Users You Don't Want by Michael Seibel

Users you don't want, and this one was Stannis. Yeah, this was fun.

Yeah, 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 you didn't intend to solve. By and large, these hijackers are users you do not want.

When has this happened to you? I mean, Justin.tv was by definition hijacked. Yeah, like we built a product to allow people to livestream their lives, and within a year it was being used to stream copper and content around the world. Absolutely hijacked from us. 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.

Yeah, it turns out that, like, you know, the reason why we even had an inkling to do Twitch was because some percentage of the hijack users were video gamers. And like, it turns out that that could have been a much bigger community if we helped it, and then eventually became one.

So this kind of like users using your product for a whole variety of things, I like to think about it more like on a spectrum. There's the user who's using your product as intended. Yeah, great, makes you feel good. Maybe there's a business there. Maybe there's, right, there's the user who's using your product in interesting ways with potential.

Yeah, right, yeah. Study those users. Those users are very important. Video gamers are Justin.tv. There are users who are using your product in ways that it's extremely clear to you that there isn't long term value, even if there is short-term value.

Right? I'd argue that those were the copper and streamers on Justin.tv. There's a short-term value because we can monetize with ads. And so once over the long term, we weren't really creating value. And then they're like hijack users. I'd argue like 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, right? And like for us, you know, that would be people like, you know, some people doing pretty horrible things on Justin.tv, right? Like that was not good.

And so, um, 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. Mmm. And like, they kind of like, no one talks about these people. They always talk about like everyone before them, yeah, right? We're just like those users that have some value and no talk about the hijack users.

And it's like, if you can identify the hijack users, you need to kick them out. Like they're actually going to hurt you more than they're gonna help you. And so I think this is important. And I think that like, it's not that the hijack users have like bad intentions.

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 on your platform, it hurts your business.

How would you classify someone who just takes up all the customer support time? It depends on how much revenue they're generating. Yeah, right? Like I'd argue that if they are generating enough revenue to justify that support time, great.

Yeah, if you have a free product, not so much. Yeah, so much probably, right? Especially if you're not learning. If they're taking up a bunch of your support time, but you're learning and like you're actually improving your product because of it, great.

But like, for example, like let's see, I'll create a hypothetical example off the top of my head. Right? Imagine that I'm at the beginning of Airbnb. Someone wanted to use Airbnb to host like drug parties. Right? Like illegal drug bars, crack parties, and they thought of these Airbnbs as like mobile crack dens.

Right? Well, let's think about this for a second. Like, so on one hand, right, they are paying. Yeah, I mean, they're paying for the big apartment. It's paying for the nice places. Yeah, right? Yeah, they got money.

All right, interesting. On the other hand, they're probably destroying this apartment, doing something highly illegal, getting the police called on them, like dead-headed it out, right? And it's like, these are hijack users. Like, these are not the types of people you want.

And like, you can be a hijack user and there's still some positive, yeah. But it says like, if there's like way too much negative, it's not worth the positive. Now, and that's actually where I see a lot—I actually see it a lot where it's like the hijack users paying.

So you'll feel bad because when the hijack user is not paying, usually you feel less bad taking them out, yeah. When they're paying, you're like, "Well, but they're paying," right? And, and it's like, no, you just okay come out.

Yeah, especially in the context of like you have these metrics you want to hit and you're like, yeah, right. Oh my God, like, this is real. This is stop. It's like shut it down.

Yeah, yeah. And you end it in a nice clean way by saying, by focusing on solving one 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 affect the product roadmap, you're going to end up doing a shitty job at solving a lot of problems.

Yeah, I mean to be clear with this, ending my whole goal is that like, like I said, there's like this spectrum of potentially useful users. Yeah, my whole goal is like, don't let the hijack users steer your product roadmap.

Totally. Like, it's really totally okay if you want to explore some of those use cases that are not these cases that you thought that are still use cases that could be good for you, huh? Then it's totally okay to explore, to experiment, so and so forth.

But it's like, don't let your app—like don't let the crack den guys on Airbnb control the product or the roadmap for Airbnb, right? And suggest two features that they like, right? Not where you want—I mean this is sort of the mentality.

It's like, 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, if the market's going to take it wherever the customers take it, wherever." Yes, it's been really hard to build product without some opinion.