This essay is an instant Kiwi classic. Few comments from me: "OKCupid, like the other acquisitions of Match.com, is now just another Tinder clone - see face, swipe left, see face, swipe right. A digital nightclub. " Never liked Tinder and couldn't nail why - 'digital nightclub' is a perfect name for its vibes. "What's wrong with such a metric? A product that many users want to use is a good product, right? Sort of." What's interesting about DAU is that it tells you whether a person opened the app, which is a binary metrics. But it doesn't tell you how happy your user was using your app, which is not a binary metric. Tech typically uses some engagement metrics to measure it, like "Time spent in the app," but I don't think it correlates with user happiness. The more time I spend on most social media the less happy I am. "Since most software products charge a flat per-user fee (often zero, because ads), and economic incentives operate on the margin, a company with a billion-user product doesn't actually care about its billion existing users. It cares about the marginal user - the billion-plus-first user - and it focuses all its energy on making sure that marginal user doesn't stop using the app. Yes, if you neglect the existing users' experience for long enough they will leave, but in practice apps are sticky and by the time your loyal users leave everyone on the team will have long been promoted." This stickinness is a great point. And you don't even need network effects to build it. I still check the same Polish news site I've been checking since 2001. It changed a lot since then, I don't like most of the changes, but the habit is there. So metrics-wise, they can prove that the new shiny features haven't alienated the old users.
I like the focus on the core dynamic of many consumer apps today, which is towards extreme simplicity at the UX level. Ivan also says he doesn't know how to counteract it. There are some potential solutions I think. First, a comment: "a billion-user product doesn't actually care about its billion existing users. It cares about the marginal user - the billion-plus-first user - and it focuses all its energy on making sure that marginal user doesn't stop using the app." That does not ring true. A billion-user product does care more about its 1,000,000,000 users than it cares about 1 new user. In fact, it likely cares about what it is optimized for, which in OKCupid's case as a privately owned company, is shareholder profit. That profit comes from whoever can be monetized better: if new users generate more profit than existing users, focus is on new users. If existing users generate more profit than new users, focus is on existing users. At a billion-existing-users, focus is likely on existing users. Second, another thought: this may be what the market wants, which the more general trend among other players in the industry suggests as well. If not just OKCupid but other companies simplify their UI/UX dramatically as well, this may be a customer-driven outcome, not a producer-driven outcome. Producers may adapt to customers rather than customers adapting to producers. Third, a potential solution to the trend towards simplified UI/UX. Decoupling the backend from the frontend is a solution to the problems posed here. That is, were OKCupid's db and data open, like in Farcaster, then anyone could build different UI/UXs. Some of those frontends would retain the old, personality-test driven UI/UX experiences; others would mimetically copy the more simplistic Tinder-like UI/UX experiences. If you want to preserve old UI/UXs, like the way a Polish news site used to look in 2001, then store the site's frontend on Arweave (pay-once-store-forever) and you'll have access to the same UI/UX in 2024 as you had in 2001, with updated content if data is stored openly and permissionly on the backend, web3-style, via some headless CMS-like tool. This promise is one rooted in Web3, and for pointing towards that solution I really like the article even though I don't agree with all of it!
hey @rolf.eth, thanks for sharing your thoughts; I got some comments :) "If existing users generate more profit than new users, focus is on existing users. At a billion-existing-users, focus is likely on existing users." I generally agree with it. One angle I tried to add by sharing my Polish news site example is the power of habit. Let's say your billion users are, on average, 80/100 happy with your product. They built a strong habit of using your product, so they won't easily break it. Let's say only if they get to 40/100 happiness will they stop using the product and rebuild their habits just to use something else. It means you can add a feature that would turn 80/100 happiness into 70/100 happiness without seeing that in your metrics. So there's this latent old user happiness decrease. And if that feature makes the marginal user more happy, the metrics will show it's a net plus for the product. This is what I experienced with this Polish news site. I'm less happy than I used to be, but I haven't reached the 40/100 happiness threshold, so I won't rebuild my habits to use another site. "Second, another thought: this may be what the market wants, which the more general trend among other players in the industry suggests as well." I'd rephrase it and say this is what the biggest/most profitable market segment wants. If you run a profit-oriented company and seek to maximize profits, you must follow the segments that generate the most money. So why run feature-rich OKCupid for 1M users when you can run a Tinder-like app for 100M users? It doesn't mean that old users stopped wanting old OKCupid. Some of them still like it. But they aren't the most profitable market segment to focus on. In the long run, thanks to free markets, I believe some 'old school dating apps' will emerge to serve them, though their profit ambitions would be smaller than those of Tinder-like apps. "That is, were OKCupid's db and data open, like in Farcaster, then anyone could build different UI/UXs. Some of those frontends would retain the old, personality-test driven UI/UX experiences; others would mimetically copy the more simplistic Tinder-like UI/UX experiences." This is a brilliant idea. I mean, in web3 this back-end/front-end decoupling is something we know very well. But never thought about it from the dating app POV - this would be something really cool :)
Love your thoughts and they make a ton of sense. Agree with all of them, thanks for digging in and sharing your clarity of thought here!
Great read. DAUs not being an ideal metric remind me of the quote "the map is not the terrain". Heard it a lot when doing work w governments and the public sector in a previous life. In a lot of governments, there was a push in the 80s towards being more metrics-driven (New Public Management), but in complex systems governments operate in, this led to a bunch of unintended consequences. There's been a shift away from that reductive thinking in a lot of public service orgs, but only in pockets, and it's hard to make conversations about UX and systems thinking fit with the old ways of doing things. I feel flattening of nuance and greater abstraction will become more common as AI proliferates. What's interesting and makes me hopeful is the rise of 'luxury software' that sometimes isn't as obsessed with DAU, and has a better relationship with their users because of serving a v niche community.
@uncledavo, that reminds me of Goodhart's Law: "When a measure becomes a target, it ceases to be a good measure." Andy Grove had a solution called pairing indicators: “Indicators tend to direct your attention toward what they are monitoring. It is like riding a bicycle: you will probably steer it where you are looking. If, for example, you start measuring your inventory levels carefully, you are likely to take action to drive your inventory levels down, which is good up to a point. But your inventories could become so lean that you can’t react to changes in demand without creating shortages. So because indicators direct one’s activities, you should guard against overreacting. This you can do by pairing indicators, so that together both effect and counter-effect are measured.” So, e.g., if you measure the number of users, you should also have a way to measure their satisfaction. Since satisfaction is typically much harder to estimate, many projects tend to focus on DAUs. I think your point about 'luxury software' as a solution is interesting because the whole luxury industry is very focused on user satisfaction. And since there are very quantifiable ways to express satisfaction with luxury products (typically by buying more and more expensive stuff), brands like LVMH can grow to multi-billion valuations. Wondering how it could translate into software.