Start with the goal, not the dashboard
Every platform offers a long list of numbers, and most of them are easy to read and hard to interpret. A useful way to avoid confusion is to write the goal first and pick the measure second. A video meant to make people aware of a brand should be judged differently from one meant to bring in sign-ups.
Broadly, marketing video serves three kinds of goals:
- Awareness: getting a message in front of the right people.
- Consideration: helping people understand an offer and form an opinion.
- Action: prompting a purchase, enquiry, donation or sign-up.
Each goal points to different marketing metrics. Using an awareness number to claim an action result, or the reverse, is a common source of misleading reports.
Numbers that describe delivery
These show that a video was shown and, to some extent, watched.
- Impressions and reach. Impressions count how many times the video was served; reach counts distinct people. Neither says whether anyone paid attention.
- Views. What counts as a view varies by platform, with some counting a very short playback. Two platforms’ view counts are rarely directly comparable.
- Watch time and average percentage viewed. These say more than a raw view count because they show how much was watched. The shape of drop-off over the video is shown in retention graphs, which help locate where interest fades.
- Completion rate. Useful for short clips, but it favours shorter videos, so comparing a 15-second clip with a five-minute film on this measure is not meaningful.
The platform’s own reporting is covered in video analytics. For awareness goals, delivery numbers are often a reasonable starting point, but they are weak evidence of impact on their own.
Numbers that describe engagement
Likes, comments, shares and saves signal that something resonated. They can be useful for comparing one clip against another on the same channel, but they have limits.
- Engagement is often easy to prompt with giveaways or provocative content, which may not relate to the buying decision.
- Rates depend on the platform and audience, so there is no universal good figure. Compare against your own earlier results rather than a number someone quotes.
- A comment can be positive or negative; counts alone do not show which.
Reading these as signals of interest, not as proof of business results, keeps conclusions honest.
Numbers that describe outcomes
Outcome numbers connect video to what the organisation actually wants.
- Click-through rate shows how many people who saw an ad or link chose to click. It says little about what happened next.
- Conversion rate shows how many visitors completed the desired action. It depends on the page and offer as much as on the video.
- Cost per result, such as cost per sign-up, divides spending by outcomes. It is only meaningful when the outcomes are counted consistently and costs include production, not just media spend.
- Return on spend compares revenue to cost. It is attractive but difficult to compute reliably for video, since many sales have several influences.
Why attribution is hard
Customers rarely act after a single exposure. Someone might see a clip on a phone, hear about the brand from a friend, and buy a week later on a laptop. Attribution is the practice of deciding which touchpoint gets credit, and every method simplifies reality.
- Last-click credits the final step, which tends to favour search and direct visits and undervalue video that created the interest.
- View-through credit counts people who saw an ad and later converted, even without clicking. It can overstate the effect because some of those people would have converted anyway.
- Multi-touch models share credit across steps, but the weights come from assumptions.
None of these is simply right. If the decision matters, a more credible approach is a controlled comparison, covered next.
Comparing fairly
The most reliable question is what would have happened without the video. A few ways to approximate that:
- Hold-out groups. Keep a similar group of people from seeing the video and compare their behaviour with those who did.
- Before-and-after comparisons. Compare results before and after launch, while recognising that seasons, promotions and news can also shift the numbers.
- A/B tests. Show two versions to similar audiences at the same time and compare. Change one element at a time, and wait until there is enough data that the difference is larger than normal day-to-day variation.
Small samples can produce impressive-looking differences by chance, so reports should say how many people were involved, and avoid declaring a winner on a handful of results.
Common ways numbers mislead
- Celebrating views that came from irrelevant audiences. A cheap view from someone who will never buy adds little.
- Comparing across platforms that define metrics differently.
- Ignoring costs. A high-performing video that was very costly to produce may be a poor investment compared with a modest one.
- Averaging away detail. An average watch time can hide that most people left in seconds while a few watched to the end.
- Judging too soon. Some videos keep earning attention over months, while others spike and fade, so the reporting window should match the goal.
A simple reporting habit
A one-page report is usually more useful than a screen of charts. For each video, record the goal, the audience, the cost, one or two primary measures that match the goal, and a sentence on what will be done differently as a result. Over time, comparing like with like (the same format, channel and goal) builds an honest picture of what works for your audience, which is worth more than any benchmark from outside.