Your Video Analytics Are Lying: Track Completion Rate Instead

Your latest video has 10,000 views.

The marketing team is celebrating. The campaign report looks impressive. Someone suggests increasing the budget for the next video.

But there's one problem.

You don't know how many people actually watched it.

Some viewers may have left after three seconds. Others may have watched halfway through. A smaller group might have stayed until the end and then visited your website.

Your analytics dashboard puts all these people under one attractive number: views.

That's why views alone can give you a misleading picture of video performance.

If you want to understand whether your video is holding attention, start looking at completion rate.

Why Video Views Don't Tell the Whole Story

Views are useful. They help you understand reach and, depending on the platform, how often content meets its view-counting criteria.

But a view doesn't automatically mean someone understood your message, remembered your brand, or watched your video all the way through.

Imagine two videos:

Metric	Video A	Video B
Views	10,000	6,000
Viewers who completed the video	1,000	2,400
Completion rate	10%	40%

Video A attracted more views.

Video B had fewer views but a much higher completion rate.

If your goal is to hold attention through the entire message, Video B is the stronger performer on that metric.

However, completion rate doesn't automatically make Video B the better business investment. Video A might generate more qualified leads or sales.

The lesson isn't to stop tracking views.

It's to stop treating views as the entire story.

What Is Video Completion Rate?

Video completion rate measures the proportion of video plays or viewers that reach the end of a video, depending on how the platform defines and reports the metric.

A simplified formula is:

Completion rate = Completed plays ÷ Total eligible plays × 100

Suppose your video receives 2,000 eligible plays, and 600 of those plays reach the end.

Your completion rate is:

600 ÷ 2,000 × 100 = 30%

That means 30% of the eligible plays resulted in a completed view under this definition.

One important detail: platforms don't all count views and completions in exactly the same way. Some use unique viewers, some use video plays, and others use platform-specific engagement definitions.

When comparing results, make sure you're using the same denominator and completion definition.

Otherwise, you might be comparing two numbers that look similar but measure different things.

The Metric I Would Track Alongside Completion Rate

Completion rate is valuable, but it doesn't tell you everything about viewer behaviour.

I would start with three metrics.

1. Completion rate

This tells you what proportion of eligible plays reached the end.

It's particularly useful for evaluating whether your video maintains attention throughout its duration.

2. Average percentage viewed

This measures the average proportion of a video's duration that viewers watch.

For example, if a 100-second video has an average watch time of 40 seconds, the average percentage viewed is 40%.

This metric can reveal whether viewers generally leave early or watch a substantial portion of the content.

Be careful when comparing platforms because replay behaviour and calculation methods may differ.

3. Retention by timestamp

This is where things become more interesting.

A retention graph shows how the audience changes as the video progresses.

Imagine a five-minute product demonstration.

• At 0:15, 90% of viewers remain.
• At 1:00, 75% remain.
• At 2:30, 42% remain.
• At 4:00, 35% remain.
• At 5:00, 20% complete the video.

These are illustrative figures, not industry benchmarks.

The sharp decline around 2:30 gives you something specific to investigate.

Perhaps the demonstration becomes repetitive. Maybe an important explanation takes too long. Perhaps viewers have already found the information they need.

The graph identifies where attention declines. It doesn't tell you why by itself.

That requires reviewing the content and considering the audience's intent.

How to Calculate Completion Rate Yourself

If you're collecting video events in your own application, you can calculate a basic completion rate from playback events.

Here's a simple JavaScript example for an HTML video player:

<video id="product-demo" controls>
  <source src="/videos/product-demo.mp4" type="video/mp4">
</video>

<script>
  const video = document.querySelector("#product-demo");

  let started = false;
  let completed = false;

  video.addEventListener("play", () => {
    if (!started) {
      started = true;

      console.log("video_started");
    }
  });

  video.addEventListener("ended", () => {
    if (!completed) {
      completed = true;

      console.log("video_completed");
    }
  });
</script>

This example detects when playback starts and when the video reaches its end.

It is a useful starting point, but it is not a complete analytics system.

For example, refreshing the page, replaying the video, opening another tab, or playing the video repeatedly can affect how events should be counted.

A production implementation should define what constitutes a play, how repeated plays are handled, and whether the reporting unit is a unique viewer, a session, or an individual playback.

It should also send events to an analytics backend rather than relying on console messages.

Sending Events to an Analytics Endpoint

You could send playback events to your backend using fetch():

async function trackVideoEvent(eventName, videoId) {
  try {
    await fetch("/api/video-events", {
      method: "POST",
      headers: {
        "Content-Type": "application/json"
      },
      body: JSON.stringify({
        event: eventName,
        videoId,
        timestamp: new Date().toISOString()
      }),
      keepalive: true
    });
  } catch (error) {
    console.error("Unable to send video event", error);
  }
}

You could then replace the console.log() statements with calls to trackVideoEvent().

Your backend would store the events, validate incoming data, apply deduplication rules, and calculate completion rates for reporting.

In a real deployment, consider authentication, rate limiting, privacy requirements, and the reliability of events sent when a user closes the page.

Also remember that the browser's ended event indicates that playback reached the end. It doesn't prove that a viewer paid attention or understood the content.

Measurement tells you what happened during playback, not what happened inside someone's head.

Why Your Video Might Have a Low Completion Rate

A low completion rate isn't automatically a sign of poor production.

Several factors can influence it.

The opening takes too long

If viewers cannot quickly understand what a video offers, they may leave before the main point arrives.

Try communicating the purpose early.

Instead of beginning with a lengthy company introduction, consider opening with the problem the video will address.

The video is longer than necessary

Some subjects require ten minutes. Others need thirty seconds.

The right length depends on the audience, the subject, the platform, and the viewer's intention.

Don't shorten every video simply to improve the percentage. A detailed training video may be successful even if not every learner watches it in one sitting.

The content doesn't match the promise

If the title promises a practical demonstration but the opening delivers several minutes of unrelated background, viewers may decide the video isn't what they expected.

Keep the title, thumbnail, opening, and actual content aligned.

The pacing loses momentum

Repeated explanations, unnecessary transitions, and long pauses can cause viewers to leave.

Review retention graphs alongside the actual video to identify sections that might benefit from editing.

The audience has already found the answer

A short instructional video may answer a viewer's question quickly, making an early exit perfectly reasonable.

This is why completion rate must be interpreted in context.

A viewer who gets the exact answer they need in 20 seconds isn't necessarily a failed engagement.

Don't Optimise Completion Rate at the Expense of Business Results

This is where marketers can make another mistake.

They discover completion rate and start treating it as the ultimate measure of success.

But consider two scenarios.

Video A: A short, entertaining video achieves an 80% completion rate but generates no qualified enquiries.

Video B: A detailed product demonstration achieves a 35% completion rate but generates several relevant sales conversations.

Which one performed better?

The answer depends on the objective.

If the goal is awareness, you might prioritise reach, retention, and brand recall.

If the goal is lead generation, you might focus on qualified visits, enquiries, and conversions.

If the goal is education, you might measure completion alongside knowledge retention or task performance.

Completion rate helps explain attention. It doesn't replace the metrics that measure your actual objective.

A Practical Video Analytics Dashboard

If I were setting up a basic dashboard for a marketing team, I would include these metrics:

Metric	What it helps you understand
Views	How much measured exposure the video receives
Average watch time	How long viewers watch on average
Completion rate	How often plays reach the end
Retention curve	Where viewers stop watching
Click-through rate	How often viewers click a relevant link or CTA
Conversion rate	How often the desired action occurs
Cost per result	How efficiently a campaign achieves its objective

You don't need every metric for every video.

Start with the objective, choose a small number of relevant measurements, and establish a baseline.

Then compare videos with similar purposes, audiences, lengths, and distribution conditions.

That makes your conclusions more useful than comparing unrelated videos based on one headline number.

What to Do With the Data

Collecting analytics is the easy part.

Using them to improve the next video is where the value appears.

Try this process:

1. Establish a baseline. Record completion rate, average watch time, and your primary business metric.
2. Inspect the retention curve. Identify major drops and sections where viewers remain engaged.
3. Review the footage. Look for slow openings, repetition, unclear explanations, or mismatches between the promise and the content.
4. Form one hypothesis. For example, "Moving the product demonstration earlier may improve early retention."
5. Test the change. Compare similar audiences and distribution conditions where possible.
6. Evaluate the outcome. Check whether attention improved without harming the result that matters to the business.

Avoid changing the opening, length, thumbnail, distribution, and CTA simultaneously if you want to understand which change made a difference.

And don't treat a small difference as proof of improvement without considering sample size and normal variation.

The purpose is to learn, not simply to produce a more attractive dashboard.

Final Thoughts

Video views aren't useless.

They're just incomplete.

A video with thousands of views can still struggle to hold attention. A video with fewer views can create meaningful engagement or business results.

That's why your reporting should go beyond the headline number.

Track completion rate. Study average watch time. Examine retention. Connect those measurements to the reason the video exists in the first place.

If you're planning a video campaign, it's worth thinking about measurement before production begins. The intended audience, message, distribution channel, and desired outcome should influence both the creative approach and the analytics you collect.

You can explore examples of professional video projects through Title Productions' case studies and find more practical perspectives in Title Productions' Insights.

The goal isn't to make everyone watch until the final second. It's to understand whether your video is doing the job you created it to do.

What do you prioritise when evaluating a video: views, watch time, completion rate, or conversions? And has a retention graph ever revealed a problem you wouldn't have noticed otherwise?