Can Olcer


Quantifying YC’s “Make something people want”

Y Combinator is one of the world’s most renowned accelerators. Their slogan and main mantra is “Make something people want”. At first, this truism might seem like a… truism. Of course, you should make something people want if you want to build a sustainable company, else no one would use it.

The following is my attempt to explain “Make something people want” in a way that it’s actionable for entrepreneurs.

My main realization is that “Make something people want” is the aggregate of advice that is given by YC. Almost every advice they give aims at this.
I’ve spent some time in the last months understanding YC. I read Paul Graham’s essays (the founder of YC), listened to talks and interviews by Sam Altman (YC’s president) and other YC partners and participated in YC’s inaugural Startup School (note that Startup School is something different than the “real” YC. I did not take part in YC.). While it shouldn’t be taken as dogma, I think they have some compelling advice.

Some of the more famous sentences repeated by YC are:

“It’s better to make something that 100 people love than something that 1,000 like.” (or any variation of that)

“Talk to your users. Talk to 100 users a week.” (or any variation of that)

“Focus on weekly growth of revenue (or the closest proxy of revenue if you don’t make revenue).”

“Why do people care about what you do?” (in other words, what’s your value proposition and what problem are you solving)

“Make that something that solves a problem you have yourself.”

“Launch as fast as possible.”

In the end, it all revolves around making something people need (and therefore want). How do you determine what people need? Easy, figure out a problem people have and provide a solution. How do you find what problems exist? No big deal, just look at what problems you have yourself or talk to a bunch of people and ask them. Evidently, it’s not as simple as it sounds.

Your H(0) is almost always wrong

Often, if you start a new company, you start with a hypothesis of a problem that exists and that your solution is a significantly better solution to that problem than existing ones. The chances that your hypothesis is correct are slim. After all, how the heck are you supposed to know?

So you build something and give it to people to use. Now, instead of just following your initial product plan, you need to talk to people who use your product. And ask them why-questions. And dig deep. Ask them like ten why-questions (like a little kid that keeps asking). Understand the underlying reason for their answers.

And it doesn’t matter if you have 10 or 100 or 1,000 initial users — as long as you have a dozen committed super fans. Make sure your product satisfies their needs. And then start expanding to onboard more and more users with slightly different needs. Like a fire that starts in the middle and grows circularly. That’s why I believe the advice “It’s better to make something that 100 people love than something that 1,000 like.” is a very non-obvious but powerful advice.

It’s not hard to get 10,000 users, but it’s hard to get 10,000 users that keep using your product over a longer period. If you start adding users (e.g., through ads or other paid channels) that leave after one week, it’s money down the drain. But if you make sure users stay, then every new user is worth much more because you know your product is solving their problems and they keep using it.

That’s why I believe that measuring the growth of your retention rate is significantly more important than measuring the growth of new users. Of course, in the end, there needs to be a right balance — but in the beginning, this is all that counts. And it’s hard because it’s much cooler to say that you have 1,000 weekly active users than saying your 6-week retention grew by 10%.

How to quantify love?

Retention! How do you measure retention? There are different ways, but the general idea is that you take an action a particular person did in your app (e.g., opened it) and see how often they keep doing it after they did it. For example, you look back six weeks and look at all the people that opened your app that week. After that, you look at every one of the five weeks until today and measure how many of those users opened your app minimum once per week. Several analytics tools help you do that easily. This is our real retention report from Mixpanel:

Mixpanel’s weekly retention report (real numbers from Humbot)

For example, in the week of July 17, 69 people opened our app. And four weeks later, 23.19% of them did it again (that week is not complete yet as of the writing of this article).

What your primary goal now is to make sure that the columns grow with time. Because this means that people increasingly keep using your product more.

So, every time you add a new feature or make a change, you can observe if it makes your product more or less attractive. In casu, our retention rates have decreased a little in the last two weeks. This is because we added a browser version next to the existing iPhone and Android versions of Humbot. Naturally, websites have lower retention rates than apps, and this lowers our average.

Retention rate is the scalar quantity of “Make something people want”.
I believe that it’s important, especially for new startups, to first get your retention rate to a place that you’re happy with. You’re doing something right if you have people using your product on a regular basis. Talk to your users to find out why they are using it. Take that info and add more features in that direction. See what happens to your retention rates. Talk to some more user (by the way, a good idea is also to talk to prospective users that you think should be using your product and find out why they are not.).

If you feel like you got a strong base of super fans, think about how adding new users and growing your weekly active users. But always make sure to keep an eye on what constitutes your weekly active users (new vs. recurring users).

PS: I am aware that this advice is not valid for every company or startup. I am talking from the vantage point of a consumer tech company. Still, I am sure that some of this can be adapted to other verticals/industries. Also, there is no right or wrong retention rate. It depends on your company’s mid-term strategy. YC recently put together an interesting survey of retention rates that might help you to get an idea. Read it here.

PPS: I did not participate in YC. This is purely from publicly available info.

More by Can Olcer

Topics of interest

More Related Stories