Worksheet 8 — Measurement card

Use with chapter 8, Measuring an intervention. Fill sections 1–5 before the change ships. Fill section 6 when the window closes.

App: ______________ Change: ______________________ Ships on: __________

1. Baseline

Metric (exact definition and source): ______________________ Last 8–12 comparable periods: ____ ____ ____ ____ ____ ____ ____ ____ → normal range ______ to ______

2. Exposure

Who experiences the change: ______________________ Can I measure the exposed group separately? ☐ yes, how: ______________ ☐ no (the total will dilute the effect)

3. Window

From __________ to __________. I will not judge before __________.

4. Confounders in the window

• ☐ holiday/season ☐ release ☐ price change ☐ featuring ☐ ads ☐ community/press post ☐ competitor launch ☐ store outage ☐ metric definition change ☐ other: ______ Comparison group (if no split test): ______________ — remember: the other platform is a hint, not a control group.

5. Verdict rules (write now, don't edit later)

Small-sample check

Counts: assumes independent events, a stable underlying rate and N above ~20. A sanity check, not a significance test.

• One period vs its long-run average N: ±2√N. N = ______ → ±______
• Two periods vs each other: ±2.8√N. N = ______ → ±______
• Better still: the 8–12-period normal range in section 1.

Rates: use the test brief's planning estimate or the store's own test result (its confidence level and minimum detectable effect, under its own assumptions).

6. Result (when the window closes)

• Observed: ______________________
• The metric changed? ☐ yes ☐ no ☐ within normal range
• The goal was reached? ☐ yes ☐ partly ☐ no
• Our action caused it? ☐ strong evidence (split test) ☐ weak (before/after with confounders) ☐ not established
• Verdict: ☐ worked ☐ did not work ☐ harmful ☐ inconclusive
• Copied to the decision journal on __________ ☐