Completion rate appears to be the most objective video metric available: the ad either finished or it did not. The number is far less comparable across sellers than it looks.

The denominator is the first disagreement

Some systems calculate completions against impressions served, while others calculate against video starts, and a meaningful gap exists between those two events.

An ad can be served and never begin playing because the user scrolled away, the player failed, or the creative did not load within the allowed window.

Measuring against starts removes those cases from the calculation and produces a materially higher figure from identical delivery.

Player behavior decides how hard completion is

In an environment where the ad cannot be skipped and the content will not resume until it ends, completion approaches the ceiling almost automatically.

In a scrolling feed where the ad autoplays silently and the user controls what happens next, finishing requires a deliberate decision to keep watching.

Comparing those two figures treats a measure of player design as if it were a measure of creative performance.

Audio and visibility conditions differ by placement

A view counted as complete with the sound off and a small portion of the player on screen is not equivalent to one watched full-screen with audio.

Different environments apply different thresholds for how much of the player must be visible and for how long, and those thresholds are not always disclosed in reporting.

Two sellers can therefore report the same completion rate while delivering attention that differs substantially in quality.

Quartile reporting is more informative than the endpoint

Most video reporting includes viewership at the quarter, half and three-quarter points, and the shape of that curve says more than the final figure.

A steep early drop points at the opening seconds of the creative, while a gradual decline across the whole duration points at the length or the story.

The curve also exposes placements where nearly everyone leaves immediately, which usually indicates the environment rather than the ad.

Completion is a proxy that can mislead

High completion where skipping is impossible tells the advertiser little, since the user's alternative was to abandon the content they came for.

Where skipping is available, choosing not to skip is genuine information, and a lower rate in that setting may represent more real attention than a higher one elsewhere.

The useful practice is comparing completion only within a consistent environment and reading it alongside outcomes further down the funnel rather than on its own.