A tool for your next product review

Did the metric improve—or did the comparison change?

Bring two groups of completion counts. Check what the change actually says, then leave with a brief you can use in a release review.

No account. No AI score. Inputs stay in this page and disappear when you reset or reload. Use aggregate counts; leave out personal or confidential details.

1. Define a completed task

Use a distinct user or team once in each group. This calculator is for a binary outcome, not revenue, averages or repeated event counts.

2. Compare like with like

Group A

Group B

3. Check the evidence

Tick only what you have verified. Leaving a check empty is useful: it identifies the next thing to resolve.

Method and sources

Completion rate = completed eligible users ÷ all eligible users × 100. Percentage-point change is group B's rate minus group A's. Relative change divides that difference by group A's rate and is undefined when that rate is zero. Values are displayed to one decimal place; calculations use unrounded rates.

The denominator and task-completion checks draw on GOV.UK's completion-rate guidance. The caution about changing populations and short-term metrics draws on Microsoft Research's Pitfalls of Long-Term Online Controlled Experiments. The sequence and brief are Product Mastery's practical adaptation, not a validated scoring model.

Prepared by Product Mastery / Response Shift. Method reviewed 25 October 2026. Placeholder phrases illustrate input format; they are not customer cases or evidence of results.

Practise the judgment behind the numbers

The seven-day plan includes a published value-metrics exercise with AI feedback, alongside prioritisation and stakeholder practice. Free-plan allowances apply.

See the three-exercise practice plan →