# Read newsletter analytics without mixing dates, subscribers and email

All scenarios and values below are original constructed teaching examples. No real account action, subscriber, reader, customer, delivery or performance outcome is established. Keep real completed notes private. These files do not publish, change settings, contact anyone or retrieve account data.

## 1. A period flow and a current stock

Constructed input: comparable seven-day article views are 14 then 21. New subscribers in the second period are 3. A separate all-subscriber panel concerns 40 current subscribers. Unsubscribe counts and the earlier subscriber total are unavailable.

Calculation: views increased by 7 and by 50% of the earlier 14. Do not infer that 7 distinct new people read, that 40 subscribed in the selected week, or that net subscriber growth is exactly 3. The panels describe different units and scopes.

Report: “Same seven-day view metric: 14 to 21; 3 new subscriptions recorded in the later period. Completion, unsubscribes and payments unavailable.” No real newsletter was analyzed.

## 2. Zero and unknown are different baselines

Constructed input A: article views are 0 then 9. Input B: the earlier value is unavailable and the later value is 9.

Decision A: report nine additional views with undefined relative percentage. Decision B: report only the observed nine views, with neither a supported absolute difference nor relative percentage. Do not replace missing data with zero.

The optional offline checker reproduces these arithmetic boundaries. It does not retrieve private analytics, verify delivery or evaluate reader understanding.

## 3. An edition email metric is missing

Constructed input: the subscriber total is visible, but an edition’s email field is unavailable. A colleague suggests dividing page-wide article views by all subscribers as a substitute open rate.

Decision: reject the substitute. The numerator, denominator, edition and observation scope are not matched. Record the field as unavailable and use only the actual appropriately scoped evidence for the question.

Record: neither zero delivery nor a known open rate established. Real delivery diagnosis and verified customer attribution remain separate tasks; no subscribers are contacted or exported.