Building fast isn’t the hard part. Knowing what to stop building is.

In the early days, everything from customers seemed like an urgent request.

We'd throw in little tweaks, integration hooks, options and workarounds because saying “not yet” was the equivalent of not listening to our users.

And the product became more complex and difficult to understand and maintain along the way.

Today, we always ask ourselves three things before building a feature:

Who is asking for it?

What does it help fix?

What becomes more difficult by building it?

Saying no has become easier.

Ironically, that hasn’t made us any less customer-oriented. But it has made our decision-making process much more thoughtful.

When do you say yes to a feature request?