Name both parts of the ratio
An illustrative team resolves 18 out of 20 eligible requests in one period and 72 out of 100 in the next. The rates are 90% and 72%. The second period has a lower rate and a larger number of resolved requests. Neither observation cancels the other; they describe different aspects of the activity.
Show the counts beside the rate where practical. “72% · 72 of 100 eligible requests” tells a reader more than an isolated percentage. Explain what makes a request eligible and which event assigns it to the period.
Do not average rates without checking weights
Suppose one group completes 9 of 10 items and another completes 1 of 2. Their rates are 90% and 50%. The simple average is 70%, while the combined rate is 10 of 12, approximately 83.3%. The two calculations answer different questions.
Use combined counts when the question is the rate across all eligible items. Use an average of group rates only when equal weighting of groups is intentional, and label that choice. A reporting tool cannot choose the intended interpretation on its own.
Keep the denominator consistent with the question
A completion measure can follow items received during a period or items completed during a period. The first is a cohort view and may change as that group progresses. The second is an activity view and may include older work. Comparing them as if they were the same measure is misleading.
Eligibility also changes when canceled, duplicate, test, or incomplete records are excluded. Record those choices in the metric definition. If exclusions increase, a rate can rise even while the underlying work stays unchanged.
Handle small and empty populations
A single item can move a small group's rate sharply. Display its count and avoid presenting a small sample as a stable trend. Where a minimum reporting population is needed, agree it for the operational context and show when the view falls below it.
When the denominator is zero, a percentage is undefined. Show “Not applicable” or an equally explicit state rather than 0%. Zero percent implies eligible items existed and none met the condition; that is a different statement.
Make comparison rules part of the view
Before comparing periods or groups, check that eligibility, time windows, units, and definitions match. If they do not, provide the distinction in the interface rather than relying on a separate document that a reader may never open.
The figures here are fictional arithmetic examples, not Liberty Trading LLC performance statistics. The broader habit is simple: keep the population close to the percentage and the calculation close to its purpose.