Video marketing metrics: what to track, and the benchmarks that exist

The video marketing metrics that matter fall into four groups: whether people saw it (impressions and click through rate), whether they started it (views and view rate), how much of it they watched (watch time, average view duration and how far through they got), and what they did next (clicks and results). Track one number from each group per video, read each against the definition the platform publishes, and compare a video with your own previous videos rather than with an industry average.

The definitions matter more than they look. A YouTube skippable ad counts a view at thirty seconds, or the whole video if it is shorter, or a click. A LinkedIn video ad counts a view at two continuous seconds with half the video on screen. Put those two numbers in the same column of a report and the report is meaningless before anyone reads it.

Views are not counted the same way anywhere

Start every report by writing down which kind of view you are reporting. Google Ads defines a view on a skippable in stream ad as a viewer watching 30 seconds or to the end of the video, whichever comes first, or clicking an element of the ad such as the call to action, a card or the logo. LinkedIn’s Campaign Manager counts two or more continuous seconds of playback while the video is at least 50 percent on screen.

For a 29 second video, those rules produce very different pictures. On YouTube an ad view means the person watched the entire thing or clicked. On LinkedIn the same person could have watched two seconds on the way past. Both platforms then report view rate as views divided by impressions, so the view rate inherits the difference. A LinkedIn view rate is likely to look higher, and that says nothing about which platform worked better.

Watch time and average view duration

YouTube Studio defines watch time as the amount of time viewers have watched your video, and average view duration as the average minutes watched among those who stayed to watch, calculated from engaged views.

Average view duration is only meaningful as a share of the video’s length. Forty seconds is excellent for a one minute video and poor for a ten minute one. Divide it by the length before you compare two videos. Watch time, meanwhile, rewards longer videos and bigger audiences, which is useful for a channel and misleading for a single short ad.

Wistia’s 2026 State of Video report, drawn from 13 million videos and 79 million hours of viewing on its platform, found that the shorter the video, the higher the engagement rate, while longer videos build more total watch time. Both statements are true at once, which is why a report needs both numbers. For how that plays out by placement, see how long a promo video should be.

Completion and quartiles

Quartile metrics show where people leave. Google Ads reports the share of viewers who initiated the player and watched to 25, 50, 75 and 100 percent. LinkedIn reports views at 25, 50 and 75 percent and a completion when a video is watched to 97 to 100 percent of its length, with completion rate calculated as completions over impressions.

This is the most practical diagnostic for a short ad. The headphones UGC example is 29 seconds and six scenes, and its offer, the code for ten percent off, is the final scene. Anyone who leaves at 75 percent never hears it. If the 25 percent figure is weak, the opening line is the problem. If viewers hold to 75 percent and then drop, the ending is too slow, or the offer arrives too late.

Click through rate: two different numbers

“Click through rate” means two unrelated things depending on where you read it. On YouTube, impressions click through rate measures how often viewers watched a video after seeing its thumbnail, and YouTube’s own help page says half of all channels and videos sit between 2 and 10 percent. It measures the thumbnail and the title, not the video.

In an ad platform, click through rate means clicks on the ad’s link, the action you actually paid for. A video can have a strong thumbnail rate and a weak link rate, or the reverse, and each points at a different fix. Label which one you mean every time.

Where each number lives

The table at the end of this article lists each metric, the definition its platform publishes, and where to read it. The short version is that organic YouTube numbers live in YouTube Studio, paid numbers live in the campaign report of the platform that ran the ad, and results live in your website analytics, ideally with a separate tagged link for each video so a sign up can be traced to the cut that caused it.

The fitness app example and the landscape SaaS feature launch example show why that tagging matters. One is a vertical creator ad built for a feed, the other a landscape launch video for YouTube or a website. If both point at the same page with the same link, their results merge and neither can be judged.

A one page report per video

Report each video on one page with the same five lines every time. The placement and the view definition. Impressions and view rate. Average view duration as a share of length, or the quartiles for an ad. The click rate on the link. The result, whether that is sign ups, bookings or sales. Put the previous video’s numbers in the next column.

Then write one sentence underneath: what the numbers suggest changing in the next video. A weak 25 percent figure means a new opening line. Healthy quartiles with a weak click rate mean a clearer offer or a different landing page. Strong clicks with poor results mean the page after the click, not the video, is where the problem sits. Change one of those things at a time, so the next report can show whether it worked.

That comparison is the benchmark that counts. Published averages mix industries, lengths and definitions. Your last five videos share your audience, your offer and your definitions, so a change against them means something. If the format itself is new to you, what a UGC ad is explains why creator style videos are usually judged on their first seconds and their close.

Common mistakes

Adding views across platforms. A total of YouTube ad views and LinkedIn views is not a number anything can be decided on.

Reading watch time for a short ad. It favours long videos. Use average view duration as a share of length, or quartiles.

Judging a video on likes. Engagement is a signal, not the result the video was made for.

One link for every video. Without a tag per video, results cannot be traced to the cut that earned them.

Comparing with someone else’s benchmark. Different lengths, audiences and view rules make other people’s averages a weak yardstick.

Build it

The video builder itself reports no views. A share link opens a read only player page that plays the video live in the browser; it is not an MP4 file and it does not count or list who watched, which makes it the right thing for review, as the share video link page explains, and the wrong thing for measuring an audience. To measure, export the MP4 and upload it to the platform where the audience is, which is where every metric above is recorded.

Videos are offered in landscape at 1920 by 1080 and vertical at 1080 by 1920, so export the shape each placement needs; the export guide covers the settings. An MP4 already rendered from the same content is reused with no second charge, 4K export needs the Starter plan, and captions are on by default, which is what a muted autoplay view needs to carry the message.