On finding a deep, narrow problem your software actually solves
A second exchange with Andreas Stuhlmüller, the probabilistic-programming researcher whose Ought question-answering project Alyssa had begun advising. Stuhlmüller had shared a Google doc spelling out the use cases his system was meant to handle — medical questions feature heavily — and asked again about lawyers. The reply zeroes in on a recurring failure mode for AI startups: a system that nominally answers questions in a domain where thousands of websites already do is not really competing unless there’s a specific, well-defined sub-population for whom the new answer is clearly better than every existing answer. Most of the body is then a long quotation from Paul Graham’s essay on startup ideas, spelling out the same point in his own terms.
For lawyers, how much money are you prepared to spend? Retaining one will typically cost thousands of dollars, even for fairly basic work. That’s a smaller concern if you already have investors, but are you funding this all out-of-pocket?
I skimmed over the Google doc, and it seems unclear which particular use cases this software would excel at. Eg., medical questions are mentioned a lot, but there are already thousands of websites devoted to medical questions. That’s not necessarily bad — it shows that there’s strong demand — but there have to be particular questions where the new solution does better than all the existing solutions. Eg., here’s one made-up hypothetical: Question: I’m a 22-year-old taking Progenitorivox, and I have pain in my chest. Could this be a heart attack?
Existing Website #1: Heart attacks are a possible side effect of Progenitorivox.
Existing Website #2: Chest pain is a potential symptom of heart attacks. If you might be experiencing any symptoms of a heart attack, call 911 immediately.
Existing Website #3: A few years ago, I [not mentioned: author is a 65-year-old smoker] was taking Progenitorivox, and I had a heart attack!
New Website: 34% of Progenitorivox users experience chest pain, but only 0.0015% of Progenitorivox users your age experience a heart attack. Therefore, the odds you are experiencing a heart attack are less than 0.01% (one in ten thousand).
In this case, there might be a ton of competitors but it doesn’t matter, since their answers all suck and your answer doesn’t. But there has to be some particular group of users for whom that is clearly true. Commentary by Paul Graham ( How to Get Startup Ideas ): “Why is it so important to work on a problem you have? Among other things, it ensures the problem really exists. It sounds obvious to say you should only work on problems that exist. And yet by far the most common mistake startups make is to solve problems no one has.
I made it myself. In 1995 I started a company to put art galleries online. But galleries didn’t want to be online. It’s not how the art business works. So why did I spend 6 months working on this stupid idea? Because I didn’t pay attention to users. I invented a model of the world that didn’t correspond to reality, and worked from that. I didn’t notice my model was wrong until I tried to convince users to pay for what we’d built. Even then I took embarrassingly long to catch on. I was attached to my model of the world, and I’d spent a lot of time on the software. They had to want it!
Why do so many founders build things no one wants? Because they begin by trying to think of startup ideas. That m.o. is doubly dangerous: it doesn’t merely yield few good ideas; it yields bad ideas that sound plausible enough to fool you into working on them.
At YC we call these ‘made-up’ or ‘sitcom’ startup ideas. Imagine one of the characters on a TV show was starting a startup. The writers would have to invent something for it to do. But coming up with good startup ideas is hard. It’s not something you can do for the asking. So (unless they got amazingly lucky) the writers would come up with an idea that sounded plausible, but was actually bad.
For example, a social network for pet owners. It doesn’t sound obviously mistaken. Millions of people have pets. Often they care a lot about their pets and spend a lot of money on them. Surely many of these people would like a site where they could talk to other pet owners. Not all of them perhaps, but if just 2 or 3 percent were regular visitors, you could have millions of users. You could serve them targeted offers, and maybe charge for premium features.
The danger of an idea like this is that when you run it by your friends with pets, they don’t say ‘I would never use this.’ They say ‘Yeah, maybe I could see using something like that.’ Even when the startup launches, it will sound plausible to a lot of people. They don’t want to use it themselves, at least not right now, but they could imagine other people wanting it. Sum that reaction across the entire population, and you have zero users.
When a startup launches, there have to be at least some users who really need what they’re making—not just people who could see themselves using it one day, but who want it urgently. Usually this initial group of users is small, for the simple reason that if there were something that large numbers of people urgently needed and that could be built with the amount of effort a startup usually puts into a version one, it would probably already exist. Which means you have to compromise on one dimension: you can either build something a large number of people want a small amount, or something a small number of people want a large amount. Choose the latter. Not all ideas of that type are good startup ideas, but nearly all good startup ideas are of that type.
Imagine a graph whose x axis represents all the people who might want what you’re making and whose y axis represents how much they want it. If you invert the scale on the y axis, you can envision companies as holes. Google is an immense crater: hundreds of millions of people use it, and they need it a lot. A startup just starting out can’t expect to excavate that much volume. So you have two choices about the shape of hole you start with. You can either dig a hole that’s broad but shallow, or one that’s narrow and deep, like a well.
Made-up startup ideas are usually of the first type. Lots of people are mildly interested in a social network for pet owners.
Nearly all good startup ideas are of the second type. Microsoft was a well when they made Altair Basic. There were only a couple thousand Altair owners, but without this software they were programming in machine language. Thirty years later Facebook had the same shape. Their first site was exclusively for Harvard students, of which there are only a few thousand, but those few thousand users wanted it a lot.
When you have an idea for a startup, ask yourself: who wants this right now? Who wants this so much that they’ll use it even when it’s a crappy version one made by a two-person startup they’ve never heard of? If you can’t answer that, the idea is probably bad.
You don’t need the narrowness of the well per se. It’s depth you need; you get narrowness as a byproduct of optimizing for depth (and speed). But you almost always do get it. In practice the link between depth and narrowness is so strong that it’s a good sign when you know that an idea will appeal strongly to a specific group or type of user.”