The second customer breaks it all

You think you have a product once the first customer says yes. It feels like you finally built something real. You start believing all your assumptions were right.

But that first customer is almost always a special case.

They probably found you through a friend. Or they were desperate. Or you showed up at the right time. They'll put up with bugs and missing features because they want you to win.

The first customer allows you to keep going. It shows you that someone needs what you're building.

The real start is when you go looking for the second one. And the third. Do it on purpose and don't wait. Every new customer breaks your assumptions in a different way. You want that to happen while you still have the energy and money to fix things.

The second customer doesn't know you and doesn't care about your story. When you try to sell to them you find out your demo only works for customer one. They need integrations you've never heard of. Security audits. SLAs. And surprise, you have none of that.

It feels like failure. It's actually really useful information. Every request shows you where the product is too rigid. So build on your assumptions but leave some room on purpose.

• Make your data model extensible for customers you have not met yet.
• Let your configuration be flexible even if it adds complexity.

Assume most companies will want some customization.

We started getting conversations about our second customers a few weeks ago and we learned the lesson the hard way.

So go find more customers. If you can, skip anyone who looks just like customer one. Look for a different data volume or team size or workflow.

That difference will teach you a lot.

Think of the first customer as a guess and the second as the first test. By the fifth or sixth you'll start seeing patterns.

Requests that repeat become your roadmap. One-offs get a polite no. That is if you can live with losing the customer over it.

You can't see a pattern with one data point. So go find that second customer before you polish the demo too much. Let them tell you what's missing.

Then fix it and go find the next one.