What Video Completion Rate Counts on Meta, TikTok and YouTube

What is video completion rate?

Completed plays divided by started plays is what everyone means by video completion rate, and that is the last thing about it any two platforms agree on. Both halves of the fraction are defined by the platform reporting them, so the same upload can produce two different percentages in two dashboards on the same day without either one being wrong. None of the four biggest video ad platforms even ships a column with that name. They ship quartile counters, and you do the division yourself.

Which platforms publish a metric called completion rate?

None of the major ones. Meta, TikTok and Google Ads all publish a count of plays that reached 100%, and leave the rate to you. YouTube Analytics does not publish a completion count at all.

Read that first row twice. Somebody who dragged the scrubber to the last second registers the same way as somebody who watched every frame.

How does Meta count a play at 100%?

Meta counts a play at 100% whenever the playback position reaches the end, whether it arrived there by playing or by skipping. The ladder below is Meta's own, from its "About video ad metrics" page.

Meta is the only platform here with a 95% counter next to the 100% counter, and that extra rung exists for a reason. On a separate page, "About ThruPlay," Meta states the rule outright: "A video's full length is considered to be at least 97% of the video, since people frequently drop off when credits roll or content fades out before the end of a video."

So Meta holds two definitions of a finished video at once. For ThruPlay, 97% is finished. For the quartile counter, 100% is finished, and a scrub counts. A fifteen second ad watched to 97% produces a ThruPlay and no play at 100%.

The two Meta pages also disagree on ThruPlay itself. "About video ad metrics" describes one fifteen second threshold. "About ThruPlay" describes two goals, a 6s one and a 15s one, each counted when "someone watches at least" that many seconds "or to 97% completion if shorter." Both pages are live. Neither points at the other.

How does TikTok count video views at 100%?

TikTok counts a video view at 100% when the video was played 100% of its length, per impression, with replays excluded. The wording matters because TikTok changed it between rungs of the same ladder.

The 25%, 50% and 75% rows all say "at least." The 100% row drops it. TikTok never explains the difference, and the page carries a "Last updated: November 2025" stamp, so this is current text rather than a leftover. TikTok also never says whether a scrub to the end counts, which is the one thing Meta does say.

The denominator is the part people get wrong here. Video views excludes replays and is counted per impression, and TikTok defines an impression session so that swiping away and swiping back "is counted as 2 impressions." On a platform built around the swipe, the denominator inflates for exactly the content people return to.

What does YouTube report instead of completion rate?

YouTube Analytics reports averageViewPercentage, described in the Analytics API reference as "The average percentage of a video watched during a video playback." There is no quartile counter and no completion counter anywhere in YouTube Studio.

The detail that trips people up sits in the YouTube Studio help text: "Average percentage viewed: Average percentage of a video watched among those who stayed to watch. This is calculated from engaged views and their corresponding watch time." The denominator is engaged views, not views. Since August 24, 2026, YouTube counts a view "the moment a video starts to play across all formats," so views and engaged views are now different numbers on the same video, and the completion figure is built on the smaller one. What YouTube engaged views count is worth reading before you trust any YouTube retention figure.

Google Ads does publish quartiles for YouTube ads, along with the formula: "Video Played to 25% Percentage = # of viewers still watching after 25% of the video has elapsed / # of times the video player is initiated." The denominator is player initiations, not impressions and not views.

Google then undercuts its own metric in the same document, warning that quartile reporting "should not be used to back into your view metric to confirm accuracy of either metric" and is "intended to be directionally helpful." One further rule catches in-feed advertisers: "for in-feed and Masthead, only completed views from clicks to the watch page are counted, and not completed auto-play views." An autoplay that runs to the end in the feed does not increment the counter.

What does the VAST spec call a completed view?

The IAB VAST specification is the closest thing to a neutral definition, and it is the one every ad server implements. VAST 4.3 defines five playback tracking events on a linear creative:

<Tracking event="firstQuartile">...</Tracking>
<Tracking event="midpoint">...</Tracking>
<Tracking event="thirdQuartile">...</Tracking>
<Tracking event="complete">...</Tracking>
<Tracking event="progress" offset="00:00:10">...</Tracking>

The spec defines the three quartiles as the creative having "played continuously for at least" 25%, 50% and 75% "of the total duration at normal speed," and defines the fourth as: "The creative was played to the end at normal speed so that 100% of the creative was played."

Two words there do work no platform metric does. "Continuously" rules out stitching together scattered seconds. "At normal speed" rules out a 2x playthrough. Meta's 100% counter explicitly allows the scrub that VAST's wording excludes, which is the largest single reason a demand side platform and Meta Ads Manager will not agree on the same creative.

VAST also offers progress with an offset attribute, which lets a buyer define completion at any point, including a percentage. Ask anyone quoting a completion rate off a VAST feed which event they fired on.

Why do two tools report different completion rates for the same video?

Because both halves are platform specific and the gaps compound. Four independent decisions sit inside one percentage:

1. Does a skip to the end count? Meta says yes in writing. VAST says no in writing. TikTok and Google Ads do not say.
2. What is the denominator? Meta and TikTok use plays or views with replays excluded. Google Ads uses player initiations. YouTube uses engaged views.
3. What counts as the end? 100% for quartile counters, 97% for Meta ThruPlay, and whatever offset you chose on a VAST progress event.
4. Do replays count? Excluded from the denominator on Meta and TikTok, which quietly raises the rate on rewatchable content.

This is the same class of problem as the gap between reach and impressions, with the same fix: pick one platform's definition, write it next to the number, and never average across platforms.

How should I calculate it consistently?

Fix the definition before you fix the target, and write both halves of the fraction into the report:

completion rate = plays that reached 100% / plays that started

Then name which counter filled each slot. Video plays at 100% / Video plays on Meta. Video views at 100% / Video views on TikTok. On Google Ads you divide nothing, because Video played to 100% already arrives as a percentage of player initiations.

Three habits keep the number honest. Compare a video only against other videos on the same platform. Segment by placement, since an in-feed autoplay and a click to a watch page follow different rules on Google Ads. And pair the rate with a duration metric such as Meta's Video average play time or TikTok's Average play time per video view, because a high completion rate on a six second clip and a low one on a four minute film can represent the same total attention. On its own, completion rate drifts toward the vanity metric end of a report.

Analytics in AdaptlyPost pull each network's own figures through each network's own API, alongside per network views such as TikTok analytics. The counting rules above belong to the platforms, and no tool changes them. What one dashboard removes is the step where somebody drops a Meta percentage into the same column as a TikTok percentage.

Frequently asked questions

Is video completion rate the same as retention rate?

No. Completion rate is one percentage describing how many plays reached the end. Retention is a curve describing where people left. A completion rate says nothing about whether the drop came at second three or second thirty, which is why YouTube ships an audience retention graph and no completion column. The broader sense of the term is covered in what retention rate means.

Does a replay count toward completion rate?

Not in the denominator on Meta or TikTok. Meta's Video plays metric "excludes replays," and TikTok's Video views says "replays are excluded." YouTube's averageViewPercentage "excludes looping clips traffic" as of December 13, 2021. None of the platforms state clearly whether a replay that reaches the end increments the 100% counter.

What is a good video completion rate?

No platform publishes a benchmark, and any figure in a third party roundup was calculated against an unstated denominator on an unstated platform. The only comparison that holds is your own video against your own videos, same platform, same placement, same period.

Does Meta count a scrub to the end as a completion?

Yes, and it says so. Every Meta quartile metric carries the phrase "including plays that skipped to this point," the 100% metric included. If you need a completion figure that excludes scrubbing on Meta, no published metric gives you one.

Why does my Google Ads completion rate not match my view count?

Because Google says it will not. The YouTube ad metrics documentation states that quartile reporting "should not be used to back into your view metric to confirm accuracy of either metric, as it's likely to vary slightly," and that quartiles are "intended to be directionally helpful."

Where does completion rate sit against engagement rate?

They answer different questions. Completion rate measures whether the video held attention to the end. Engagement rate measures whether people did something afterwards. A video can run to the end and collect no comments, and a video can be abandoned early and drive every click you got that week.

What denominator does Google Ads use for its video quartile metrics?

Google Ads defines each quartile as viewers still watching after that point divided by the number of times the video player is initiated. That denominator is player initiations, not impressions and not views. The same page warns against using quartiles to back into your view metric, since they are less accurate and likely to vary slightly.

Does TikTok say whether a scrub to the end counts as a view at 100%?

TikTok's documentation never addresses it. The 25%, 50% and 75% metrics all carry the phrase "at least," but the 100% metric drops it without explanation, and the help page carries no note on scrubbing either way. Meta is the only one of the four platforms that puts its position on scrubbing in writing.

What is the VAST progress tracking event for?

progress fires at a timestamp the buyer sets with an offset attribute, rather than at a fixed quartile. That lets whoever configured the VAST tag define completion at any point in the video, including a percentage that has nothing to do with 25, 50, 75 or 100. Anyone quoting a completion rate off a VAST feed should be asked which event produced it.

Why do Meta's two help pages give different ThruPlay definitions?

"About video ad metrics" describes ThruPlay as one flat fifteen-second threshold. "About ThruPlay" describes two separate goals, a 6-second one and a 15-second one, each triggered by watching that many seconds or reaching 97% completion if the video is shorter. Both pages are live at once, and neither one corrects the other.