Example weekly issue. Replace with your own series.
Deleting a feature is the cheapest performance work you can do. This one had been live for three months, had nine tests, two paragraphs of documentation, and a section on the landing page that I had rewritten twice.
It also had three sessions in ninety days, all of them mine.
What made me finally look
The analytics were flat and I assumed the feature was just young. Then I checked the number of people who had ever clicked the button, not the number of page views it collected. Three.
What I removed
- The route and its tests.
- The marketing copy.
- The docs page, redirected to the parent section.
- The config flag, which nobody had ever turned on either.
What it cost
One afternoon, mostly spent making sure the redirect was right. The build got smaller. Nothing else changed, which is the point.
The rule I keep relearning
Usage is not traffic. A page that gets visits and no clicks is a page that is failing politely.