Describe the counted action
A conversion rate is not self-explanatory. Write down the action in the numerator and the population or opportunities in the denominator. For a fictional enquiry form, completed requests divided by form visits asks a different question from completed requests divided by all site visits. State whether repeated actions are included and which period is covered. Avoid borrowing a familiar label from another dashboard without checking whether it uses the same definitions, scope and rules for counting activity.
Keep comparisons compatible
Before placing two rates beside each other, compare their definitions and coverage. One period may include a different audience or a changed form. Collection gaps may affect the denominator differently from the numerator. A percentage can look precise while those underlying choices remain uncertain. NIST's experimental design guidance provides a foundation for planning comparisons carefully; this worksheet does not replace a suitable statistical analysis or establish that an observed change was caused by a marketing edit.
Create a small metric card
For an important rate, record its purpose, formula, source, exclusions and known limitations in one place. Use a fictional example to check the arithmetic and make clear that the example is not a reported business result. Ask a colleague to explain the rate using only the card. If they reach a different interpretation, revise the definition before sharing the chart more widely. A useful number lets the team understand the question it answers and the neighboring questions it leaves unresolved.
NIST: introduction to experimental design ↗
This article combines general educational guidance with original planning exercises. It does not claim first-hand field testing or an endorsement from the linked organization. Sources checked October 2026.
