Paul Graham: How to find an idea for a startup (part one)

A recent article by Paul Graham about finding ideas for a startup. Grem is itself a cult personality in Silicon Valley and is therefore worthy of attention. But in my opinion, this article is one of the best of its kind. Yes, it’s just lazy about Customer Development and Lean Startup, but it has Gram’s deep inner philosophy and summarizes his extensive experience as the founder of YCombinator, who has thousands of startups communicating a year.

The article is very long, so I took the liberty to break it into 5 parts, so as not to tire anyone. Go:



The best way to find an idea for a startup is to not think about it. Find the problem, and better - if you have it yourself.





The best startup ideas have 3 things in common: there is something that the founders need; they themselves can create it; and only to a few other people this seems worthy of realization. Microsoft, Apple, Yahoo, Google, Facebook have all gone this way.

')

Problems

Why is it so important to work on a problem that you have yourself? One of the reasons is to make sure the problem really exists. It sounds obvious that you only need to work on the problems that really exist. Nevertheless, this is the most common mistake made by startups - to solve a problem that no one else has.



I did it myself. In 1995, launched a project designed to bring all art galleries online. But the art business doesn't work that way. So why did I spend 6 months working on this idiotic idea? Because I ignored users. I invented a model of the world, which did not correspond to reality, and worked, proceeding from it. And I did not notice that the model is incorrect, until I tried to convince users to pay for what we developed. Even after that, it took a long time until I realized everything. I was so attached to my model of the world and I spent so much time developing the software that users had to want it!



So why do so many founders build projects that nobody needs? Because they start with trying to come up with an idea for a startup. Such a way of thinking is doubly dangerous: it will not just bring no good ideas, it will generate bad ideas that seem plausible enough to be deceived and start working on them.



In YC (YCombinator) we call them “made-up” or “sitcoms”. Imagine someone from a television series decides to start a startup. A writer would have to come up with something. But it’s hard to come up with a good startup idea. This is not something that can be done upon request. Therefore, the screenwriter will propose an idea that will ring true, but in reality it will be bad.



For example, a social network for pet owners. It does not seem to be absolutely mistaken. Millions have a pet. Often they care a lot about them and spend a lot of money on them. Of course, many of the owners would like to have a website to chat with other owners. Probably not all of them, but if 2-3 percent of them constantly visited the site, you could have millions of users. You could show them targeted ads and perhaps offer some premium features.



The danger of this idea is that if you interview your friends who have pets, they will never say that they will never use this service. They will say: "Yes, I can imagine that I am using something similar." Even when a startup is launched, it will sound true to many people. They themselves do not want to use it, at least right now, but they can imagine that other people are using it. Aggregate this response among the general population and as a result get zero users.

Well

When a startup is launched, there must be at least a few users who really need what you are doing. And not just people who may see themselves ever using your service, but who do not need it right now. Usually this initial user group is small. The reason is simple: if this group, which urgently needs something, was large and it could be built by the forces that a startup would normally need to develop the first version, it would probably already exist. Thus, you need to donate something: you can build something for a large audience, which is not very necessary; or for a small audience that really needs it. Choose the second! Not all ideas of this type are good startup ideas, but almost all good startup ideas relate to this type.



Imagine a graph where the x-axis reflects the number of people who want your product, and the y-axis how much they want. If you mirror the Y axis relative to the X axis, you can imagine the company in the form of holes. Google will be a huge crater: hundreds of millions of people use it and they really need it. A newborn startup should not expect to dig a pit of such a volume. Thus, you have two options regarding the shape of the hole with which you start: either you dig a wide but flat hole; or narrow, but deep, like a well.



Fancy start-up ideas usually refer to the first type. Many people are moderately interested in a social network for pet owners.



Almost all good startup ideas are of the second type. Microsoft was well when doing Altair Basic. There were only a few thousand users, but without it they would have to program in machine language. 30 years later, Facebook had the same figure. Their first site was only for Harvard students, only for a few thousand, but these users wanted this site strongly.



If you have an idea for a startup, ask yourself: who needs it right now? Who needs it so much that he will use even the ugly first version of a product made by a two-person startup he has never heard of? If you don't have an answer, maybe you have a bad idea.



But you do not need the narrowness of the well itself. You need its depth. Narrowness is obtained as a by-product of depth optimization (and speed). And you will get it — narrowness — almost always. In practice, the relationship between depth and width is so strong that it is always a good signal when you know that an idea strongly attracts a certain group or type of users.



And although the demand in the form of a well is almost a prerequisite for a good startup idea, it is not sufficient. If Mark Zuckerberg had built something that only Harvard students would like, this would not be a good startup idea. Facebook was a good idea because it ran on a small market while there was a fast way out. Colleges are quite similar in the sense that if you built facebook that suits Harvard, it will work for any other college. So that you can spread the idea quickly to all colleges. After you have all college students, you will get all the rest simply by firing them into the system.



Same with Microsoft: Basic for Altair; Basic for other computers; languages ​​other than Basic; operating system; soft; IPO.



Continuation (part two) follows tomorrow. Just right to digest and / or sleep with thoughts.