A video has many views, but no inquiries. Another was watched less, yet after it people ask about the product. To choose the topic of your next Reels, first decide which of these results you need right now. The overall counter cannot tell you what happened after the view.
Separate the three questions
How many people did the video reach? For this, look at account reach if the metric is available. It helps you avoid confusing the size of the audience with repeat plays.
How long did they watch? Total time depends both on the number of views and on attention to the video. Average time helps evaluate the viewing of an individual video, but comparing a short and a long tutorial still requires caution. Meta separately explains that total Reels time includes repeat plays. Description of Meta metrics.
What did they do afterward? Saves, substantive comments, profile visits, and inquiries are needed for different tasks. Choose the action in advance, otherwise it is easy to declare success the metric that happened to grow.
Report names and availability may vary. For comparison, record the exact metric from your interface, the video length, and the date the data was captured.
Situation 1: many views, few product-specific questions
Suppose a lighting store filmed a beautiful light change in a room. The video attracts attention, but it is unclear from it what is being sold: a lamp, a design project, or an installation service. Questions like "what's the music?" are pleasant for the author, but they do not help someone choose a light fixture.
The next video can be devoted to one item: show it turned off, turned on, and next to a person for scale. In the caption, include the model name and how to ask about the specifications. Check whether there are more questions about the product. Views may not increase at the same time: you are testing a different task.
Situation 2: the topic is useful, the explanation drags
In the training example, two videos of the same length talk about setting up lighting. One begins with the story of buying the equipment, the second begins right away with a failed shot and its correction. The second has a higher average watch time.
This is a reason to test the opening of the next video. There is no need to declare it proven that any stories hinder viewing: the audience, the traffic source, or the topic itself could have differed. Keep the format and try starting the next similar tutorial with a demonstration.
If a retention graph is available for the video, match noticeable drops to specific scenes. A drop next to a long intro gives a more precise hypothesis than the advice to "shoot more dynamically."
Situation 3: fewer views, more inquiries
A video like "How to choose light for a desk by a window" may be more useful to a buyer with a specific task than a flashy general roundup. If suitable inquiries came in, expand the scenario: a desk without a window, a narrow tabletop, video calls in the evening.
Write down the content of the questions without personal data. They will suggest which example to add to the next post. Do not carry over the number of inquiries from one successful video to your entire future plan.
Keep a short decision log
For each post, it is enough to note the topic, the first frame, the duration, the available metrics, and one note: "In the next video we'll show the dimensions earlier." Capture the data at the same time after publication and mark ads, collaborations, external placements, and ordered activities separately.
If you use a Reels views service, do not mix its volume with an assessment of organic interest. Check purchases and inquiries using your own data, not the size of the counter.
Move comparable actions into the calculator
For two videos with the same purpose, record the duration, the same snapshot age, the exact reach, and the same types of actions. Training example: a short video received 50 selected interactions with a reach of 1 000 accounts, a long one — 90 with a reach of 1 500. The ratios are 5% and 6%. These are not completion percentages and not the share of buyers.
In the ER calculator ](/ru/instrumenty/er/), select the reach and one numerator composition for both videos. Save the CSV along with the duration and measurement time in your log. If reach is unavailable, do not substitute it with views: choose a different denominator and label it explicitly. Comparing two observations does not prove an advantage of the long format; look at retention and the desired action separately.
✦ Choose a quantity and place an order →
After several issues, the log will show recurring questions and successful ways of explaining. These are the ones worth carrying over into the next video; a random view record by itself does not replace a script.