# Compare short-video completion rates without mixing durations

Record duration alongside the completion measure and its actual denominator. Compare like definitions and suitable duration groups before treating a rate difference as a content decision. Completing a brief clip is a different viewing task from completing a much longer explanation.

Comparison question:
Platform and source report:
Exact completion definition:
Exact denominator definition:
Observation window:
Video A duration and topic:
Video A completion numerator and denominator:
Video B duration and topic:
Video B completion numerator and denominator:
Duration group or matching rule:
Audience and traffic differences:
Other viewing measures available:
What the comparison supports:
Next decision:


"record_type","video","duration_seconds","completed_measure","denominator","descriptive_rate_pct","definition","window","context"
"fictional example","A","20","50","100","50","Matched example measure","same example window","Different duration from B"
"fictional example","B","45","40","100","40","Matched example measure","same example window","Do not interpret duration difference as a controlled content result"
"your observation","","","","","","","",""


## Check comparability first

"Dimension","Record","Why it changes interpretation"
"Duration","Actual runtime and chosen group","A longer video requires more viewing to complete"
"Completion","The source's exact numerator","Different reports may measure different viewing events"
"Denominator","The reported starting or view measure","A percentage is meaningful only with its base defined"
"Window","Equal observation age or agreed reporting period","A fresh upload and a mature upload have different exposure histories"
"Audience","Traffic and distribution context","Different audiences can respond differently to the same explanation"


## Read the fictional rows

The examples give A a 50% descriptive rate and B a 40% rate, using the same assumed denominator definition. A is also 20 seconds while B is 45 seconds. The rows describe observations; they do not isolate an editing change as the cause.

When your source does not provide the required numerator or denominator, retain its reported metric and definition instead of inventing the missing counts. Use viewing duration or another available measure as a separate observation, not a substitute with the same label.

## Report the decision honestly

- [ ] Duration and definitions are recorded.

- [ ] Different sources are not silently combined.

- [ ] A zero denominator is marked unavailable.

- [ ] Audience and publication differences remain visible.

- [ ] The conclusion stays within what the comparison measured.