Not all features add value — but every feature adds cost

Are you working in a feature factory? Quick test. At the end of a quarter, when leadership asks how it went, does your team answer with a list of what shipped — or with what changed for customers?

If it's the list, you've got a factory on your hands. And the factory has a flaw built right into the floorboards: not all features add value, but every single feature adds cost.

The cost you don't see on the roadmap

Every feature you ship doesn't just take time to build. It takes time to maintain, forever. It adds surface area for bugs. It makes the next feature harder to build, because now there's more to reason about, more to not break. Complexity compounds like debt.

So when you ship a feature nobody uses, you don't break even. You go backwards. You've added permanent cost in exchange for zero value. Do that for a few quarters and you've got a bloated, fragile product that's expensive to change and confusing to use.

Here's the number that should stop you cold

Teams that rigorously A/B test their ideas have found that, in the best case, only about one in three ideas produces a measurable improvement. In the typical case, the hit rate is more like 10–15%.

Sit with that. The factory model — optimize for output, ship everything on the roadmap, measure velocity — is running flat-out to produce a backlog where most items don't move the needle. You're not just building the wrong things sometimes. You're statistically guaranteed to be building the wrong things most of the time, then maintaining them anyway.

The shift: pursue goals, not features

The best product companies don't optimize for throughput. They pursue goals and optimize for outcomes. The roadmap stops being a list of features and becomes a list of problems to solve and behaviors to change.

This is a real shift, and it shows up in four concrete places:

• How you brief the team. Not "build the export feature" but "help users get their data out so they trust us enough to put more in." The first dictates a solution. The second invites the team to find the best one — and to prove it works.
• How you measure success. Not "did we ship it" but "did the behavior we cared about actually move." Output is a leading vanity metric. The outcome is the truth.
• What you celebrate. A factory celebrates launches. A value generator celebrates a metric moving — and is comfortable killing a finished feature that didn't earn its keep.
• How much you build. When the goal is an outcome, "build nothing this sprint because the test told us not to" becomes a win, not a failure.

"But we'll look slower"

This is the fear, always. If we stop measuring output, won't it look like we're doing less?

You'll ship fewer features. You'll deliver more value. Those are not the same thing, and the gap between them is exactly where your competitors are wasting their quarters. While they fill their product with features that 85% of the time don't matter, you ship the ones that do — and spend the time you saved on the next real problem.

The factory feels productive because it's busy. Busy isn't the goal. A product customers actually love, that's cheap to evolve because it isn't drowning in dead weight — that's the goal. You don't get there by shipping more. You get there by shipping right.

So change the question at the end of the quarter. Stop asking "what did we ship?" Start asking "what changed?" Everything downstream of that question — what you build, what you measure, what you celebrate — starts to fix itself.