Measure the observed error rate in a reviewed sample of attribution assignments. The calculator reports errors per 100 assignments, correct assignments, and an approximate 95% interval so teams can separate measured quality from sampling noise.
The displayed 95% range uses p̂ ± 1.96 × standard error and is clipped to 0%–100%.
Use the main result together with the supporting bounds, counts, or capacity figures. The estimate is only as reliable as the input definitions, sampling process, and operating assumptions.
Use a representative audit sample. An error rate from hand-picked or unusually difficult cases may not describe production traffic.
38 errors in 500 audited observations gives an observed error rate of 7.60% and 76.0 errors per 1,000.
Use a written audit rule and count only observations that fail it. Keep the rule consistent across the entire reviewed sample.
A normalized rate makes samples of different sizes easier to compare while preserving the underlying observed percentage.
No. Errors must be between zero and the total number of audited observations.
It is an approximate range for the underlying error proportion if the audited observations form a representative independent sample.
Not necessarily. Confirm that sampling, traffic mix, definitions, and audit difficulty remained comparable.
| Metric | Definition |
|---|---|
| Error rate | Errors divided by audited observations |
| Passing count | Audited observations minus errors |
| 95% interval | Approximate sampling range |