Why Story Points Inflate Over Time (And How to Stop It)
If your team's velocity keeps climbing quarter over quarter without a matching increase in shipped output, you're probably not getting better at estimating. You're experiencing point inflation, and it's one of the most common ways velocity stops meaning anything.
How it happens
Inflation rarely comes from one bad decision. It creeps in through small, individually reasonable choices:
- A story that would have been a 3 last year gets called a 5 "to be safe," and nobody pushes back because being generous feels lower-risk than being wrong.
- New team members anchor on whatever number a senior engineer throws out, without an independent read on the original reference stories.
- Stories genuinely get more complex as the codebase grows, but the team never recalibrates what a "3" represents against that new baseline.
- A manager treats velocity as a productivity target, so the team (consciously or not) starts sizing work to hit the number rather than to reflect effort.
Why it matters
Once points inflate, velocity becomes a moving target. A sprint that "delivers 40 points" this quarter isn't comparable to one that delivered 40 points a year ago, which means burndown forecasts, roadmap commitments, and cross-team comparisons are all quietly wrong. The number still looks precise. It just isn't measuring the same thing anymore.
How to reset it
Don't try to force velocity back down; that just punishes the team for a drift that wasn't really their fault. Instead:
- Re-anchor with reference stories. Pick two or three recently completed stories the team agrees were clearly a 2, a 5, and an 8. Use them as the explicit comparison point in the next few sessions, the same way you would when calibrating a new team's scale.
- Separate velocity from performance conversations. If leadership is using velocity as a productivity signal, inflation is the predictable result. Decouple the two and the pressure to inflate mostly disappears on its own.
- Re-baseline after major team changes. New hires, a new codebase, or a shift to a new domain are all good moments to explicitly reset the scale rather than let it drift, similar to how you'd handle onboarding new team members into estimation.
Point inflation isn't a discipline failure. It's what happens when a relative scale is left uncalibrated for too long. A short reset conversation every couple of quarters keeps the numbers meaning what they're supposed to mean.
