Why 80% of software features are never used — and what the other 20% have in common
Pendo found 80% of software features are rarely or never used and 12% drive 80% of usage. It's rarely a functionality gap — and the 20% that stick share three traits decided before any code is written.
Pendo found 80% of software features are rarely or never used, and just 12% of features drive 80% of usage. It's mostly not a functionality gap — and the features that stick share three traits, all decided before a line of code is written.
The 80/20 of features is brutal
- 80% of features are rarely or never used. An average of just 12% generate 80% of daily usage (Pendo, 2019 Feature Adoption Report).
- Public cloud-software companies invested up to $29.5 billion on rarely/never-used features.
- The ratios barely move by company size — Pendo treats them as a proxy for the whole industry.
Cheaper building doesn't shrink that 80%. It grows it. Lower cost to ship means more shipped, and the hit rate doesn't automatically improve.
It's usually not a functionality problem
- Roughly 70% of features go unused because of poor usability, not missing functionality.
- Trying a feature once isn't adoption. Early activity often reflects curiosity, not value — teams mistake early clicks for a win.
- What customers ask for and what they need diverge. People describe problems through a solution they can already imagine, so a feature built to the literal request misses the actual job.
Why features die
- Habits win. Most work runs on autopilot; a feature that demands a new habit is fighting the strongest force in the workflow.
- People pick the least risky option, not the best one — the familiar, the one everyone already uses.
- Environment matters. Great features fail when they ignore the ecosystem the user already works inside.
What the 20% that survive share
Keep reading
Spec Hallucination: The 15-Minute Claim Trace for AI-Written Specs
AI writes specs that sound sourced but often aren't. The Claim Trace is a 15-minute audit that labels every claim Sourced, Inferred, Assumed or Invented before engineering builds from it.
What Counts as Evidence? The Risk Receipt for Value, Usability, Feasibility, and Viability
AI made feasibility cheap to prove. Value, usability, and viability still need their own evidence — and a great prototype demo counts for less than it feels like. Here's the Risk Receipt framework for checking what you've actually retired.
What Product Managers Should Never Delegate to AI
AI can summarize evidence and propose options, but some product decisions should remain human-owned. Here are the judgments PMs should never delegate without review.