Phantom features

The cheapest way to find out if anyone actually wants it.

Building a product, or even a feature, is expensive. It takes a lot of time and money, and most of the time your assumption about what users need is wrong. When you’re short on time and money (the definition of a startup?) you have to find a shortcut.

So what do you do?

1. Ask users whether they need it. You can, but it’s usually not enough. Users don’t really know what they want, and they’ll say “yes” because why not, especially if the product is offered for free. Experienced people will tell you the distance between saying “yes, I’d buy it” and pulling out a credit card and paying is the distance between heaven and earth.

2. Build an MVP and test it. You can, but it’s not a hit. First because it takes quite a bit of time and money. And second, when the MVP crashes and doesn’t attract users, you will always start finding excuses, because you’re already invested (“it doesn’t have all the features”, “it needs another explanation screen”, “it needs more design”), and you’ll try to make it work.

3. And my favorite: a phantom feature. I have no idea if that’s the official name. You let the user think there’s really something there, take them through the whole flow, and then say “good to know, we’ll get back to you”. That way you’re actually testing on real users.

What does that look like in practice? Eytan Levit once told a story about doing something like this at Mixtiles. They wanted to test a new product, so they built a real landing page with a purchase page, launched an ad campaign and everything. At the end they told users the product was out of stock (and gave them some compensation). That way they could simply test the product’s economic viability.

For more, these are the slides from a talk I gave on this.

This post started as a thread on X.