Stop Building. Start Removing.

Startups obsess over adding value. Often the bigger lever is reducing the effort it takes to experience it.

At DoorDash, there was a quarter where we stopped building entirely. No new features. No new products. Every engineer focused on fixing tech debt, reducing friction, improving what was already there. It was one of the highest-impact quarters the company had. Not because we shipped anything new, but because we made everything we’d already shipped actually work the way it should have.

Most founders I talk to are focused on the same question: what do we build next? Especially now, when you can point an AI agent at your codebase and ship something new by lunch, the bias toward building more has never been stronger. But the question that moves the needle more, and almost nobody asks, is: what can I remove? What’s standing between my customer and the value I’ve already built?

One of the companies I work with measured this recently. It used to take a new user twenty minutes to reach the moment where the product clicked. They spent two weeks focused on nothing but compressing that number. It’s now three minutes. Same product. Same features. Retention shot up overnight.

Every feature you ship adds surface area. Every surface area adds complexity. Every complexity adds friction between the customer and the thing they actually came for. Conversion slips, retention softens, and the instinct is always to build more to fix it. Sometimes the fix is to remove.

Your customers don’t experience your product. They experience the distance between themselves and your product.

Measure the time it takes a new user to reach the moment where they get it. Put that number on the board deck next to revenue and retention. Then do this: sign up for your own product today. New email, new browser, no saved state. Time yourself. Then watch five real users do it without helping them. Don’t say a word. The places where they pause or click the wrong thing are worth more than your entire roadmap.

Default aggressively. Front-load the value, back-load the setup. Kill every screen that doesn’t earn its place. In an AI world, you can do in a weekend what took DoorDash a full quarter.

The best product isn’t the one that does the most. It’s the one that asks the least of the person using it.