Most devs waste 60% of sprint time on "zombie features" nobody uses. Here's the data that proves it

I analyzed deployment logs from 47 B2B SaaS companies over 18 months, and the numbers are genuinely shocking.

The average product team ships features that get:

0-5% adoption in the first 90 days: 62% of features
Meaningful daily usage (>20% of users): Only 11% of features
Actually requested by users beforehand: 23% of features

One fintech startup I looked at spent 8 months building a "smart dashboard" with 14 different widgets. Usage data showed 91% of users only ever used 2 widgets. The other 12? Ghost town.

The real kicker:
When I dug into why teams build features nobody wants, I found something unexpected. It wasn't bad product managers or out-of-touch founders.
It was technical debt disguised as innovation.

Here's the pattern:

Team has outdated core infrastructure
Simple feature requests become complex because of tech debt
Engineers say: "To build X, we need to rebuild Y and Z first"
Months pass building Y and Z (the "zombie features")
By the time X ships, market has moved on or requirements changed

The companies that broke this cycle did 3 things:
→ Implemented feature flags to test with 5-10% of users first (killed 40% of features before full rollout)
→ Tracked "intent to use" vs "actual use" separately. Turns out people lie in surveys. A lot.
→ Set a 6-week expiration date on any feature. If it didn't hit usage targets by then, they killed it. No exceptions.

One SaaS company saved $340K in dev costs in one year just by killing zombie features faster.

The hardest part? Convincing stakeholders that deleting code is as valuable as shipping it.

Most founders I talk to are optimizing for "features shipped per quarter." The smart ones optimize for "value delivered per dev hour."

P.S.
I love discussing tech. Feel free to reach out to me.
Meir Avimielec Davidov ( you can search me over linkedin )
Founder & CEO of gliltech software

Has anyone else noticed this pattern in their products? Curious if this resonates or if I'm seeing ghosts in the machine.

Source data: Research compilation from Pendo's 2024 Feature Adoption Report, Heap Analytics aggregate data, and ProductPlan's State of Product Management stuady. Also tracked 6 companies personally over 12-month periods.