Feature Flag Rollout Percentage Calculator
Turn a feature-flag rollout percentage into real user counts, and preview a staged ramp from canary to full launch.
How it works
Exposed users = total users × rollout% · A staged ramp typically steps 1% → 5% → 10% → 25% → 50% → 100%
A percentage rollout hashes each user into a bucket so the same people consistently see the feature as you ramp — it is not a fresh random draw each request. Multiplying the percentage by your user base tells you the real blast radius: a 1% canary of 100k users is still 1,000 people, enough to surface errors before a wide launch. Ramping in stages lets you watch error rates and latency at each step and roll back instantly by lowering the percentage.
Worked example
With 100,000 users at a 10% rollout, 10,000 see the new feature and 90,000 remain on the old path. A 1% canary would expose 1,000 users first — enough signal to catch a regression cheaply.
Frequently asked questions
Does a percentage rollout pick new users each time?
No. A good flag system hashes a stable user ID into a bucket, so a user in the exposed set stays exposed as you hold the percentage, and only additional users are pulled in when you raise it. That consistency is what makes staged ramps and A/B tests valid.
What is a good ramp schedule?
A common pattern is 1% → 5% → 10% → 25% → 50% → 100%, pausing at each step to watch error rate, latency and key metrics. Small early stages catch catastrophic bugs cheaply; later stages confirm the feature holds up at scale.
How is this different from an A/B test?
A rollout is about safe delivery — gradually turning something on for everyone. An A/B test holds a control group to measure impact. You can combine them: roll out to a percentage, keep the rest as control, and compare.
Related calculators
This calculator is for educational purposes. Double-check important results before acting on them.