The Exchanges

Q Anyway, enough about Jules. So to give listeners a little bit of context on yourself, can you just give us like a 55 second overview of all of the wonderful things that you've done in your career?

A Oh, we'll do real fast. So big highlight about me is I'm originally from Trinidad and Tobago, an island in the Caribbean, came to the U.S. for college, double E, got seduced by consulting, and did that for a couple of years, worked in oil and gas, electric power, heavy industries, love that stuff. But also like writing code on the weekends for fun. So I thought I should move into tech and I did. Worked at Intuit and helped develop their first iPhone app, which was, you know, a thing back in the day. Worked at a startup. Growth team at Facebook for four years, working on user acquisition, which was really fun. And I get kind of like strong, performative experience I had. Quick stint in biotech, and then worked on marketplace at Lyft. So rider pricing, real-time driver incentives. Matching writers with drivers. And then a lot of the operational tools that we use to manage our marketplace. And so that's a bit of my journey in maybe 45 seconds.

AI assessment note: “Worked at Intuit and helped develop their first iPhone app”

Q so many questions along these lines. Okay. I'm going to pick a couple. One is Facebook is famous for this kind of activation milestone of getting 10 friends or seven friends, whatever it was. Like there's some number of friends you got to get and the good things will happen. Were you involved in that? Do you have any insight into like how that came to be? Is that real?

A That decision came before me. I saw it. I understood the data and I worked on this problem. What I thought was brilliant about that was not the, it's not the metric. It was the designing it to be understood and communicated. What I think is fabulous about it is that you're talking about it now because it's memorable and it got people to take the right actions to start chasing the gold. There was literally nothing magic about the number or the date, but basically it was a way of saying like, Get people as many friends as possible, as fast as possible. And if you said that generically to someone, they'd be like, yeah, I kind of get it, but yeah, go do that. When you create a discrete number and a discrete time, and there is a concrete goal to chase, and there's a number and a graph that everybody can look at and see, we are going to go make that thing go up. The organizational effect of that is galvanizing. So what I thought was brilliant about it is, as I've heard the story is, you know, this is all secondhand. There was a lot of debate about what the number should be, what time frame should be. And at some point, Zakja said, 10 friends, 14 days, go. And it just, just got people past the academic debate of like, all right, got it. As many friends as possible, as fast as possible.

AI assessment note: “That decision came before me. I saw it. I understood the data”

Q And is that actually how you divide up those bets broadly? Is that like a rule of thumb you have, or is that just numbers you're putting out there?

A Those are just numbers I'm putting out there. It's always going to be a gut call beast on where you are. I think depending on the stage your product is at, it should be a different set, a different bias. Very early on when you're building a product, you kind of know what the big things are. You've talked to enough people, you have enough, just go build it. You should not be playing around with experiments. It might be a hundred percent cannonballs. Just go knock the big pieces out. Don't worry. It'll work. Also the cost of experimentation is time. So if you're experimenting on every little thing and waiting for the data to come in and then also screwing up some other part of the product, because your experiments on fifty-fifty, it's just not worth it. Just bang the big things out. As you get more mature, the balance needs to switch in the portfolio. Probably, you know, probably aren't that many big cannonballs anymore. Probably just one. And this will be a lot of the refinements that you need to work on. And by then you have the scale that the time to experiment isn't as high and the cost of experimenting is lower. So it's fine. It's good to do it that way.

AI assessment note: “Those are just numbers I'm putting out there.”

Q this topic is the idea of these kind of R and D ish teams at larger companies and companies that have been around for awhile. I know you're relatively new there and this kind of may be a new thing, but I'm curious, is there anything you've learned about how to set up teams like this and investments like this, these kind of long-term horizon bets R and D teams?

A Yeah, I think it's really Good coming to this from being on the other side of it. If you think about where I've been, I've been on growth and on marketplace, which is as far as you get from seeing like we're on the new stuff kind of team. And what I've seen happen a lot is organ rejection. That like this thing looks so different to the rest of the body and the rest of the organization that you get some form of rejection of the ideas entirely. So I think what I've learned is a few things. So first is the rest of the company needs to see what you're doing as being core and critical to the mission. It can't seem like these guys are just playing off in a corner on something that isn't related to what we are doing every day. Cause I think that leads to some of the like resentment. Cause you can imagine any team internally is fighting for resources and they look at this group as having resources that they can't get. They're like, ah, we gotta get rid of that because they're not helping us do what we are here to do. So you have to be part of the core mission. Otherwise you're gonna have problems culturally with that. So I think that's one thing. The second is it has to be everyone's success. So if you end up doing something on one of these R and D teams, it should just be the R and D team that wins. Everyone should feel like they win. And that is kind of related to that first goal I w…

AI assessment note: “So I think what I've learned is a few things.”

Q that they're doing bad things to the world. And imagine as a PM, that's just like a challenging place to be. And the fact that you've been at three different places. I imagine you've learned some stuff about how to operate as a product leader at companies full of chaos and fires and bad PR and things like that. So is there anything to share about what you've learned there?

A I think the biggest thing is that as a PM, you are a leader. You have to provide a buffering or damping effect on the team. And that goes two weeks. Sometimes we're doing stuff that everybody thought was amazing. This is the best thing we've ever seen. And you kind of got to bring people back down to earth and go, look, that was cool, but we got a lot more stuff. We are really not there on providing the value that we want to provide to people in the real world. So slow your roll and recognize that there's a lot more to do. And then when it's terrible and the press is telling you that like, you're the worst thing that ever happened in the world, kind of have to also go back and say, guys, slow down. We're not anywhere near as bad as what they think. You see and know what we're doing, and they're going to misunderstand us sometimes. And so pull your team up at this point in time and keep charging forward with the mission. I think some controversy is necessary. So I may be in a different point on that one. I don't think you're going to have any meaningful influence on the world without changing some pattern of behavior. And if you're changing a pattern of behavior, there's somebody who's invested in that pattern of behavior, and that's going to create some conflict. The most fun news stories to read involve conflict. So that's always going to make for a great story and put you in …

AI assessment note: “You have to provide a buffering or damping effect on the team.”

Q So the main thing you said is just treat ops as a first order component when you're designing the software. Is that the big learning?

A I think it's not just treating ops as a kind of first order requirement. The bigger picture for me was like, When I look across my career is the algorithms need people to help make judgment calls. And so I saw it really, I got a heavy lesson in it at Lyft. But when I look back, I recognize it was there at Facebook too. It just wasn't in my domain. There is always a judgment call that has to be made between how often are there going to be ads versus how often are we going to show organic stories from your friends and family? How often are we going to show content that you might be interested in that's not quite in that group? How often might we want to show you things that help you find your friends or help other people find their friends? And that is a judgment call that varies for different markets in different situations. And there may be algorithms behind the scene that are making that call for every single person in real time, but there still have to be people applying some strategic judgment to that. And I wasn't in the position of needing to do that at Facebook, but once I saw how much I needed to do it at Lyft and I kind of looked back at history, I saw that it was there too. But I think there are too many people who don't see this. And believe that there's an algorithmic solution to everything. I think as a product manager and especially product managers working on syst…

AI assessment note: “I think it's not just treating ops as a kind of first order requirement.”