- Startups often attract a variety of users with different problems when they open up their product to the public.
- Some users may hijack the product to solve their own problems, which may not align with the startup's intended purpose.
- While hijack users may bring short-term value, they can ultimately harm the startup's long-term success and should be identified and removed.
- It's important for startups to focus on users who are using the product as intended or in interesting ways with potential.
- Users who create no value or actively decrease value should be kicked out, even if they are paying customers.
- Startups should not let hijack users steer their product roadmap and should instead focus on solving one problem well to make a small group of users happy.

Michael Seibel is a partner and the CEO of YC. He cofounded Justin.tv, which was in the winter 2007 batch and Socialcam, which was in the winter 2012 batch. https://twitter.com/mwseibel

In this video Michael comments on his essay Users You Don't Want https://blog.ycombinator.com/users-you-dont-want/

The YC podcast is hosted by Craig Cannon - https://twitter.com/craigcannon

***

Topics

00:00 - Users You Don’t Want

1:30 - The spectrum of how users are using your product

3:00 - Users that take a lot of customer support time

5:50 - Don’t let the hijack users control the product roadmap

right, when you're just getting started many startups will take every user they can get They have a strong idea of a problem and they want to attract as many users with that problem as possible Unfortunately when you open up the barn doors, you get all sorts of people with all sorts of problems Some of them will try to hijack your product to solve a problem

I mean, Justin.TV was by definition hijacked.

We built a product to allow people to live stream their lives, and within a year it was being used to stream copyrighted content around the world.

It's absolutely the definition of a hijacked product.

And what's interesting is that I'm a little afraid of my phrasing here.

I wish I didn't say that these hijack users are not users you want.

Because it turns out that sometimes they are.

It turns out that the reason why we even had an inkling to do Twitch was because some percentage of the hijacked users were video gamers.

And it turns out that that could have been a much bigger community if we helped it and then eventually it became 1.

So this kind of like, users using your product for a whole variety of things, I like to think about it more on a spectrum.

There's the user who's using your product as intended.

Maybe there's a business there, maybe there isn't.

There's the user who's using your product in interesting ways with potential.

Study those users, those users are very important.

There are users who are using your product with ways that it's extremely clear to you that there isn't long-term value, Even if there is short-term value.

I'd argue that those were the copyright and streamers on Justin.TV.

There was short-term value because we could monetize with ads and so on and so forth, but long-term, we weren't really creating value.

I'd argue Those are users who are using your product and they're creating no value, and they're probably actively decreasing your value.

For us, that would be people like, some people streamed pretty horrible things on Justin.TV.

What's interesting is that now when I talk to startups at YC, they don't know how to recognize these hijack users and just kick them out.

And they kind of, no 1 talks about these people.

They always talk about everyone before them, which those are users that have some value.

And they're talking about the hijacked users.

And it's like, if you can identify the hijacked users, you need to kick them out.

They're actually going to hurt you more than they're going to help you.

And I think that it's not that the high-tech users have bad intentions even.

But some, it's just they want to use your product to do what they want to do, and that is literally going to hurt your business.

There are some things that, like, if people do it on your platform, it hurts your business.

Edith Warren How would you classify someone who just

takes up all of the customer support time?

Peter Collins It depends on how much revenue they're generating.

I'd argue that if they are generating enough revenue to justify that support time, great.

If you have a free product, not so much probably.

If you're taking a bunch of your support time, but you're learning and you're actually improving your product because of it, great.

But for example, I'll create a hypothetical example off the top of my head.

Imagine that in the beginning of Airbnb, someone wanted to use Airbnb to host drug parties, like illegal drug parties, crack parties.

And they thought of these Airbnb is like mobile crack den, right?

Well, let's think about this for a second.

Like, so on 1 hand, right, while they are paying,

they're paying for the big apartments, they're

On the other hand, they're probably destroying those apartments.

Doing something highly legal, getting the police called on them, like da da da da da, right?

These are not the types of people you want.

And you can be a hijack user and there's still some positive.

But if there's way too much negative, it's not worth the positive.

I actually see it a lot where it's like the hijack user's paying, so you feel bad.

Because when the hijack user's not paying, usually you feel less bad kicking them out.

When they're paying, you're like, well, but they're paying.

Yeah, especially in the context of you have these metrics you want to hit, and you're like, even during my seed.

Yeah, and you end it in a nice, clean way by saying, by focusing on solving 1 problem really well, you're betting on making a small amount of people very happy.

If you let any user that walks in the door the product roadmap, you're going to end up doing a shitty job at half solving a lot of problems.

Shen Yeah, I mean To be clear with this ending, my whole goal is that, like I said, there's this spectrum of potentially useful users.

My whole goal is don't let the hijack users steer your product roadmap.

It's really totally okay if you want to explore some of those use cases that are not the use case that you thought, that are still use cases that could be good for you.

Then it's totally okay to explore to experiment so on and so forth.

But it's like, don't let the crack den guys on Airbnb control the product roadmap for Airbnb and suggest the features that they want.

At the end of the day, you have to have some opinion about your product.

You can't just be like, I don't know, the market's going to take it wherever,

Yes, it's really hard to build product without some opinion.


Speaker 0: Users you don't want. This

Speaker 1: 1 was contagious. Yeah, this was fun. Yeah, let's bring it back All

Speaker 0: right, when you're just getting started many startups will take every user they can get They have a strong idea of a problem and they want to attract as many users with that problem as possible Unfortunately when you open up the barn doors, you get all sorts of people with all sorts of problems Some of them will try to hijack your product to solve a problem

Speaker 1: When has this happened to you? I mean, Justin.TV was by definition hijacked. We built a product to allow people to live stream their lives, and within a year it was being used to stream copyrighted content around the world. It's absolutely the definition of a hijacked product. And what's interesting is that I'm a little afraid of my phrasing here.

I wish I didn't say that these hijack users are not users you want. Because it turns out that sometimes they are. It turns out that the reason why we even had an inkling to do Twitch was because some percentage of the hijacked users were video gamers. And it turns out that that could have been a much bigger community if we helped it and then eventually it became 1. So this kind of like, users using your product for a whole variety of things, I like to think about it more on a spectrum.

There's the user who's using your product as intended. Great, makes you feel good. Maybe there's a business there, maybe there isn't. There's the user who's using your product in interesting ways with potential. Study those users, those users are very important.

Video gamers and Justin.TV. There are users who are using your product with ways that it's extremely clear to you that there isn't long-term value, Even if there is short-term value. I'd argue that those were the copyright and streamers on Justin.TV. There was short-term value because we could monetize with ads and so on and so forth, but long-term, we weren't really creating value. And then there are hijack users.

I'd argue Those are users who are using your product and they're creating no value, and they're probably actively decreasing your value. They're harming your network. For us, that would be people like, some people streamed pretty horrible things on Justin.TV. That was not good. What's interesting is that now when I talk to startups at YC, they don't know how to recognize these hijack users and just kick them out.

And they kind of, no 1 talks about these people. They always talk about everyone before them, which those are users that have some value. And they're talking about the hijacked users. And it's like, if you can identify the hijacked users, you need to kick them out. They're actually going to hurt you more than they're going to help you.

And so I think this is important. And I think that it's not that the high-tech users have bad intentions even. They can, some do. But some, it's just they want to use your product to do what they want to do, and that is literally going to hurt your business. There are some things that, like, if people do it on your platform, it hurts your business.

Edith Warren How would you classify someone who just

Speaker 0: takes up all of the customer support time?

Speaker 1: Peter Collins It depends on how much revenue they're generating. I'd argue that if they are generating enough revenue to justify that support time, great. If you have a free product, not so much probably. Especially if you're not learning. If you're taking a bunch of your support time, but you're learning and you're actually improving your product because of it, great.

But for example, I'll create a hypothetical example off the top of my head. Imagine that in the beginning of Airbnb, someone wanted to use Airbnb to host drug parties, like illegal drug parties, crack parties. And they thought of these Airbnb is like mobile crack den, right? Well, let's think about this for a second. Like, so on 1 hand, right, while they are paying,

Speaker 0: they're paying for the big apartments, they're

Speaker 1: paying for the nice places. Right. Yeah. They're, they got money. Right.

Interesting. On the other hand, they're probably destroying those apartments. Doing something highly legal, getting the police called on them, like da da da da da, right? And it's like, these are hijack users. These are not the types of people you want.

And you can be a hijack user and there's still some positive. But if there's way too much negative, it's not worth the positive. And That's actually where I see a lot. I actually see it a lot where it's like the hijack user's paying, so you feel bad. Because when the hijack user's not paying, usually you feel less bad kicking them out.

When they're paying, you're like, well, but they're paying. And it's like, no. You still kick them out.

Speaker 0: Yeah, especially in the context of you have these metrics you want to hit, and you're like, even during my seed.

Speaker 1: Yeah, right? Oh my God, this is stuff.

Speaker 0: Shut it down. Yeah, and you end it in a nice, clean way by saying, by focusing on solving 1 problem really well, you're betting on making a small amount of people very happy. If you let any user that walks in the door the product roadmap, you're going to end up doing a shitty job at half solving a lot of problems. Edith W.

Speaker 1: Shen Yeah, I mean To be clear with this ending, my whole goal is that, like I said, there's this spectrum of potentially useful users. My whole goal is don't let the hijack users steer your product roadmap. It's really totally okay if you want to explore some of those use cases that are not the use case that you thought, that are still use cases that could be good for you. Then it's totally okay to explore to experiment so on and so forth. But it's like, don't let the crack den guys on Airbnb control the product roadmap for Airbnb and suggest the features that they want.

It's not where you want

Speaker 0: to go. Peter T. Leeson This is sort of the mentality. At the end of the day, you have to have some opinion about your product. You can't just be like, I don't know, the market's going to take it wherever,

Speaker 1: the customers take it wherever. Yes, it's really hard to build product without some opinion. Yeah. You

See all Y Combinator transcripts on Youtube

Users You Don't Want by Michael Seibel