We audited our product and found 40% of our features were used by less than 5 people.

by•

Last week, I finally sat down with our engineering team to do a "feature audit." We wanted to see which parts of our SaaS were actually driving value and which ones were just bloating our codebase.

The results were horrifying. Out of the 30 major features we've built over the last three years:

  • 12 features (40%) were used by fewer than 5 active users.

  • 4 features had literally zero usage in the last 6 months.

  • Every single one of these features still required us to maintain the code, fix bugs, and handle support questions.

I realized we had been building features based on the loudest 1% of users—the ones who send 20 emails a day demanding a specific integration or a custom workflow. We built those things to make them happy, and in the process, we turned our product into a bloated, slow, confusing mess for the other 99% of users.

Worse, every time we need to update our core product, we have to spend weeks untangling the spaghetti code of these dead features. I estimate we're spending $3,000/month in engineering time just maintaining things nobody uses.

Now I want to do a "feature sunset"—actually rip these features out of the product. But I'm terrified of the backlash from those 5 users who might leave, and I'm worried about breaking something critical in the codebase.

For founders who have done a feature audit: How did you actually kill dead features without breaking your product or your relationship with your power users?

Did you just hide them? Did you announce a deprecation schedule? I need to figure out how to clean up this mess before it drags us down.

5 views

Add a comment

Replies

Best

I felt this with my own calendar app. I built it because Fantastical didn't do what I wanted, and every time I thought about adding a competitor's feature, I could feel it getting bloated. For a calendar, being easy to read matters more than anything, so that's where the effort goes. My other filter is simple: I don't build what I won't use myself.