UA Ledger · Measurement Lab · draft for review

Scale a holdout outcome rate before subtracting conversions

Synthetic teaching inputs. No real campaign results or industry defaults. Change the values to inspect the model. Calculations run locally; this page sends no data and does not retain inputs after closing.

Use eligible assigned people in each group and a shared conversion definition. Divide holdout conversions by holdout population, multiply that rate by treatment population, then subtract the resulting expected count from treatment conversions. Keep the rate difference visible alongside the count difference. Do not use reached users in one arm and assigned users in the other.

Calculated outputs

Assumption sensitivity

Only cc varies; all other inputs are held constant. This is not a probability range.

Equations

• Scaled holdout conversions: cc/cn*tn
• Descriptive excess conversions: tc-cc/cn*tn
• Rate difference percentage points: 100*(tc/tn-cc/cn)

Additional constraints: (tc<=tn) && (cc<=cn). The field minimums also apply.

Limits

Random assignment, comparable observation and limited spillover are unverified. The worksheet does not model sampling uncertainty, noncompliance, household clustering or interactions with concurrent campaigns.

Confidence intervals for a proportion · primary documentation checked 19 September 2026.