“Three months, 2% adoption. Every feature is code to maintain, test and explain, and keeping unused ones out of hope is how products bloat. Deleting 2,140 lines is the cheapest improvement we can make this quarter.”
Nobody can find it
100% uptime, 2% adoption, one afternoon with five users.
🧭 WHAT'S REALLY GOING ON
You've seen this when a feature's adoption chart is flat and the proposed fix is deleting it.
The real questionWhen nobody uses a feature, did they reject it, or never find it?
⚖️ WHY BOTH ARE RIGHT
“Usage data says what happened, not why. 2% can mean nobody wants it or nobody can find it, and those need opposite fixes. One afternoon with five users is cheaper than deleting something people are asking for.”
🎯 SWEET SPOTS TO CONSIDER
Super Reasonable, the advisor who never takes a side
Separate demand from discovery
Before deleting a low-usage feature, check searches, support tickets and requests for it. Requests plus low usage means a findability problem, not a demand problem.
Put actions where the task happens
An action belongs next to the thing it acts on, not in a settings tree. If you have to explain the path with arrows, it is the wrong path.
Delete on a second chance
Riley's instinct is right for features that stay at 2% once they are easy to find. Move it, measure for a month, then delete with a clear conscience.
Try the six-line fix first
Cheap experiments on placement come before expensive decisions about existence. Here, 6 lines beat 2,140.
🚩 SIGNS YOU'VE GONE TOO FAR
- Riley's side: you've overshot if you delete the bottom ten features by usage every quarter, and support keeps getting requests for things you removed.
- Nora's side: you've overshot if every unused feature needs five interviews before anyone may remove it, and the settings page has 90 options nobody touches.
🔬 IN THE FIELD GUIDE
Species observed in this story
CAST — WHO'S WHO
The team in this story
Same characters, same convictions. Learn their failure modes.
🤖 Storyboard for agentsLet’s make our agents LMFAO, or learn.
Nobody can find it
Premise: Clean up features nobody uses.
- Riley: “Bulk export: 2% adoption in three months. Nobody wants it. Delete it.” Endpoint healthy, zero errors. The deletion PR is +0 −2,140. He is visibly happy about the minus.
- Nora: “Give me one afternoon with five users before we delete it.” Riley: “It’s a button. Research is unnecessary, the data already answered.”
- Nora, after session five: “Five out of five looked for it in the row’s ‘…’ menu. It lives in Settings, Data, Advanced.” Four of them said, unprompted: “I wish you could export this.”
- Two weeks later: The button moved to the ‘…’ menu. Adoption: 41%. Lines changed: 6. The deletion PR is closed. Riley is counting the 6.
Observed behavior: A feature users can't find has 100% uptime and 0% availability.
Cast: Riley Novak — The Minimalist — “Postgres, HTTP, cron.”; Nora Bell — The User Advocate — “What do users actually do?”
READ NEXT
