Video Completion Rate 2026: Seven Definitions Audited

01

Video completion rate definitions at a glance

Download the video completion rate definition dataset. Each row preserves the published definition, denominator, replay treatment where disclosed, scope or methodology change, official source, and verification date.

Sources: Campaign Manager 360, Google Ad Manager, LinkedIn Campaign Manager, TikTok Ads Manager, and YouTube Analytics, checked September 4, 2026.

Verification date:

• 2026-09-04

02

Seven definitions compared

The last three rows are included because they are frequently mislabeled as completion rate in secondary articles and dashboards. YouTube's average percentage viewed and Wistia's average engagement can be high when no individual viewer completes. TikTok's 100% view metric is a numerator until an analyst chooses a denominator. Vidyard's published benchmark uses end-retention wording but does not share the same formula as the Google ad systems.

03

Campaign Manager 360 uses video plays

Google's Campaign Manager 360 benchmark defines video completion rate as:

Video completions ÷ video plays × 100

The page describes an industry benchmark using the last 90 days. This play-based denominator asks: among the times playback began under the system's play definition, how many reached completion?

If 1,000 plays produce 420 completions, the Campaign Manager-style VCR is 42%. Do not compare that number directly with an impression-based rate unless starts and impressions are identical—which they rarely are.

04

Google Ad Manager uses video starts

Source: TapVid public research dataset; verified 2026-09-04.

Google Ad Manager's interactive-report metric lists completion rate as video completes divided by video starts. This is conceptually close to a play-based rate, but event definitions can still differ. One system may count autoplay as a play, require a minimum visible state, or deduplicate replays differently.

The safe data-model choice is to retain the raw numerator and denominator event names. Store `video_completes`, `video_starts`, and the derived rate rather than importing only a column called `completion_rate`.

05

LinkedIn uses impressions—and changed history

LinkedIn Campaign Manager currently defines Completion Rate as completions divided by video impressions. Its help page says the calculation changed in March 2025 from completions divided by plays to completions divided by impressions. LinkedIn applied the current definition to historical data as well.

This creates two analytics risks.

First, a LinkedIn completion rate will usually be lower than a start-based rate when many impressions never begin meaningful playback. Second, a previously exported report can disagree with a newly exported historical report even though the campaign events did not change. The formula changed.

Record the extraction date and metric-definition version beside any LinkedIn benchmark. A year-over-year chart without that metadata may show a methodology break as creative performance.

LinkedIn also defines a completion as a video ad that played through 97–100% of its length, including watches that skipped to that point. “Completion” therefore does not always mean continuous viewing of every prior second.

06

TikTok publishes a 100% view count

TikTok defines Video views at 100% as the number of times a video was played through 100% of its length. Plays are counted separately for each ad impression, and replays are excluded. The same documentation defines 25%, 50%, and 75% milestone counts.

The official page describes counts, not one universal “completion rate” formula. An analyst could divide 100% views by impressions, initial video plays, two-second views, or another event. Each result answers a different question.

If a dashboard derives TikTok VCR, name the denominator in the metric label—for example, `100% views / impressions`—and keep the original fields available. Never import a derived rate as though TikTok published that exact formula.

07

YouTube average percentage viewed is not completion rate

YouTube Analytics defines average percentage viewed as the average percentage of a video watched among people who stayed to watch. For Shorts, the metric is calculated from engaged views and their corresponding watch time, and loops are excluded from engaged views.

A 60% average percentage viewed does not mean 60% of viewers completed. Ten viewers could each watch the first 60%, or some could finish while others leave early. The metric describes average consumption depth, not the distribution of completions.

Use YouTube audience-retention and milestone information when the question is where viewers leave. Use average percentage viewed to compare consumption depth among similar formats. Do not relabel it VCR.

08

Wistia average engagement is also a different metric

Source: TapVid public research dataset; verified 2026-09-04.

Wistia calculates average engagement as total time played ÷ (total plays × media length). Its published benchmark sheet reports 50% for videos below one minute and progressively lower averages for longer buckets.

That statistic is useful for understanding how much of a hosted video was consumed. It is not the fraction of viewers reaching the final frame. Rewatches and partial viewing can influence time played even when completion count is unchanged.

The companion video engagement benchmark study keeps Wistia and Vidyard figures separate and shows why label normalization should happen before analysis.

09

Published completion benchmarks need their source context

Vidyard's 2025 Video in Business Benchmark says 65% of viewers stay engaged to the end for videos under one minute and 20% for videos over 20 minutes. The underlying report covers anonymized customer data and 943,305 videos created in 2024.

These figures are useful within a business-video and sales-video context. They are not a substitute for an ad-platform VCR because the population, placement, viewer intent, and denominator differ. A prospect who clicks a one-to-one sales video starts with more intent than an impression served in a feed.

When quoting a completion benchmark, preserve five fields:

1. platform and product; 2. metric formula; 3. length bucket; 4. sample and observation period; 5. traffic and placement context.

If any field is unavailable, say so rather than filling it from another report.

10

Build a completion-rate metric contract

Use a data dictionary that makes comparison explicit:

For internal comparisons, calculate a common metric only when the raw events are semantically compatible. If every platform provides impressions and 100% milestone counts under acceptable definitions, an impression-based rate can be built. Label it as an internal normalized metric and keep platform-native rates beside it.

11

How to use VCR without optimizing for the wrong thing

Completion is most useful when the final seconds contain necessary information. If the CTA, price, or legal qualifier appears only at the end, low completion means many viewers never saw it. The first fix may be information order, not a shorter animation.

For product and explainer video, pair VCR with:

• average watch time or percentage viewed;
• first-quarter drop-off;
• clicks or conversion assists;
• approved product-claim exposure;
• the actual downstream task, such as signup, demo request, or completed setup.

Keep the source copy and product assets stable when testing pacing. TapVid makes approved assets, script, and scene mapping visible before output and allows a specific scene to be revised while preserving the rest. That helps teams create controlled variants, but the product does not guarantee a completion rate. Performance must be measured with the actual audience and placement.

12

Methodology

Source: TapVid public research dataset; verified 2026-09-04.

We selected first-party platform or product documentation that defines a completion-related metric or a commonly confused watch-depth metric. One row represents one reporting system and its current public definition. We transcribed the numerator, denominator, replay rule, methodology change, and scope when stated.

We did not force the systems into one formula. TikTok's 100% milestone remains a count. YouTube average percentage viewed and Wistia average engagement are marked as non-completion measures. Vidyard's public end-retention benchmark is retained with its own wording. All pages were checked September 4, 2026. Dataset version: 1.0.

13

Limitations

This is a documentation audit, not an independent event-level validation. Platform implementations may contain conditions not fully described on public pages. Definitions can change, and some pages apply changes to historical data. Organic video, paid video, hosted business video, and one-to-one sales video have different populations. The dataset does not provide a universal benchmark or prove that any length causes higher completion.

Before using the table in reporting, verify the current source, export the raw numerator and denominator, and save the extraction date.

14

References and citation

@techreport{chen2026videoCompletionRate,
  author = {Kenneth Chen},
  title = {Video Completion Rate 2026: Seven Definitions Audited},
  institution = {TapVid Research},
  year = {2026},
  month = {September},
  url = {https://tapvid.ai/blog/video-completion-rate},
  note = {Version 1.0}
}

• Campaign Manager 360 completion benchmark
• Google Ad Manager interactive report metrics
• LinkedIn Campaign Manager video metrics
• TikTok video play metrics
• YouTube content performance metrics
• Wistia guide to video metrics
• Vidyard 2025 Video in Business Benchmark
• TapVid definition dataset

Recommended citation (APA): Chen, K. (2026, September 4). Video completion rate 2026: Seven definitions audited (Version 1.0). TapVid Research. https://tapvid.ai/blog/video-completion-rate

BibTeX:

@techreport{chen2026VideoCompletionRate, author = {Kenneth Chen}, title = {Video Completion Rate 2026: Seven Definitions Audited}, institution = {TapVid Research}, year = {2026}, month = {September}, url = {https://tapvid.ai/blog/video-completion-rate}, note = {Version 1.0} }

15

Frequently asked questions