00:14:00 · David Senra
[00:14:00] S0: I wanna tell you about the presenting sponsor of this podcast, Ramp. I have been reading a lot about SpaceX lately. SpaceX is one of the most valuable businesses in the world. And one of the main themes in the history of SpaceX is constantly attacking and questioning your cost. Ramp helps many of the most innovative businesses in the world do exactly that. The median company running on Ramp cuts their expenses by 5%. And one thing SpaceX has demonstrated is that a religious dedication to controlling costs can help actually increase revenue because you can pursue opportunities you couldn't otherwise. And we see that in the ramp data too. The median company running on ramp also grows their revenue by 16%. So when you're running your business on Ramp and your competitors are not, you have a massive competitive advantage that compounds over time. Ramp is the only platform designed to make your finance team faster and happier. Many of the top founders and CEOs I know run their business on Ramp. I run my business on Ramp, and you should too. Go to ramp.com to learn how they can help your business save time, save money, and grow revenue. That is ramp.com. Okay, but go back. How do you identify talent when the talent doesn't have experience? Have you done that?
00:15:10 · Luca Ferrari
[00:15:10] S1: I think you can think about life in general as, when it comes to people and accomplishments, you control certain inputs. You know, how much you work, what you do when you work, for example, to make it super simple. And then there's a bunch of elements you don't control, called them boundary conditions, sometimes knowable, sometimes unknowable, sometimes fixed, sometimes shifting. And the combination of those ingredients leads to outputs or results or call it what you will. Outputs results are fairly easily observable. Sometimes there is a gigantic amount of inputs and that go into achieving a certain output and those inputs go into it for a long period of time. Whereas the output can be very simple. Know, you, the company, achieved a certain amount of revenue, whatever. You won a certain tournament, you know, no matter the field. And so it's a lot easier to just look at the results, the outputs. It's the convenient, sometimes lazy way. But life typically in most pursuits, most endeavors is so complicated. The amount of inputs, the amount of people contributing different inputs, amount of variables you don't control those boundary conditions are such that if you just look at the outputs, sometimes you get a massively distorted picture of what the person contributed. A lot of it could be luck. A of it could be actual human performance, but not by that individual, someone else who just you, you know, you failed to know was involved, maybe wasn't as front facing. Now, the more extensive someone's track record is, the more results likely correlate with actual talent. Take an investor, you can get lucky one year, two years, four years, but unlikely thirty years. I don't think anybody would question that Warren Buffett is almost certainly insanely good at investing. You can never prove it definitively. It could have been lucky for decades, but that's astronomically unlikely. Right? But if you find any hedge fund who delivered 50% performance in any given year, it could easily be they got lucky with two stock picks. Maybe they were... Those were terrible ideas. Maybe they picked them for the wrong reasons. But, you know, whatever, the boundary conditions changed and they made a lot of money.
00:17:22 · Luca Ferrari
So when you pick someone who's very experienced, decades of of work, probably you can get away with just looking at the results, maybe some reference calls, very likely you'll get a reasonably accurate assessment. But if you have to pick someone who hasn't even graduated yet, or maybe he's been in the workforce for a year or two, then you don't have the luxury of using this, like the sample is too small. Exactly. You need to find different ways. Something we do is making extensive use of testing. So we develop tests that people go through, that we have found over time proxy pretty well their mental capacity and faculties, for example. Have over time developed we've really built a science out of studying people's track records, including academic records and personal projects and similar things, whereby in some of the applications we identify over 100 different signals. And through those we predict their long term potential. It's not entirely dissimilar from what a stock trader who is algorithmically oriented would do. More quantitative hedge funds they would identify and test hundreds or even thousands of signals. Many of these are only very marginally predictive, but in aggregate, they make you predictive enough that you can succeed because you're just better than others. So we have all these signals, some of which are completely obvious, you know, someone's GPA tells you something, you know, higher GPA is better than lower GPA. It doesn't prove anything. It's not definitive, but it's a good sign. Some are more subtle, we keep working and investing scientifically in identifying and and measuring these signals. So over time, I think we have developed a competitive advantage in finding people who, despite their minimal track record, are very likely to excel.
00:19:08 · David Senra, Luca Ferrari
[00:19:08] S0: Okay. So this is one of the notes that we talked about at lunch, and this is one of the things I text you, where you essentially said you need... A lot of other companies kind of... Don't really... Do know how many companies have brilliant people in HR? And you're like, no, you actually need brilliant people in HR. You said that you've... You can make hiring a science, that you had at the time a team of like 50, and you actually, these are engineers. These are not typical people that you find. Right? And then you say, you compare one to several years of performance, the signals from their CV, which is what you just described, or these 100 signals, signals rather, and that you centralized hiring and firing. And I guess these are you call them talent managers, and they're in charge of both entries and exits. The hiring and firing is centralized for all the companies that you own. [00:19:51] S1: Yeah. It's very unusual. And by the way, the signals are not just from the CV, but they could be from email exchanges with our recruiting team. From the tasks we ask a candidate to go through anything, really. Anything qualifies as a signal, we just care that it's predictive. It doesn't have to be something that, let's say, intuitively immediately makes sense as long as we can prove it's likely not a statistical fluid production. [00:20:15] S0: Give me an example of that. I'm a little confused.
00:20:16 · Luca Ferrari
[00:20:16] S1: Well, I mean, for example, one of the qualities that we value in people is... Because what we do is such a team sport. Need to be somewhat collaborative. You don't have to be the nicest person in the world, you know, like, but if you are arrogant or dismissive of others or just an asshole, that that doesn't work typically. Unless you're a freaking genius, we might, you know, accept it occasionally, but it's almost nobody is. So for most people, you need to be nice enough. And we find, though, that when they're interviewing even assholes, especially because these are smart assholes, because they've already passed the more cognitive oriented test, they're generally quite nice. Because they know that if they come across as super arrogant, they're not gonna get an offer. We found that whether the interviewer felt that the interviewee was, you know, open to criticism and reasonably pleasant to talk to, wasn't a good predictor of whether they were actually collaborative on the job. So we have a role, call it like a, it's almost like customer support. People help you with more the logistics of your application process of scheduling interviews. So it's more of a support role, which clearly does not come across as in any way connected to the final assessment. How people interact with those is a lot more predictive of how they actually are as human beings. And so we found that people who were curt and sometimes, of course, even disrespectful occasionally, that's rare. Ultimately, that predicted poor behavior in a social context much better than how they interacted in an interview. You know, again, one of hundreds of signals in and of itself is not definitive, but it helps form an accurate picture at the end of the day. And that that would be for collaboration, but then we would have others for hardworking attitude, whether you're creative, whether you're logical in your thinking, perseverance. Identifying what's important is not rocket science, you could imagine what's important to performance. The difficult part is spotting it through these subtle signals.
00:22:16 · David Senra, Luca Ferrari
[00:22:16] S0: And the interesting part is that [00:22:17] S1: you just said, how many companies do you own right now? We bought a little over 50 businesses over time. [00:22:24] S0: But you could see, if you own 50 businesses, there's multiple different ways. You're a conglomerate. The hiring could be pushed down to the actual individual company level, and you're like, no, no, this is so important. And the talk I had Almost all matters. I was going say, maybe the most important, so you just said it's almost all that matters, it is the most important. So I'm going to centralize this. Are you also then the centralization allows you to kind of take the insights that you learn from one of the 50 and disperse it to the other ones? Is that what happens?