Keith Rabois on How to Find and Grow Talent

We're here today with Keith Raboy who was an early executive at LinkedIn, PayPal. He was COO at Square. He's been an active angel investor and investor at Khosla Ventures. He's also chairman at Opendoor here where we're filming. Keith, we're really excited to have you here today, so thanks for being with us.

Pleasure to be with you. Great.

So today I wanted to start by talking about How do you identify talent? And so, you know, whether that's either hiring from the outside or finding people within a company for a new role, what are the things that go into looking for what might be a good fit for a role, a talented person, whether it's for founders when you're thinking of an investor or thinking about joining a company, whether it's for managers or individuals, just how do you think about finding good talent?

Well, I think there's two differences. Internally within an organization, it's a lot easier because you have a lot of data points. externally identifying talent based upon an interview or very random or short compressed interactions is very difficult. But it's one of the most important things a founder does and an executive does to build out his or her team. Peter Thiel taught me this a long time ago in 2000. My first week at PayPal, we went for a jog around the Stanford campus. And he explained to me that basically if you were going to succeed as a startup or succeed as a leader, you needed to be able to find people that were undiscovered that the rest of the world didn't know about. Fundamentally, if someone has a lot of data points on their resume that suggests they're amazing, people, organizations like Google, Facebook, Netflix, Amazon are going to try to attract them and they're going to pay more money than a startup can ever afford. So you have to be able to find people that organizations like that don't know how to process, they don't know how to evaluate, they don't know how to assess. And that's a skill in and of itself. It's a little bit like drafting elite athletes out of college to play baseball, football, basketball. that's what you need to become proficient at. So the way you do it is, I think you have to first decide what kind of role you're hiring for. Are you looking for a value creation role, like i.e. like create more value from scratch, like 10x more value, or are you trying to preserve value? So you're already doing well and you just want to make sure nothing gets screwed up. Those are very, very different profiles. The hardest one, of course, is the 10x value creators.

And that's where all the art is and that's the magic of startups. And so, you know, you talk about when, For small startups, it's so critical to have every hire be good. But at the same time, you need to find these sort of undervalued assets because you're competing with Google and Facebook and they've got better salaries and better benefits and they've got more cachet. And so how do you as a startup, when it's both extra important that your 50th employee be great, how do you do that? How do you attract those great people? What does it take?

Well, I think the selling point is the opportunity, the opportunity to grow, the opportunity and challenge. to run something yourself, the missionary zeal of the organization, the degree of flexibility, the number of people that can kind of tell you no, usually is pretty small in a startup versus in a large organization. There's 10 to 50 people that can easily veto your projects and ideas. So I think that's part of the selling point. But the evaluation and the assessing, you need people that are ready to run that really want the challenge. There's some people who learn, like sort of like to learn to swim by being thrown in the water and kind of drowning a bit. And then there's other people who need like, a lot of instruction. Basically, you want the people who like kind of enjoy learning to swim by being thrown in the water. And that's one of the filters. The second thing is people ultimately you need are tenacious. The tenacity of going over a wall, under a wall, through a wall, making friends on the wall, figuring out why the wall doesn't matter. That's the core skill that most startups need. And you can usually find that on someone's back in someone's background from things they did in college, high school. They sort of have that baked into their personality. And you're only really extracting that from an interview. It's a little harder to tell sometimes from a LinkedIn profile, but once you meet somebody, you can find that relentlessly resourceful characteristic. The second thing you're looking for is sort of just pure intellectual horsepower. Most companies are solving problems that other people haven't solved and can't solve or don't think they can solve. And sometimes that comes from pure insight married with tenacity of not giving up. So it's a bit of the perseverance and IQ married together. If you have both of those propelled by something to prove usually, is the best possible formula. That's what you're really looking for when you're interviewing someone is signs of that, signs of unique insight, signs of something to prove, and an incredible resourcefulness.

Got it. Those are very much aptitude based, which I think it sounds like one of the things you can essentially do is find these diamonds in the rough based on aptitude that others may not have recognized. What about the cases where you might need this sort of more experienced person? How do you decide I'm going to make an aptitude based hire here versus like, I really just need somebody who's going to be experienced.

And when do you make those kind of trade-offs? So generally, I believe in the people with aptitude, not experience. For example, at Square, we intentionally hired almost nobody who had any payments background or any financial services background. I think there was between three and five people in the organization that did out of an organization of now thousands. At PayPal, I think the only person who knew anything about payments as far as I can remember is our general counsel. And that's actually a good point. There are some roles where experience is valuable. So when you're in the risk reduction sort of phase or the risk reduction role like a general counsel or CFO sometimes, you're looking for someone with experience. But if you're trying to create something, forge something from scratch, experience is often a handicap. You can also benefit from the experience by hiring consultants or advisors or interviewing them. So you get a lot of the benefits in the first 30, 60, 90 days of working with someone who has industry expertise. For example, at Open Door, we allow you to sell your house in 30 seconds online. We don't have very many people in the organization of about 125 people that have actually done anything in real estate before. In fact, probably half the company has never bought or sold a house.

That's a great framework. So basically the way you could think about it is when you need upside, that's when you can hire for aptitude. And when you're protecting downside, often you want to hire for experience. Maybe one last question that I have about hiring. What's the way that you sort of look back and sort of refine your process? So this is whether you're either investing in companies or passing on companies and then looking back at what that process was. Or I know at Google they've got a very rigorous process about looking back at their hires. How do you recommend people look back at that process of hiring and deciding when they made a good decision, when they did something wrong? What's the way to refine yourself on that?

Sure. So I think hiring is itself an independent skill that you get better at by doing more of. So sometimes people want to take their best engineer and have them interview all the engineering candidates, which may work. but that's not exactly the same. Shipping code, high quality code is not exactly the same as assessing engineering talent and the potential to create disruptive innovation. The same thing is true of designers or business people. So I think developing a skill in noticing who in an organization actually has predictive value is a really important thing. Amazon has this concept called bar raisers, which are people who have historically proven to be very proficient in evaluating candidates, and they have to approve basically all offers, and they can reject an offer sort of by themselves. so I think what you should do is on every hire, you're kind of getting data of how accurate are you, and then look mostly at the anomalies, either the great hires or the mistakes, and figure out more of the characteristics, the common denominators in those cases, and try to either edit those out or refine them in the future or find more of the stars. I think that's an experienced process. It can take years to get better and better out because it depends on how many people the velocity of hiring. If you're hiring hundreds of people in a year, you can get better at it fast. Usually you're hiring a lot less than that, especially by team and by executive. So it takes a long time. It can take an entire career, which is kind of unfortunate that by the time you get good at it, you may be done with it. But it is an experiential product. And I think you definitely find some common characteristics of what mistakes particularly made, although it's important not to try to be a zero defect hirer. I think if you're looking for upside, people with potential, you're going to make some mistakes by definition, because otherwise you're too conservative. Because some of the people, some of the most interesting people that create the most innovation have a screw loose one way or the other. And if you overreact to that sort of screw loose, you're gonna miss some of the most important, most talented people sort of on the planet. So you have to be willing to make some mistakes. Of course, you have to correct those mistakes and not procrastinating on fixing them is an important skill in and of itself. But I always look back at all the best hires and try to figure out what was in common about them and then the mistakes. What were the kinds of things I missed that I should have picked up on? Or maybe just what's the baseline of mistakes and is that good enough?

I've heard people say things like if you're as an investor, if you're making money on every single investment, you're doing something wrong. I've never actually heard it that way with hiring, but I think that sounds really, really good.

You need to apply both the false positives and the false negatives. And interestingly enough, very few companies track the false negatives very well. So the companies that they didn't-- the people that didn't hire that could have been great. Those actually arguably are more important than the mistake in hires you make. But very few companies are religious about that. The CEO of eShares wrote a very interesting and provocative blog post about this that really nailed the concept of tracking the mistakes that you didn't hire. just as equally as passing on Uber if you're an investor.

Okay, so now you've got all these people at your company and arguably one of the highest leverage things you could do once you've got your team is to make your existing team better. And so the concept of sort of developing talent, particularly in these cases where you're hiring a lot of people quickly and people are, like you mentioned, they're doing things for the first time. How do you kind of, in your own experience and when you're working with startups, how do you recommend companies develop talent effectively and not just through the haphazard experience or is it just you grow and you learn and you swim or is there other kind of systematic ways that you can develop talent predictably?

Well, I don't know how predictable you can make it in a startup. A startup by definition is somewhat chaotic and shouldn't have too many processes because processes tend to freeze in place lots of things. So when things are working perfectly, you want lots of process because you want to freeze that, bake that, sell it sort of. But in the beginning and even in media startups, you're trying to innovate all the time and so you want a degree of flexibility that feels a little chaotic. And in fact, you want to probably filter for people who can thrive in a somewhat chaotic, less predictable environment. The first thing you look for is who shows signs of promise that are extraordinary. And there's ways to pick up on that. One way is who do people in the company go to talk to? That they don't have to. Sort of so you can watch like whose desk becomes popular. And when you see a distillation of people approaching somebody, particularly they don't report to that person, it's probably because they find that person to be insightful and helpful, and that's a great sign. So one thing that's also about having an open office, which I highly recommend for a long time, is you can actually monitor this and kind of see the ebbs and flows of people and look for common characteristics there. Second thing is what you want to do, and David Sacks taught me this in the beginning, is constantly expand the scope of responsibility for every single employee, and everybody will break at some point, but you want to keep pushing the envelope until they can't discharge that level of responsibility successfully anymore. So you can start with something trivial. You know, I've talked about before about giving people like ordering smoothies and seeing if they can order smoothies correctly so that they're delivered on time, they're fresh, they taste good, and then expand the scope from there to something a little bit more complicated to something a little bit more complicated.

And just that way you're maximizing everybody's potential. So when you've got, say, you know, you've got maybe the sort of employees that are in the middle of your bell curve and you've got these exceptional employees, Do you generally recommend putting more effort into sort of improving the whole pool or do you take your stars and put extra effort into those?

So the best thing to do is actually to double down on your stars. There's an old concept from golf, kind of an introductory chapter in a kind of an interesting book that asked this kind of amusing question about Tiger Woods. And the question is, somewhat apocryphally, how often does Tiger Woods practice his sand wedge? the answer is never, never ever ever has Tiger practiced a sandwich. The reason why is he spends all his time avoiding getting in the sand. And to some extent that's what you want to do is figure out where the high lever, high leverage activity is. And if you believe that you have people who have 10x value creation, improving their abilities has magnificent effects versus trying to fix people that aren't going to ever create 10x value. So the best thing you can do is double down on your stars, even though it's counterintuitive because there's a lot of management drag that goes into helping the people that are suffering and struggling versus, but that's not going to create most of the value for the organization. And your job is really to create impact, impact for the organization measured, you know, in terms of whatever KPIs the organization settles on. And the best way to do that is to make the top 10 or 20% more effective and then move more people over time of the bell curve shift into this sort of pool if you can. So the bell curve is not totally static. It's a concept that feels static, but you can shift the sort of the distribution so there's more people in the top, you know, sort of category.

Yeah, and relatedly, I think, you know, a situation that I've talked to a lot of founders about or just general, you know, through sort of the ecosystem, a problem that you see people run into is you're growing really fast and you've got somebody who is doing their job they're not doing fantastic, but you've got so many other things to worry about that you sort of leave somebody in that role. And your choices are you can either move them out of the role or you can sort of part ways with them, or you can just kind of let them go and avoid, you can kind of let them keep running in their job and sort of let them, you know, sort of just do, do what they're doing and not have to go through the recruiting and the hiring again. And you see this trade off happen a lot. And it's very easy to say that, you know, we want to only have top talent here. And, you know, there's the Netflix thing of adequate performance is met with a general severance package. But it seems like in practice a lot of times it's very difficult to actually make that decision when you're faced with, man, we're growing fast, I've got a lot of other fires burning. How do you encourage people to keep moving fast and not letting those imperfections stop them, but also maintaining a high.

Enough bar and sometimes making those trade-offs? Yeah, so you have to triage all the time. If you're a founder or CEO, there's always going to be firefighting and crises, and you have to figure out which are the highest leverage activities for you. and depending on what that role is and what the initiatives that person is responsible for, it may be a very high leverage activity for the organization to upgrade that versus in some cases it may be a moderate leverage and you may have to defer that. One way to think about it is kind of a metaphor from sports is most managers in baseball usually make the mistake of leaving their starting pitcher in a little too long versus calling for the relief pitcher. And that's kind of how you have to think about it. The starter got you so far, but you need a different skill set to go forward. And there are people who are awesome at that next phase. Most relief pitchers, especially closers, are incredibly proficient at what they do. And they close like 90-some odd percent of the save opportunities they have. So you can get people that are incredible and knowing, being willing to do that, understanding that, and being proactive about it, the hardest part is being proactive. Once things start showing up, they're broken, the team's broken, morale's broken, and metrics are missed. It's really hard to start fixing that because there's a lag in terms of when you can find somebody, identify them, close them, get them up to speed. So the best leaders are actually thinking two or three steps ahead and can see when this team, this person is going to start struggling and are already recruiting, sourcing, interviewing, and assessing people so that they have somebody ready to step up to the plate when the time's right.

And so tactically, we kind of have two ways that we can improve sort of the development of our talent. One which you mentioned is putting people into stretch roles and letting them just grow through that experience. And then the other obvious one is kind of giving people good feedback. So there's sort of this coaching element and then there's just throw you in the deep end. So on the coaching element, I generally believe that people are pretty bad at giving good feedback for various reasons. I think you see people are unable to give constructive feedback clearly or when they do it becomes ad hominem. How do you give good feedback? We all know it's really important but most people are really bad at it. How can people be better at coaching others, whether they're people who report to them, whether they're people who are peers, whether they're giving feedback to their own manager. How do people give good, effective feedback?

So I think the frequency of feedback matters and the timing of it matters. People tend to procrastinate and tend to defer it, mostly like writing an essay in college, because it takes a lot of work to do well and the news isn't always pleasant and therefore you tend to hope it kind of goes away. But it almost never does. Very few times when you want to give someone constructive feedback, Does the problem remedy itself? Does the person learn how to do something the way you want it done in the first place? Just sort of suesponde. The best way to do it though is tied to whenever the object of the feedback is most relevant. So if someone does a presentation, for example, or makes an argument about a certain initiative, it's better to tie the feedback to that specific problem because then you can help their brain adjust oh, I can see how I could have done this better and why that would matter versus a generic statement three months later that, you know, you're not analytically rigorous enough, blah, blah, blah, your slides aren't well-designed, all these kind of things where you have something tangible and concrete to point to, that's actually a better way to learn. Feedback can also be done through osmosis. I think other colleagues, it's not just the executive or the person you report to that provides feedback. It can be very well-rounded feedback, but I think also an attention to what's the goal? Knowing where you want the person to get to helps you construct the feedback. So if you don't understand the destination, it's very hard to give people feedback. It's a little bit like a startup. When you see an early stage company, if you know what the company is supposed to look like in a year when they're raising their Series A or their Series B or their IPO, it's easy to help the CEO triangulate to where they need to get to. So understanding where this person wants to get to, where they need to get to, helps you recast the feedback into a constructive line that they can draw for themselves. Yeah, that makes total sense.

And particularly as the company grows, I think when you're saying here's the things that we want you to see you accomplish, as a company grows this obviously gets very challenging. Things get chaotic quickly. You grow past a certain point and you no longer know everybody's name and this is this weird dynamic that startups go through where at 50 people you might know everybody's name and 100 people you don't and at 200 you certainly don't. So as you're sort of going through this, what's the way to sort of keep, how do you keep people focused and aligned when there become so many competing priorities? You've got multiple people jumping onto specific projects. Like how do you make it so that, you know, it's more like the early days when there was, you know, one goal, one team, one dream and then later it's, wow, there's so many things we could do and we're all gonna splat into a million directions. How do you, how do you drive.

That sort of focus and alignment for people. I think you need to subdivide. You need to subdivide teams. You know, Jeff Bezos talks about pizza teams, and I think that's a good idea. I like small, vertically integrated teams that can have accountability for specific set of metrics and allow them to focus on it. So they have ownership mentality. You can see who's doing well, what skills cap, what skill gaps are existing in the organization. Those people go home, and when they're showering or when they're dreaming at night, they're thinking of problems. So I think subdividing into accountability is important and then rolling those sub teams up together to macro goals is pretty important. And that requires the founder and CEO to be very thoughtful about organization design, about which sub teams can be created that can act mostly independently of the other teams and they can therefore run as fast as possible and create a lot of value as fast as possible in parallel with a lot of other teams. And that's one of the harder challenges. Every organization every organizational design decision has some trade-offs. There's trade-offs in skill, there's trade-offs in learning, there's trade-offs in people's political standing, there's a bunch of these kind of important ingredients. But ultimately, getting a healthy organization allows you to accomplish a lot very fast, as well as allows you to expand the scope of responsibility for your best people and identify which new people you need to recruit because it's very transparent as opposed to a monolithic organization where you can kind of hide and mask people. You're very exposed. So I think that that's the best process. Secondly, unfortunately, isn't much software. You would think like in Silicon Valley that you could adopt software and tools and data. There isn't a lot out there. I mean, that's one of the benefits of Lattice is that you can actually help solve this problem. But there really isn't. I mean, a lot of this is still done best with one-on-one meetings every week or every other week. You know, I subscribe to the views of Andy Grove and high output management. Everybody should read that book. You should hand out copies of it. I mean, it's really old school. The book was written in 1982, but it's still better than most of the tools and software that exists. So I think you have to go back to first principles, which is managing by time and attention. There's also another concept you can take too far, but basically focus, which is allowing most people to narrow their scope, allows them to improve their execution. Peter used to have this philosophy, Peter Thiel at PayPal, that everybody was allowed to do only one thing. and he would only talk to you about that one thing and your review at the end of the year would only mention that one thing. And so I think that's, you know, an extreme version, but it helped avoid the distraction of multiple initiatives and I'm going to solve a second tier problem because I know how to do it versus the important ones that are really critical for the organization. That's one version. Apple, of course, was notoriously focused under Steve where they absolutely avoided doing a lot of things so they could do one or two things extremely well. I think that's actually healthy for a startup. If Apple with all of its resources, all of its brand, all of its cult following can only do about one thing simultaneously, it's pretty crazy for too many organizations of 100 people to be doing multiple initiatives. But you see it happen all the time because there's this kind of myth of option value and option value is a really dangerous concept. More often than not, life is path dependent, not option value, not option value dependent. So I think picking one, two or three three things that everybody's going to be successful at is more important than trying to do ten things and watching what works.

Absolutely. I also want to go back into something you touched on about organizational design, which I think people actually don't talk about a lot and it's actually fascinating. And I think from the experience I've seen, a lot of times the way an organization ends up being designed is somewhat by happenstance where you hire in maybe leaders and they build a team under them and sort of that's your organizational design. I think one of the trade-offs I think a lot about and when I talked companies about their overall goal alignment and how they're structuring things is one of the natural trade-offs with aligning a whole company is that you build some natural competition between different groups. So whether this is between sales and marketing or between sales and engineering or between, say, marketing and finance, you end up giving different metrics to different groups and then those can be in competition with each other, which It doesn't mean obviously that having alignment and metrics is something that you shouldn't do, but that's a trade-off that people need to be thoughtful of and address. And so it's an interesting point that you bring up about making sure that you're thoughtful about the organization design. Because I think there's not a lot of great frameworks. And I don't know if the answer is there are these five frameworks and you should pick the one based on which you are, or there's a million different frameworks and every company just needs.

To build their own. I think it's a little bit custom brew, although there's some conceptual frameworks you can leverage. The reason why it has to be somewhat custom is every market opportunity is somewhat different and where you stand in that market opportunity is very different. So for example, if you're trying to capture a market and it's not competitive but you're the first one, you can do things differently than if you're in a hyper competitive market and competing with someone like Travis and Uber. It just requires a different organizational structure or a different culture to be successful in that environment versus you're the first one pioneering something new that no one's ever thought of doing before. Obviously, if you're building a SpaceX, you're going to emerge with a different engineering culture, a different precision culture than someone who's doing photo sharing. If you lose someone's photo, it just doesn't matter sort of thing. So it depends on what you're trying to do, where in the market, what kind of product, all those kind of things matter. It's not really matters what your talent base is. I don't think you can take from scratch, like first principles and say, oh my god, I have an open whiteboard and I can get anybody in the world I want. That's not true. You can't write a spec that says I want a third baseman who's going to bat 300 with 30 home runs, 147 RBIs, and win the Gold Glove, because there's only one of those people in the world. So I think you have to figure out what skills, unique advantages you have in at least involve that in your considerations of what organizational design to use. And then it evolves over time because as you add scale, as you add zeros to your headcount, things that would have worked very well through a lightweight process, highly osmosis oriented process will not work with a thousand employees. Like as you pointed out, at 50 employees, if you're the CEO, you should know everybody's name, you should know what they're working on, and you probably should be able to know how well they're doing their job. At 100 employees, you'll probably lose the ability to calibrate how well each of your employees is doing their job. You still should know their name and what they're working on. 150 employees, you may start losing the ability to know their name, but and once you lost their name, you're certainly not going to know what they're working on. So at some point, you have to have processes designed to reflect, you know, the human brain and limits of the human brain. And so you have to change all the stuff all the time. There's always going to be those some seams in the organization. The question is where to place those seams, where there's potentially some inconsistency in metrics, there's some inconsistency in style. No organization is perfectly consistent. Even like Apple under Steve Jobs, if you'd actually looked at their HR software, it was all It was clear that Steve had never logged into HR software because the design wouldn't have passed any level of scrutiny. So every organization makes compromises, but knowing where those compromises are, placing them intentionally, and then moving them and shifting them.

As things evolve is pretty critical. Yeah, it makes total sense. All right, there's one last topic I want to see if we can squeeze in here, which is the topic of company culture and specifically transparency. I know this is something that you and I have talked about before, and this is something that I think a lot of people want to be transparent. They recognize the value of transparency in a transparent culture. And then when the rubber meets the road, they sometimes decide, well, we're not gonna share that thing. Or, you know, let's leave that one out. And so I think in the middle things, it's very, it's obvious, you know, this is something we can share. When it's hypersensitive, people sometimes don't necessarily make the decision that they know, you know, logically they should make. then maybe sometimes you have cases where people overshare on the small detail then, but maybe that's a little bit less interesting. So the question I sort of have is for the well-meaning teams that want to be transparent and then decide not to be, what's sort of the miscalculation that they're making? Like why do they end up not sharing everything with the company and not being an open book? They're usually afraid, it's usually fear.

So the kinds of things that don't get shared in those cases are like how much money is in our bank account. you know, what was our revenue? Things that are, let's say, lower than ideal. And they're afraid that people will panic, that they might look for new jobs, that they'll be distracted. But the problem is the countervailing compromise is if you don't give people enough information, there's two things that go wrong. One is they're not just not going to make the right decisions. Like one of the reasons why CEOs get frustrated more often than not with parts of their team is people make decisions that are not in alignment with what what the CEO wants. The biggest reason for that is usually lack of information. The rest of the team doesn't have the full picture. And so what you want are tools, software, meetings designed to spread that information so everybody's working from the same sort of plate. And then therefore you can expect everybody to make the same caliber quality in aligned decisions. Second thing that also goes wrong is you breed a little bit of distrust. So if you want people to trust you, you have to actually earn it. and the more information you give people, ironically, the less an organization usually leaks to the media. It's sort of perverse. Jeff Weiner actually, I've LinkedIn pointed this out, I think most starkly, which is if you tell people, here's a lot of confidential information, I'm trusting you with it, it's really important that we don't share this. It tends not to get shared, but if you hide it and then people discover it, they tend to leak it. So trust plus decision making quality are the two most important reasons. that I believe in transparency. As I've talked about, I believe it's sort of radical transparency. I believe you should share your entire board deck with the entire company right after the board meeting. I believe that the only real question is do you share compensation information? And I still believe that someone will figure out how to do that successfully. But I understand why people are sensitive to that.

So maybe except for some of these people issues, it seems like on things like, you know, revenue is a little bit lower than we wanted or cash is lower than we you want it to be on those types of things, you think people are generally just not.

Making the appropriate trade-off?

Yep.

Actually, I think complete transparency will lead to better problem solving, more ownership mentality by your team. Basically, everybody in these companies is a shareholder and they have the right to the information. So I think sharing it is mostly upside and very limited downside, especially if you frame it appropriately. That's great. All right.

Well, I think that's all the time we have. So, Keith, thanks a Tun for being with us. I really enjoyed the conversation. Thanks again.

Sure, pleasure to be here.