Video Completion Rate: The Formula, Real Benchmarks, and How to Read It

Most people learn video completion rate in about thirty seconds, then spend months optimizing the wrong thing. The metric is simple. Reading it correctly is not.

Here is the whole formula, and then the part nobody tells you: a 95% completion rate can mean your ad is working, or it can mean nothing at all. It depends entirely on what kind of ad it is and what happened after the view. Let's fix that.

What video completion rate actually measures

Video completion rate (VCR) is the percentage of video plays that ran all the way to the end. If 1,000 people started your video and 600 watched it through, your VCR is 60%.

That's it. It tells you one thing: of the people who started, how many stayed. It does not tell you whether they cared, clicked, or bought. Keep that separation in your head, because it's where most of the mistakes come from.

The video completion rate formula

VCR = (Completed Views / Video Starts) x 100

Some platforms use impressions in the denominator instead of starts. That distinction matters more than it looks:

• Completed views / video starts measures how well the video held people who actually began watching. This is the version you want for judging creative.
• Completed views / impressions mixes in people who were served the ad but never really started it. This drags the number down and muddies what you're measuring.

When you compare VCR across campaigns or platforms, confirm you're using the same denominator each time. Comparing a starts-based VCR to an impressions-based one is how teams end up in pointless arguments about "why did completion drop."

A worked example, starts-based:

• Video starts: 20,000
• Completed views: 9,000
• VCR = (9,000 / 20,000) x 100 = 45%

Nothing complicated. The complexity lives in what a "good" 45% or "bad" 80% actually means.

Video completion rate benchmarks by platform

Benchmarks are useful as a sanity check and dangerous as a target. Here's the current lay of the land, pulled from ad-platform reporting:

CTV reporting from MNTN puts average completion between roughly 90% and 97%, with non-skippable pre-rolls often clearing 98%. YouTube ranges come from Pixability's advertiser benchmarks. The gap between a 95% CTV number and a 20% Meta feed number isn't a quality gap. It's a format gap. Which brings us to the real problem.

The three ways video completion rate lies to you

1. Non-skippable inflation

A 6-second bumper or a non-skippable CTV pre-roll hits 90-95% completion because the viewer has no choice. The number tells you the format worked, not that your creative did. If you brag about a 96% VCR on non-skippable inventory, you're crediting your ad for a floor the format guarantees.

The tell: high VCR that barely moves no matter what creative you run. If every ad in the placement lands in a tight 90-96% band, VCR has stopped being a signal. Judge those placements on clicks, view-through conversions, and brand lift instead.

2. Autoplay and muted starts

On feed placements, a "start" can be triggered by an autoplaying, muted video the user never chose to watch. That inflates your denominator with junk starts and tanks VCR, or, depending on how the platform counts, inflates completions from people who scrolled past with sound off. Either way, feed VCR is noisy. Treat it as directional, not precise.

3. Format confounds length

A 90% completion on a 15-second TikTok and a 90% completion on a 3-minute YouTube ad are not the same achievement. Longer runtime almost always means more drop-off. Wistia's benchmark data shows videos under one minute average around 65% engagement, while longer content falls off from there. Never compare VCR across videos of very different lengths without adjusting your expectations for runtime.

The read that actually matters: VCR against conversion

VCR alone is half a sentence. Pair it with what happened after the view, and it starts telling you what to do. Here's the diagnostic I'd hand any performance marketer:

The two off-diagonal cells are where the money hides. High VCR with low conversion is the most common trap: teams see the pretty completion number, assume the ad works, and pour budget into something that entertains but never sells. Low VCR with high conversion is the one teams kill by mistake, because the completion number looks ugly while the ad is quietly doing its job.

Rule of thumb: never optimize VCR without a conversion metric next to it. On its own, VCR rewards ads that are pleasant to finish, not ads that make people act.

Where they dropped off: reading the quartiles

Most ad platforms report completion in quartiles: how many viewers reached 25%, 50%, 75%, and 100%. The overall VCR is just the 100% number. The quartiles are where the diagnosis lives, because the shape of the drop-off tells you what's broken.

• Big drop before 25%: hook problem. Your first few seconds don't earn the next few. This is the most common and most fixable leak.
• Steady bleed from 25% to 75%: the middle is boring or too long. The premise held but the payoff took too long to arrive.
• Drop right at 75-100%: usually the CTA or end card. People got the message and left before, or right as, you asked for the click. Sometimes fine, sometimes a sign your ad is a few seconds too long.

Two ads can share the same 50% VCR and need opposite fixes. One lost half its viewers in the first three seconds (rebuild the open). The other held everyone until the last quarter (tighten the ending). The single VCR number hides that. The quartiles expose it.

How to actually improve completion rate

Once you've diagnosed the leak, the fixes are mechanical:

• Front-load the hook. If you're losing people before 25%, the fix is almost never the whole video. It's the first 3 seconds. Open on the payoff, the problem, or the pattern break, not a logo.
• Cut runtime to the story, not the slot. A tight 15 doesn't need to be 30. Every second past the point you've made your point is a chance to lose someone.
• Match the format's native feel. Sound-on, vertical, fast for TikTok and Shorts. Ads that look like ads get skipped; ads that look native get watched.
• Test more openings, not more budget. Completion problems are creative problems. Throwing spend at a weak hook just buys you more people leaving early.

That last point is where volume matters. If your VCR is low across the board, you don't have a media problem, you have a creative problem, and the answer is more shots on goal: more hooks, more cuts, more variations to test. That's the whole reason a marketplace like Viralix exists, matching brands with campaign-ready AI video creators who can turn out ad-grade variations fast, so you're testing ten hooks instead of defending one. If you're running any kind of creative testing program, completion-by-quartile is one of the cleanest early signals of which hook to scale.

The one-line version

Video completion rate is completed views divided by video starts, times 100. A "good" number depends entirely on placement: 90%+ is routine on non-skippable CTV and meaningless as a creative signal, while 20% can be normal on a muted feed. Read VCR against conversion to know whether to scale or fix, and read the quartiles to know what to fix. Never chase completion in isolation, because the highest completion rate in your account might be your worst-performing ad.

---

Related reading: [CTV Advertising: The Brand's Guide to Connected TV Ads](https://viralix.video/blog/ctv-advertising), [Video Hooks That Stop the Scroll](https://viralix.video/blog/video-hooks-stop-the-scroll), and [How to Build a Creative Testing Framework for Video Ads](https://viralix.video/blog/creative-testing-framework).