have you shifted your opinion? Like have you counter shifted because of the PayPal experience? Or was it something that like gave you confidence going into PayPal that you'd be able to execute all these strategies successfully in order to make PayPal successful in the long-term? Or are there just like times where you sort of make exceptions to this like partner dependency rule now based on seeing some confidence in, people's ability to manage that?
Keith Rabois: [00:20:15] Well, I think we were burned by all these experiences at PayPal. And if anything, my allergy is a reflection of that. Maybe it's possible, but you know, we have an unusually dense pool of talent at PayPal and sort of use a lot of heroic efforts and a lot of incredible talent to minimize the risks associated with all of these platforms dependencies on eBay, partner dependencies with Visa and MasterCard, regulatory dependencies being a financial services institution.
So I, I think being burdened by those maybe makes you develop sort of an allergy to doing those again. Nevertheless, I think there are times and places for all of them if you have an unusual insight into how they both calibrate the risk associated with these and then how to sort of manage your way out of them.
So for example, in the Slide case, which we can talk about at length, but fundamentally Slide was predicated on building applications on top of initially MySpace, and then later Facebook where we had massive platform dependencies. We never really solved it, and in fact, it burned us very badly, which is why no one remembers Slide other than my soccer teammate. Then the second, you know, financial services version though, the Visa MasterCard platform dependency or product dependencies has changed a lot. Back then, Visa and MasterCard were private organizations.
So they're basically country clubs and affiliated members of banks. So they're basically in their organization the only thing Visa and MasterCard did was enforce rules across the banking system. And their entire incentive structure was designed to not add any risk of something going wrong. A few years later, they became public companies, quite large public companies, actually.
And so now they operate more like traditional businesses. They have their own needs to grow. They want incremental volume. They want incremental revenues, much easier to deal with it. And you know, when I went through this kind of transition in my own life, personally, of dealing with Visa and MasterCard at Square, it was quite easy compared to my experience, you know, only a few years earlier with PayPal partially because those organizations that completely transformed themselves culturally, structurally, so that they became partners for Square, they became investors in Square. We helped bring incremental volume into the Visa/MasterCard system.
So Square basically converted paper transactions or cash based transactions into electronic digital transactions that Visa and MasterCard were getting paid for. And we were able to frame that argument successfully. Partially based upon my experience with PayPal, with what arguments didn't resonate, but also just the leadership changes and structural changes at those organizations.You know, Amex was always complicated both, in the PayPal days actually, it was easier to deal with Amex, because they were more flexible cause they wanted market share. Square days, they actually were in some ways the most difficult partner for Square even though they had the smallest market share, because they actually saw the strategic threat of Square, to their direct salesforce.
But in any event, I do generally try to avoid businesses that have massive platform dependency. I think sometimes the platform dependency can be overrated. Like you'd need to have a theoretical approach of minimizing awkward dependency over time, and if you invest in that early enough in the business trajectory, by the time platform dependency becomes critical, you may already have solved the problem.
Yelp started having a massive platform dependency on Google. Obviously SEO is key part of the growth strategy, but with the transition to mobile, particularly on iOS, Yelp users who were addicted, tended to use the Yelp app, which allowed us to eliminate Google from at least the retention loop and the engagement loop, for a majority actually of the uses of Yelp.
So by time, we were able to escape the platform dependency in some cases with Yelp. SEO is a little bit like that, where you're playing for time, because Google doesn't usually move overnight in reorienting its search results, but over a decade, you definitely want to try to eliminate your dependency on Google, on the regulatory side though, I actually embraced it, as you know Delian, I prefer to invest in companies that are heavily regulated. I think it's to many people, regulated industries, whether financial services or healthcare as examples, are typically thought of by other investors as a black box and the investor doesn't really understand what's underneath the hood and therefore they have to outsource their diligence or feedback to lawyers, versus deciding for themselves from first principles, whether this risk can be solved, whether it's acute, whether it's fatal or not, and I'd prefer to make that decision myself. Because I still know enough law to be dangerous and hopefully have enough business instincts to combine the two together.
So I love times when there's a lot of perceived regulatory risk when a founder is pitching us, because I feel that if I calibrate the risk correctly and make the decision in my own brain—I have an alpha or competitive advantage or part of an advantage against all the other investors. So the more regulatory risk in some ways better for me as an investor, doesn't need to be better for me as an executive. I do remember a lot of sleepless nights, and you know, working a lot of hours to solve, like some of those regulatory risks, but they don't terrify me as an investor at all. In fact, some of the best investments that I've ever made have absolutely confronted substantial regulatory risks, some of the ones I’m currently working on have lots of regulatory risks, but I would prefer to be doing more of that rather than less.
Delian Asparouhov: [00:25:59] You know, after you left PayPal, about two years later you ended up joining LinkedIn, which is probably the opposite of all of these, which it has zero platform dependency, zero regulatory risk. I guess, what got you excited to go there? Like why not? I guess like another financial services, regulatory thing where you could apply your legal background, et cetera, you basically joined the equivalent of a social network, which obviously I assume has relatively limited, like let's say like, you know, legal needs. So what got you excited about that opportunity?
Keith Rabois: [00:26:23] Yeah. So in fact, it's not accidental. I think Reid Hoffman, when he was starting LinkedIn was looking for something that had none of these risks, because a lot of Reid’s portfolio over his experience at PayPal was dealing with a lot of these issues, which were very acute, and borderline crisis all the time. So he wanted to create a business that was completely insulated from this and found the perfect one. So yeah, it was somewhat nice. I didn't lose any sleep at LinkedIn, ever. So that was probably nice.
Yeah, the only question at LinkedIn, you know, we had been a viral growth strategy that was basically email-based viral growth and we had a retention strategy that was mostly email-based at the time. Revenue strategy that was mostly under our control, although we did look at some modernization options, explore them that would have been more complicated from a regulatory perspective. But fundamentally at the end of the day, LinkedIn was designed to be, you know, self-sufficient animal with no risks anywhere. The risk is really just user traction, user growth, branding, adverse selection. So we did have substantial concerns about building a marketplace of talent. and would you get adverse selection? It's very easy to convince people that don't have jobs to join new platforms to find new opportunities. It's a lot more challenging to find managers, directors, successful people who have the opportunities to join the platform.
So that was the fundamental risk to the company. As matter of fact, In June of 2003, when the product launched I used to wake up every morning and go through every single profile that joined LinkedIn. Literally every single profile.
And the reason was, I did this for months, was to check on the quality of the people joining and to make sure that they would be a balance between people who had opportunities and job seekers. And the only way back then maybe—these days are fancy now—there might be a way to algorithmically determined the quality of users joining, but it was much easier and faster for me to just look at every single profile and get kind of an anecdotal impression, but it felt like we were doing a quite good job.
And there's some very conscious decision-making behind the product, behind the marketing, behind the framing that was designed to offset the risk of adverse selection. So for example, LinkedIn would never use networking as a verb, because there's a stigma to networking as a verb, at least in some places.
And so we were very conscious about vocabulary and design to ensure that we could have a healthy and balanced ecosystem.
Delian Asparouhov: [00:29:05] And it sounds like one of the first like super successful projects that you had early on, there was actually the public profile, which I didn't actually realize wasn't something that, you know, was a part of LinkedIn from the launch. So how did that come about? And when was it clear that that was going to sort of be successful and significantly impact the growth rate of the company?
Keith Rabois: [00:29:20] Yeah, actually, public profiles were a little bit later, than when I joined, at least a year later. Actually it's a spark that came from Yelp of all things. So I joined the board of Yelp, in 2005, which is basically creating, you know, public profiles for businesses. The magic of Yelp was that every business is effectively a very long tail search query.
And if you're a startup, competing on long-tail search queries is actually a pretty easy SEO strategy until someone else figures that out. And so I had this epiphany one day at dinner, Jeremy and Ross the co-founders of Yelp, that, 'Oh my God, the same principle applies to people. People are the ultimate long tail business.' So, I mean, yes, there's a few very successful people-think Elon Musk or something where you're not a long-tail business, but 99.9% of the workforce is a micro-business.
And I was like, 'Okay, well, we should just be indexing every single person as if they're a business on Yelp.' So I went in the office the next day and it's like, here's what we need to do. Here's why it should work. Boom. And then, because it was such a great idea, but it wasn't on the pre-existing roadmap, we actually had to find a way to build it, engineering wise, because LinkedIn moved at sort of a colossal pace of execution. So things that weren't on the roadmap were really hard to add regardless of the quality of the idea. But fortunately Reid appreciated the idea, and was willing to fund it. So what we did is we built a separate engineering team. I hired this guy as a sort of part-time VP of engineering to build out a whole SEO team. I took one of my deputies, David on to manage it and build out the SEO team that would create public profiles that could be SEO’d successfully. And they did such a phenomenal job that even today, 15 years or so later, if you Google someone's name, that's like a normal person, odds are by a lot that the LinkedIn profile will be the first search results against Facebook against anybody else, because it was so well architected from the very beginning. So yeah, we had to find an orthogonal way to build it, but once we shifted it was clearly working, you know, I think I mentioned it as early as 2007 or 2008 it already 20-25% of all LinkedIn's traffic, and I think that's probably even scaled up a bit from there.
Delian Asparouhov: [00:31:41] So the public profiles existed. It was the indexing on the public profiles and like, you know, creating SEO around them that was really causing the growth. There wasn't like, let's say like social viral effects of it. It was really just this long tail search is what really enabled it.
Keith Rabois: [00:31:55] We didn’t have officially a public profile, but because there were some concerns about privacy we created this hybrid, public profile in conjunction with the SEO strategy where you could primarily control what information was indexed by Google.
It turns out like most of these settings almost, nobody alters them, but we needed to build that flexibility just to avoid anybody, particularly in Silicon Valley echo chamber freaking out.
Delian Asparouhov: [00:32:24] So beyond the user growth of the company, you obviously focused on the revenue streams as well. It sounds like one of the, you know, sort of two most successful projects was offering super precise targeting somewhat similar to, you know, direct mail and then offering the premium subscriptions to individual users rather than corporations. Can you talk about sort of both those revenue streams? Did you have familiarity with direct mail beforehand? Were you just trying to replicate that. How did those revenue streams scale up and maybe what were other revenue experiments you tried early on, but you know, it didn't necessarily work as well?
Keith Rabois: [00:32:52] Yeah so the original reason I was hired, which is worth double clicking on was the company had reasonable, moderate, viral email-based traction of about 1.1 million-going on 1.2 million users, but didn't know how to monetize. And because of the legacy of the nuclear winter of Silicon Valley post 2000-2004, there was significant concern at both the executive level and particularly at the board level of LinkedIn, how will we make money?
We get asked this question all the time. You get asked by your parents, get asked by your board members, it's like a universal refrain. And the lack of conviction around our ability to make money led to me being hired specifically to generate revenue streams orthogonal to jobs. So everybody's sort of knew that one way or the other we would try to launch a jobs forum. It was so obvious that if you had lots of white collar professionals, what you'd do with that to make money is you offer jobs. But at the time there was three to four very powerful incumbents to the job space, HotJobs, Monster, CareerBuilder, and Craigslist to some extent, the low end of jobs. The question people had was how are we going to be differentiated and would that work?
So basically I was hired to create a hedge against the jobs strategy not being super successful. And that's what we did. So basically first thing I did is I brought over two of my former colleagues with me. And all we did is sit down and think through like what assets do we have, how can we monetize it and what would be the impact.
And we came up with this memo, which I hopefully I can find. It would be interesting if I still have it, of like top 10 ideas and then rank them in order of probability of success, et cetera, magnitude of success. And then we wound up pursuing a few of them actually. We would up pursuing at least three, like pursuing meaning live at least three of the eight or nine.
We may have-I don't think we ever tested the others. We didn't really do—to backtrack into your question a little bit more specifically—never really tested these. It was all first principles logic, like so analysis, here's the assets we have, here are the key ingredients, this is why this should work.
Here's the light. Here's the magnitude if it works. Versus like some AB testing, we didn't do any of that on these. So basically we wound up pursuing—what was called internally the low-end subscription product, i.e. not super high in recruiters who we charged a very high price point to—executive recruiters, et cetera.
So that was something that Matt Kohler was working on, and Josh Hellman, I know actually helped build, but that was the jobs product. And then a low-end subscription product was something for a normal person. So that would be like roughly $9.99 a month that a normal person could subscribe to that wasn't a recruiter.
And then we also wound up launching other versions of revenue, including targeted advertising. And the direct mail idea was there's lots of ways to target by geo on the internet, even at the time. There's ways to target by kind of demographics, male, female sort of stuff, age generally. But what makes direct mail magical is you can also target by occupation and income. And there was nobody to do that on the internet, and even today, there's basically not really great ways to do that.
The direct mail has been doing this since the 1950. So if you want to mail to a bunch of lawyers, doctors, or Dermatologists there's ways to do that. And, I never really worked in the direct mail industry, but my political background actually was very helpful because politics revolves around direct mail. That's the way to raise money and the best way to persuade voters in many ways is through direct mail because of its targeting capabilities.
So it was very obvious that some of that money should eventually shift online and there should be a scalable way to do it. Well, what's the unique asset we had on LinkedIn? We knew what everybody did for a living by definition. So that was the goal, basically replicate that targeting particular on the professional income side, capabilities of direct mail but with lower production costs, natural feedback loops, et cetera. And so that's what we built. But yeah, I may have to try to locate this and maybe publish this memo as well as my Day 1 memo I've kicked off.
Delian Asparouhov: [00:37:18] And did you actually, with this like sort of direct mail equivalent on LinkedIn, go out and then find the early sort of like, you know, customers that would actually spend their advertising budgets on there?
And then did it actually look like, you know, political campaigns, things like that? Or who were some of the—I think I understand the low end subscription, of course, that was just individual consumers. And so you just market to them versus the other one actually is obviously more of a B2B service.
And so what were you responsible for actually going out and finding those early customers for that advertising service?
Keith Rabois: [00:37:41] I can't recall whether we did or we just built it first. You know, the part of the more Keith Mantra of just build it and then sell it. I don't really remember. I'd have to actually debate this with like someone like Reed or Steve or David.
I can't remember the sequencing there. Again, it might've been logically driven of let's just build this because X, Y, and Z. You know, it was irrational to continue to do it through direct mail. We have this competitive asset versus everybody else on the internet, eventually over some time horizon that will shift. What actually wound up happening though is it didn't really work out that way in practice.
Where it actually wound up being very attractive was at the high end B2B customer segment of decision-makers at larger organizations. So in fact, where the product was used most successfully was-I wanted to reach a CIO. I have like some fancy security, something I need to reach the CIO at this organization.
Well great. I can find every CIO on the planet on LinkedIn. So it became more B2B, but very specialized version of B2B versus mainstream direct mail replacement in practice. And, so it actually didn't turn out to be—the infrastructure and the product was the same—but the use case actually turned out to be different than at least I think we spec’d out in the initial memo.
Delian Asparouhov: [00:39:06] Right, and so at both companies, you know, LinkedIn and PayPal you managed the corporate comms team, which we haven't really talked about. Was there a reason that sort of fit into, let's say like your scope, is it because of the background in like lobbying and speech writing, which are somewhat related to, you know, storytelling? Is that what sort of made you take it on and then sort of, how did you think about it across, you know, both companies? Was it primarily just partnering with outside firms? Did you build out large teams? Was it local, national, federal, or the particular trade publications that either, and you know, what were some of the, like, say major initiatives across the two?
Keith Rabois: [00:39:35] We can have a whole podcast on that topic of running comms teams.
But no, fundamentally the way I got started in this process with comms was- we really glossed over this but Peter's general approach to management of PayPal, certainly, as it manifests itself with me was to find projects and initiatives and topics he was unhappy with, and then when he was unhappy with the performance or, you know, focus, on a particular topic he’d basically throw it at me to solve.
So it started with the Mastercard/Visa problem, then the eBay problem and the treasury problem, and then there was some marketing challenges, and then he was frustrated with some other things and certainly frustrated with comms team. So basically they just became frustrated with our financial services teams, which is basically how we worked with partners in terms of business relationships, payment, and what the terms of those deals were with like Wells Fargo and things like that.
So basically I have this constellation of responsibilities that has accumulated over three years where it was basically just a by-product of Peter's frustration. So at one point, he was frustrated with communications, so it became my problem to solve. The first thing we did do is we immediately switched to the in-house versus outside, which is something that I highly recommend. We immediately fired our consultants. The guy who actually ran PR for me, Vince Sollitto, who is very good at what he does—he’s wound up running comms for Governor Schwarzenegger, then for Yelp subsequently. But at first he was frustrated that we were going to fire the external consultants—like ‘I need resources to go do this successfully if we're going to get rid of these consultants. So I need X, Y, and Z in terms of hiring.’ So then my job as the executive was to make sure we had enough internal budget basically, so that he can correctly resource the initiative. And basically by using my credibility to get the budget, then you can hire, or actually what we did is mostly reassigned people internally. So we found some under-leveraged resources, used them to fill out his team. Do you want to build a pretty sizable internal team that supported all comms initiatives, but then we have full the benefit once we have full control of it, like, so Peter was happy because you could see it was being closely managed by me and by Vince.
And that we can be quick, very agile, very responsive because we understood the business strategy very granularly, you could switch, you know, in a split second and be responsive to some macro thing or some micro thing. So it had a lot of benefits, but my responsibility as executive was really fundamentally first get the resources to be set up for success.
And then to edit really at the end of the day, the work product. But fortunately I had Vince quarterbacking the day to day stuff, and he is incredibly proficient at all the mechanics of running PR.
Delian Asparouhov: [00:42:28] And then at LinkedIn, was it basically just running sort of that same strategy? I imagine LinkedIn also just had much fewer let's say like, you know, comms issues or it was less important than it was at PayPal…
Keith Rabois: [00:42:38] It was more of an upside potential at LinkedIn where we could use comms as a way to drive marketing. Think of it as more closely akin to marketing than to traditional corporate comms. And yeah so at some point the same thing was true. There was enough internal frustration with our comms strategy. It wasn't being effective as measured by business metrics. So given that I had previously managed it, it was pretty easy to transfer it. And then in that case actually I had to change the leadership of the team, but then things worked pretty well. We were able to make a lot of improvements pretty quickly, and then, you know, once you’ve managed multiple teams and comms at different companies, you start seeing the gamut of problems and differences, and then know, obviously we did this with Square.
Jack is very, very capable of running the comms and certainly is very insightful in you know, the messaging and the value of contract comms, but mechanically, I ran the comms team for two years at Square as well. So you'd become fairly proficient at it. It's like any other muscle you develop proficiency, by doing it and seeing what works, what doesn't.
But each company had different opportunities as well as threats. And so part of the art is using comms as not a one size fits all solution, but to tie it to very specific business objectives. I mean, my overarching philosophy on a lot of these functions that are sometimes treated as G&A functions is all of their output in all their goals needs to be tied to very specific business metrics.
So whether it's government relations and lobbying in the initial conversation we had about Paypal, or comms, often the conversation was just do these things because someone told them they should be doing these things, but they're not directly linked to very specific business goals. And at PayPal, because we didn't have a choice because we're stressed all the time.
And at LinkedIn because you know, Reed is very smart, and at Square because in the beginning we had to be very creative. We didn't have a lot of resources at Square, until after we proved that various pieces of the puzzle would work. It was critical that we tie everything together and leverage all the G&A functions, even legal, to be part of the business strategy.
And that's mainly my overarching philosophy that I try to install at Founders is there is no such thing as like this part of the organization that doesn't get valued and measured and tied to core strategy. If so we shouldn't be doing it. If that's true, you shouldn't be doing it because there's other things to do with your resources, time, people, executive time and attention.
So you need to fuse everything together to be holistic. And that's, you know, that was probably the art of my role at PayPal is tying pieces that were probably a little bit too loosely affiliated with Peter's top level goals and making them incredibly tightly aligned.
Delian Asparouhov: [00:45:28] So from LinkedIn, you ended up hopping into Slide, which you touched on briefly before. And I know you always joke about it as your least successful startup despite getting bought by Google for almost, you know, 200 million. It seems like you guys had a density of talent, maybe not quite at PayPal's level, but definitely some pretty high quality talent there. And you've talked a little bit about like, you know, the platform risk and, you know, not digging into that, but maybe what do you think like, you know, sort of, cause the quote unquote, let's say like, you know, failure of Slide of not just like, what was the platform risk, but obviously you had talent, like why did that talent not see that platform risk or address it or, you know, figure out, let's say, a way around it?
Keith Rabois: [00:46:03] Great question. I think it's partially cultural. So yes, we wound up having, a fair amount of dense talent at Slide, as you might imagine. You know, Max definitely understands how to recruit talent, how to assess talent and assemble talent, you know, the key leadership at Affirm.
Square Cash is run by Brian Grassadonia who worked for me at Slide. The CEO of Postmates worked for me at Slide. Jared Fleesler, who is not bad in his own right also at Slide. There's Rishi who is running FutureFit very successfully, you know, worked as a PM at Slide. There's a lot of very talented people, let alone on the engineering side. Lots of great leaders on the engineering side. One of the best designers in Silicon Valley was our lead designer. There was talent all over the place. The biggest issue culturally was the company was predicated on what I now believe, and maybe even Max agrees, sort of an approach to building startups—which is bottom up data analysis. So what Slide was supposed to be was the future of entertainment built on social platforms and the way you decided the future entertainment in the Slide vision of the world was: you use data to decide what to produce. So users would kind of vote with their feet and you’d create that.
And the narrative to that argument was, 'well, you could avoid the blockbuster hit problem of some random editor or some random publicist, some random, you know, screen Hollywood person that green lights a script, and that becomes a blockbuster and there's no way to replicate it.
We would use data to decide what to produce and then presumably that was a scalable way to create success. Turns out bottom up building of a company, I don't think really works, but we were pretty committed to it. In fact, a lot of that talent, you know, joined because of that mission and, you know, sort of sex appeal behind that.
So I've got an engineering team there's philosophy behind this that is pretty cohesive. And I think that led to me sort of over-steering the next time—Square is very much a top-down visionary founder. Completely divorced from the data, like at the end of the day, Jack had a vision that world should have this, here's what it would do. Here's why it would be good. Okay, now we can make this successful.
Not like users are voting with their feet, telling us what to build, you know, et cetera. So I think I immediately was attracted to some of that top-down vision of the world after suffering through the bottom up vision. But basically we were not capable of fixing that. Because it was kind of, completely inconsistent theory of startups. So once you have a hundred people, I don't think you can easily rip out the foundation and say, 'you know what? We're totally wrong. We're going to have a top down vision. ' Also the DNA of the company was not set up for success in a top down or visionary way. I think you wind up with a different composition of people and functional organization, and there was just some tactical mistakes that we all made, myself included, being a little slow to react to some things, switched offices at the wrong time with the wrong layout, which magnified some of these problems. A lot of contributing factors, but net-net, it was an interesting three and a half years. Certainly a lot of learning. A lot of, team building, which was quite useful in terms of talent just being—off the cuff, like 5% of people, you know, that caught on to interesting things again in conjunction with me.
So not necessarily a bad experience, but I think from a company standpoint, from a vision standpoint, from an outcome standpoint, I think all that was a failure. You know, one way Max framed it many years ago in probably 2004, possibly 2005 was every company you built to be more successful than the last one. And if not, it proves you haven't learned anything. So to some extent, you know, we wanted to build something that'd be more successful than PayPal. We obviously failed at that. So by that definition it's a failure. Hopefully Affirm will fix that.
Delian Asparouhov: [00:50:24] Yeah. I think Max is doing pretty well at this point. He's going to, I think, create something much bigger than a PayPal or Slide.
Keith Rabois: [00:50:33] I think there's also a founder product, sort of matching criteria. Whereas Max in financial services is a perfect founder. In entertainment they're species of entertainment that Max is actually pretty proficient at and interested in. But he's certainly not a general purpose frivolous entertainment person. So I'm not sure the founding team was predicated for rebuilding entertainment, because they probably are in the video game space more so, but a lot of movies, pop culture.
In fact, for me personally, I had sort of stopped watching pop culture, like TV, movies, even music, to some extent between, let's say 2000,-2005. The first thing I did when I joined Slide was I started watching movies and TV again, because I felt like to be successful in an entertainment company you need to be completely attuned with what pop culture was and where it was going, but that five-year vacuum was painful to try to fix. So I was like binge watching movies and really got back into music for the first time in a while really having to invest in myself again, to try to catch up. And I don't think there's a substitute for real-world experience if you're going to build a pop culture company.
Delian Asparouhov: [00:51:51] Speaking of let's say like, you know, founders/executive team, you know, company fit. I feel like Square in some ways was like the perfect potential role to like design for Keith though, you know, had financial services components. It had some interesting regulatory components.
Top-down visionary CEO that cared a lot about talent. One of my favorite stories you've always told me about from there was that you refuse to let the product launch, until you had the ability to sort of like underwrite merchants instantly under the platforms that they wouldn't have to, you know, sort of wait.
And I think JPMorgan Chase was one of the big blockers. I guess can you talk about like, you know, sort of, how did you convince, let's say, you know, Jack and the rest of the company, that this was the right call, delaying the launch, how did you end up sort of getting JPMorgan Chase onboard? I feel like that sort of project and initiative was, so let's say indicative of everything else that you had sort of learned over the course of your career.
Keith Rabois: [00:52:38] Yeah. So truthfully Jack already had the right instinct on this. He knew that the right thing to do was we needed immediate gratification and we needed to control our destiny by making the underwriting decision ourselves. It was incredibly controversial internally, and obviously very difficult to persuade our partners to allow us to do this because basically the only way to plug into a Visa/MasterCard was through a bank, and so we needed someone like JPMorgan Chase to sponsor us and use their credibility, to allow us to connect to the plumbing of Visa and MasterCard.
So Jack had the right to answer. He just was unable to stitch it all together to get approval. And as soon as he explained this to me, you know, I immediately appreciated the value behind the statement in which was immediate gratification, you can imagine it's very difficult to do a viral loop when you can't literally get the product while someone's explaining it to you. So we used to have examples later where a taxi, you know, someone would see a Square in a taxi or a passenger would literally require a driver to sign up for square to pay for a trip.
You can only do that if you have immediate instant approval, instant underwriting and gratification, I once convinced my trainer in the middle of my workout to adopt Square. I sort of forced him to that. I refused to bring checks any more, but I had to be able to do the live, you know, setting him up while we're working out or it never would have happened and then he converted 80% of his clients to Square.
So he understood the magic and the distribution of that. The underwriting is at the end of the day, there's two elements to a financial service's success. Basically it's underwriting and distribution. So the immediate gratification is part of the distribution strategy, the underwriting, you obviously have to have control of the decision-making of who gets approved.
Also, when you're working through a large financial institution, their incentives are going to be reversed, right. They are going to be very risk adverse and arguably a startup wants to be risk positive. In addition, you have adverse luck. The more financial services, you know, the more complex it gets, but also the simpler it gets is you have massive adverse selection when you put friction in front of a lot of normal people, because the normal people will walk away. The fraudsters will be very motivated to go through the friction. This was a direct lesson from PayPal, which was, the whole key to PayPal was we basically washed out fraud by having lots of real users. There's two techniques to fraud. You can restrict the use of the application, but you wind up perversely with more fraudsters because normal people are like, 'I'm too busy I've got other things to do with my life, forget this stupid thing.’
The better technique is to reduce the barrier so that normal people would get through the system really quickly. Cause there happens to be more good people in the world than bad people. So again, every normal person doing all their transactions, yes, you'll have some fraudulent transactions, but they’ll be a small fraction of the total user base, which is mostly what you care about is the percentages.
Percentage of transaction, percentage of volume. So in any event from PayPal days, I sort of immediately grokked the value proposition of underwriting. The strategic benefit of controlling it first versus sending it to JPM and having to deny users that felt risky. So then the question was how to persuade them, and we basically kept waiting for launch.
I mean, Jack announced the company, I think. - well, he announced the company at a generic level December of 2009. In March of 2010, specifically described the product, raised money in December of 2009. But wasn't able to launch the product over the summer, going into the fall when I was recruited and joined, mostly because of this one issue with JPMorgan Chase and he was unwilling really to relent on it because he understood the strategic import of it.
So my first job really at end of the day, was solve this. Figure out, you know, either how are we going to convince them? Is there a 20-80% solution, like somehow or another, you need to launch a product right? At the end of the day, morale was pretty bad in the company actually, others had pre-launch fatigue and sort of using Jack’s phrase for this. Because until your users get to test your product and get that reaction and delight and the drive if this is a success for them, there's not a lot of positive feedback loops in the organization. And so one of the benefits of having worked with partners, like JPMorgan Chase before was understanding what their concerns actually were.
And some of it was just reframing truthfully, like using the right vocabulary. Sometimes choosing the wrong words drives particularly large conservative organizations crazy because you say X and they hear Y. So part of it was reframing the problem for them, and that actually led to a huge amount of success.
Some of, it was empirical. We actually did a study that showed the delta in what decisions we would make under our sort of algorithm versus theirs. And that it would have passed their credit committee. So there's various pieces of the puzzle. There's also a hedge really that almost no one knows about. Then what I ultimately decided to do was basically get tentative approval to do this and just start shipping, in very low doses of 10,000 Squares, you know, batches. And even though this was a temporary approval, just figure it out that once they started seeing business success in the volume, that they'd be less paying attention to the perceived downsides, unless we screwed up, unless you know, the risk is real. And so basically it changed the, upside versus downside equation. And we literally raised a series B from our friend Roelof at Sequoia without official approval to launch. And I don't think anybody knows this, including Roelof, but basically, we really hadn't yet really moved to permanent blessing until post series B. But we were shipping a lot of Squares by then.
Delian Asparouhov: [00:58:50] That's a fantastic story. I'll make sure that I send that clip to Roelof. Well, speaking of funny stories, I sometimes bring this up when I explain to people sort of why I got hooked into Silicon Valley, but in, I believe it was late July or early August of 2012, Square experienced for the first time, a flat or maybe slightly down month. And I remember you gave this presentation at all hands when I was an intern, there's basically going through like the cohort analysis of like, here's our annualized volumes.
Here's you know why for the first time we were experiencing dips, just because in August, there's actually like less transactions. I guess can you talk about sort of how data driven decision-making like you talked about not wanting to do bottoms up, but at the same time, obviously you rely a lot on data in order to be able to explain the underlying story of the company.
Can you talk to me a little bit just as this particular example, since it influenced me so much? Cause I was like, Oh my God, this guy's amazing. I want to have whatever his job is. Sorta how you went through that analysis, why it mattered so much, why you presented it at all hands?
Keith Rabois: [00:59:50] So several elements to that answer, I think Square's an interesting blend of design and data.
So one of the reasons why I think a lot of the progeny from Square has done well in terms of building future companies from Fair, Opendoor to DoorDash is there's a culture that married the product design driven-thinking for principles, design thinking—with data. And neither one alone, I think is sufficient to build a company from scratch.
But I think the combination of the two and having proficiency at both allows you to have an unfair advantage. So one of the values of data is, I always use the metaphor of a startup is really like producing a movie, and then you're creating a trailer, which is your marketing message, but then you're selling tickets. If you're selling tickets at a phenomenal rate everything is great.. If you're not selling tickets at a phenomenal rate, you need to measure different ways to report, to figure out what's wrong.
And so one of the ways you find out, you know, whether the ticket selling is going correctly, you measure it and you measure it in different ways. You just ambiguate data so you get early warning signals when something changes. Because the art is looking for anomalous data. It actually, interestingly enough, some of the best uses of data are actually on the unexpected upside.
So you see something that's working even better than you expecting. You're embrace it. This actually happened at Square. And one of the reasons why Square is the company that it is today is we noticed something very early. In fact, Jack actually noticed it first as we were going from roughly 10 or 11 beta users to like roughly 40 and then roughly 40 to 400, when we got permission to ship some Squares.
There was a consistent pattern of each day getting slightly more users and actually in a predictable ratio actually. And Jack would ask me this question and we used to sit next to each other with our desks, like next to each other, he just pointed it out on the dashboard, you know, why is this happening?
And I actually didn't know initially. And I thought about it for second. And it's like, maybe it may actually be the viral loop because when you see a consistent ratio of growth on a per day basis, you know, internally that you haven't changed anything—we didn’t spend any more money on marketing, we hadn’t had any PR during that period of time.
There was nothing that could be driving it because we weren't doing anything to drive it. So I had this like, thought that, well, maybe it's related to the percentage of Squares that are in the real world. And so once you have a hypothesis, then best use of data is actually to validate that hypothesis. So if that's true, you should see a certain ratio, and so I had, another colleague of mine do the math and turned out, it is a very consistent relationship between transactions of the day before and new signups the next day. And so that validated to me completely that this was a loop and the logic of the loop was, as people saw Square in the real world, 1% of those people would sign up for Square.
So then once we figured that out, then there's another element to what Jack noticed was actually a little bit more subtle was when we would have a PR spike. So occasionally Jack went on TV, or you know, get a high-profile interview or something, and so we would see a spike in signups, but what was most interesting about that was it would never recede all the way down to the prior level.
It would recede down like any PR spike, but it would recede to a new plateau height. So that's very unusual, actually in the history of startups. And then try to diagnose what was driving that was in fact, this viral loop. So we would get that spike of sign-ups that would push more Squares into the world. That would drive more transactions. That would mean more people would encounter those Squares. And so the new baseline rate of signups, would increase every time we do that. Which led to a business strategy for our, especially, comms of we need to get Jack on TV or Jack in the media more because it's not an identity project, it's actually a very successful marketing strategy.
And it actually led to me doing more interviews. So, Jack can basically do most of the TV and high profile stuff, and I would do that printed media because we wanted to get as many stories in the wild. We even had a launch strategy to attract media attention because it would basically lift our natural rate of growth every time we do this. So it was basically a very cohesive, but it's all started from Jack noticing this anomalous data, one or two pieces of it. And then me trying to triangulate what the hell was driving it underneath. So the example you talk about in summer of 2012 was for the first time this relationship between transactions and users to new users started to decay. And it happened fairly quickly, like over a month or two but there was no obvious driver, but it's very scary because this was the growth engine that propelled the company for two years to arguably billions of dollars of value. And all of a sudden the engine- the loop is no longer working. And most of the superficial sort of potential explanations just upon inspection didn't actually prove to be true.
So it drove me crazy. It obviously drove Jack crazy, because you don't want to run a company where your viral loop is like sort of messed up all of a sudden. So we kept trying to dive into what was actually driving it and how much we should be concerned as well as what we could do obviously, too relift it. And that led to this epiphany of, you know, here's, what's actually driving it. Which required a fairly significant set of analyses and led to my presentation to the company for two reasons. One, obviously the people at the company we're concerned about like a somewhat slowing down, growth rate, companies are about building momentum. You know, accelerating not decelerating. And secondly, we had this philosophy from the very beginning that the foundation of Square was going to be both transparent and that we would be mentoring and teaching people to build their own companies is a very conscious decision from day one.
The day I joined the Square, August 21st of 2010, we had a very conscious decision that we wanted to recruit and train people who were prototypical proto-founders in my mind or entrepreneurs in Jack's words and that we would mentor and groom them to run their own company. And so, we would spend a lot of time teaching things, teaching concepts.
So Jack would bite off particular topics and address them on our weekly all hands meeting. And then, you know, in so far as I can bite off topics and try to teach them, I would teach them. I remember Gideon, who's our initial board member from KP, saying I feel like he would attend our weekly meetings the first year or so after I joined, and I remember him saying, 'I feel like I'm going to grad school and like learning how to build a startup.' And we're like, 'yeah, that's the whole point.'
So that's one of the reasons why we would share, inseminate, and discuss that information at that granular of a level in front of the whole company. It's a high leverage activity using the high alpha management framework because you're basically teaching everybody like at the time probably 300 people then maybe 250 people, how to be better at their jobs and how to build companies better.
And so that investment of that time, you know, 20 minutes of presentation for me, or 15 and maybe 5 hours of preparation time in terms of slides is totally worth improving the performance of 200 to 300 people.
Delian Asparouhov: [01:07:21] Yeah, but I think those slides basically single-handedly convinced me to stop just wanting to be an engineer drop out of school, get into business. Cause I was like, "I have learned statistics from my dad. I learned how to build things. I did not know that you could wed the two and become very successful in business by being a really good statistician.'
Keith Rabois: [01:07:36] Yeah, well, there you go. It's all my fault. Don't tell Peter this part.
Delian Asparouhov: [01:07:44] Yeah. You get credit for me dropping out. So one of the other things that I feel like I really like admired at Square and I think they've done phenomenally well since then, is just like their willingness to just constantly have sort of like skunkworks projects. Like when I joined it was like the Square bar thing. And then that didn't end up working so they shut it down. And then, you know, the Square gift cards that didn't work, and we shut down and we like, you know, tried to do a second version of Square cash and that didn't work, and we kind of like shut that down. Like what do you think was, culturally different, let's say in comparison to PayPal and LinkedIn, which were both very successful organizations, but really at the end of the day now, you know, 20 years later, both of those are still basically in their first act, like it's just, you know, first core product. Versus Square you know, I don't know the exact way that it's being publicly valued, but like both Square Capital and Square Cash, I think are significant contributors of the market cap of the company despite being, you know, somewhat orthogonal to very orthogonal to the original, you know, core business.
And both of them have existed in prior versions of Square, but failed, like what made Square, willing to keep pursuing these types of skunkworks projects, despite even some of the very early failures and like, you know, distractions of some of the ones that didn't work?
Keith Rabois: [01:08:48] Yeah. I think so most of the credit goes to Jack. I think it's very founder driven, like in terms of willingness to fail, willingness to invest in innovation, a more generally design driven company. Like if you think about Apple is the quintessential scientific company. What does Apple have to do every one to three years? It ships completely new products from scratch.
And so that culture, that DNA comes naturally to designers, more so than to people who are like optimizing use in metrics that already exist. So it's like a big step function forward. So nowadays, there's a great book that I highly recommend people called “Loonshots." It's right here standing on the bookshelf right behind me. It explains how to do both, subtitled "How to Nurture the Crazy Ideas That Win Wars, Cure Diseases, and Transform Industries."
And it basically describes how you have organizations that leverages and amplifies current business while investing in significant steps forward-innovative steps forward. And basically at the end of the day, it's a structural organizational problem. It's actually something that Vinod Khosla who actually replaced me at the board of Square explained to me back in 2011. I remember driving down to his office on Sand Hill road from Square in the city to discuss this topic. And he basically convinced me you have to have separate organizations. He was like, it's impossible for an executive to basically be told you need to hit your monthly quarterly metrics and, you know, make those happen or, you know, everybody's going to freak out—and innovate from scratch. Like almost no executive can handle those two things. So you're kind of put in whiplash to try to manage them to do both. So you need to separate these, and you probably need somewhat different DNA on each team. You know, it's like, I always talk about football metaphors.
If you need to throw the ball down the field and you need to run the ball. And typically the people who are actually running the ball are not the same people who are great at throwing or catching 40 yard passes and even the blocking techniques that are taught, and the composition of your linemen is somewhat different.
So at the end of the day, you know, Vinod had a very strong philosophy about this, which reinforced Jack's philosophy that we needed to do both. And so we set this up. We did fail and like Square wallet did fail pretty miserably. Took another draft out of Cash App, which was rethought from scratch, but based upon some learnings from why did Square Wallet fail, and how can we avoid those mistakes? And then, you know, fortunately Cash App, which probably does account for at least half of the market cap at Square, did work. It didn't work right away. It took significant efforts to get it to a place where it was working. But it's very much a top down philosophy from Jack and the board Vinod about investing in high asymmetric upside bets, and staffing them with talented people with the right DNA to potentially accomplish the goal.
Delian Asparouhov: [01:11:51] Well Keith, thanks for coming on the podcast today. It was super fascinating to talk about your operating history, as opposed to just, you know, you know startups, yeah. Thanks so much for coming on.
Keith Rabois: [01:12:05] Sure, pleasure.
Delian Asparouhov: [01:12:06] Thanks for listening everyone. If you'd like to support the podcast, please sign up for a paid Substack subscription, which we use to pay for transcripts, mics and other improvements. If you have any comments or feedback on what kinds of questions I should ask, who should come on the show or anything else, please do let me know. Have a great rest of your day.