Lenny Rachitsky: uncertainty

Transcript context

…a little broader than that. Right? There's lots of really great resources. There's amazing folks on the internet who will go down and really dissect... I love Eugene Wei's writing. I love Kevin Kwak. I always learn something because they take these apps like Figma or TikTok or whatever it is, and then they really go very, very deep with their own observations, what works, what patterns do we see across different apps that are successful and that aren't. Right? And this is all helping us to understand what are these again, the key decisions and what impact does it lead to that helps us become better at then making those intentional decisions in the product. So, that's a huge part of it. I think another thing then is of course you have to try and validate. So, one thing we can do is of course we look at opinions, we look at reactions. That's data. Right? That's the qualitative side. I think the other side is quantitative. So often if you are building products and you have the opportunity to run experiments, to do AB tests, or if you're working on one team but lots of other people in lots of other teams are also doing AB tests, it's so interesting to then be able to ask people, ask the product manager on the other team about what they're learning about their products, and to really be able to look at specific decisions and what causally happened as a result. That's what I love about AB tests, and I think being really deep in the data and really going back to can we infer some sort of causal relationship because we're... Correlation or causation, but with causation with AB tests, can we actually pick up some of these learnings, can we look at patterns, and can we take some insights away that helps validate and confirm a lot of the hypotheses that we had about product, and just ingesting as much of that as you can also helps develop your instinct for what works and what doesn't. Right? I always find people often have this, oh, design and user experience is on the other side of the coin. It's like it's a totally different industry, and they're at odds with each other, being data informed and being quantitative versus being very designer-y and subjective and caring about those aesthetics. I just think that's totally wrong. I think that there really... One helps confirm the assumptions or the other. Right? No, it is true that looking at a bunch of numbers isn't often going to tell you exactly the leaps of faith that you need to make to start something new, but they surely can help you validate whether a number of your assumptions about how people work or the way the world works are true or not. And so, I know of a lot of really brilliant product thinkers who got that way not necessarily because they came through the route of subjective observation, but because they went and they were so disciplined about always studying what happened, what was the impact in the numbers and people and so forth, and then eventually you marry that of course with, well, why might that be the case, and you get into the qualitative side and the observation, but these two both support each other in helping to build a really great product sense. That's awesome. There's so much material there that we could go on and on. On that last point, I wanted to kind of double-click on it a little bit. So, say you're founder and you're like, "Man, I have all these really clear vision and ideas of where I want to go with my product," and your team's like, "Oh, I don't know if this is right. What if we do a little more user research or run some experiments?" Do you have any advice to the founder of just when to rely on their gut and experience and just go with that versus doing more research, getting more data? That is a really great question. One of the most common pieces of advice for founders, and I actually also had to remind myself constantly of this one is the more you know your customers, the more you can really close your eyes and just imagine everything about their life and what they're doing on almost like a minute to minute basis, probably the better you're going to do in terms of coming up with something that's going to meaningfully solve a problem for them. Right? And so, that comes from a couple of different places. The first is, look, if you're the person you're building for, you're the target audience, awesome. You probably do have a lot of stuff that is instinctively known to you, and maybe in those cases, your team doesn't have that experience and they maybe can't feel the same level of conviction you do, and they might be asking you, ""Hey, well, can we validate?" and all of that. Right? It's always good advice, but sometimes, you're so deep in it and you can... You're this person or you know this person, or you did this job that probably can trust your instincts and your gut quite a bit. I remember early days at Facebook, that was us. Everybody who worked at the company was either a college dropout or a recent college grad, and we were building a product for college students. I mean, we were the perfect... It was like for us by us. We understood exactly what this audience wanted. If we didn't, we would call up some friends. I mean, this was just pure target a demographic for what we were building. But eventually if that's not true, and it evolved at Facebook, and it evolves for companies. You might start out that way, but eventually, we started to open up to the world. We started to add people in different countries. The percentage of people that were college grads who were like us who were using the product started to shrink, became a smaller and smaller percentage of actually all core Facebook users, so therefore our intuitions started to become less and less reliable. I remember in spectacular fashion, I think this was in 2008 or '09, we had a string of failures, big kind of launches that were failures, and I think it was because we reached the end of our intuition for the user base at that particular moment. That's true for founders as well. Sometimes you're building a product in a domain where you weren't the target audience. Right? I feel this right now for myself. I'm building an analytics product. I was never a data analyst.…

Stored transcript either side of the excerpt. The highlighted words are the published quote; the surrounding text is unedited source, never generated.