Reading your analytics

§1What the numbers are for

Analytics are not a report card. They are the only conversation you get to have with the people who saw your Short and decided what to do with it. Most of them never comment and never subscribe, but every one of them either watched or swiped, stayed or left, and YouTube Studio records those choices for you.

YouTube is unusually direct about what matters for Shorts in the feed. Shorts are ranked based on performance and viewer personalisation, and the signals it names are the percentage of viewers who chose to view, the average view duration and the average percentage viewed¹. The system also looks at whether a viewer chooses to watch, ignores the Short, or taps "not interested"¹. Those are the numbers worth learning to read. Almost everything else in Studio is context around them.

One technical point helps before you start. Since 31 March 2025 the public view count on a Short counts every start or replay, with no minimum watch time, and the older, stricter count is now called engaged views⁶. Partner Programme eligibility and Shorts revenue sharing still use engaged views⁶. So when the headline number looks large, remember that it is the generous count, and that the behaviour signals below are where the real story sits.

§2The order to read them in

Read the signals in the order a viewer experiences your Short. Each step only matters if the one before it went reasonably well, so reading them in sequence tells you where the problem actually is.

1. Viewed vs. swiped away. Studio reports how often your Short was shown in the Shorts feed, and the percentage of times viewers chose to view rather than swipe away². This is the verdict on your first moment: the opening frame, the first line, the promise. If people swipe, nothing after it had a chance.
2. The retention chart. This shows how much of the audience is still watching at each point. A flat line means people watched through that part, and a dip means people left at that moment³. Find the steepest drop and watch your Short at that exact second. Usually you will see why: a slow sentence, a repeated shot, a promise the video took too long to keep.
3. Average view duration and average percentage viewed. These summarise the chart in two numbers¹. Use them to compare Shorts with each other rather than to judge a single one.
4. Traffic sources. Where the views came from: the Shorts feed, search, your channel page, or elsewhere. A Short that lives almost entirely in the feed is being tested on strangers, while one found through search answers a question people were already asking.
5. Returning viewers and subscribers gained. These tell you whether the Short built anything lasting. A modest Short that brings people back and earns subscribers is often more valuable to a channel than a larger one that did neither.

Take a documentary Short about the night a city's great fire started. Suppose most people choose to view it, but the retention chart falls sharply at the point where the narration lists three dates in a row. The opening works and the topic works; the list of dates does not. That is one fix for the next script, not a reason to abandon the subject.

§3Compare against yourself, not the internet

Sooner or later you will read that a certain viewed percentage means a Short is going viral, or that retention below some number is a failure. YouTube does not publish any such figure. There is no official good percentage for viewed vs. swiped away, and no official threshold for retention. Numbers like that come from individual channels in particular niches, and they travel without their context.

The comparison that actually helps is your own baseline. Studio lets you look at your recent Shorts side by side, and that is where meaning appears. If your last ten Shorts held people for a certain share of their length and this one held them for noticeably less, something about this one is different, and you can go looking for it. If this one did better, that is even more useful: you have found something to repeat.

Compare like with like where you can. A Short published in a new subject, a new style or a new language is effectively a new experiment, and its first numbers say more about whether that audience exists than about your craft. YouTube itself notes that viewer preferences vary by format, and that perceived dips after a change usually come from how viewers react to the new format rather than from any penalty⁴.

A simple way to make the baseline concrete is to keep a small notebook or spreadsheet beside Studio. For each Short, write the subject, the angle, the first line, and two marks: whether the choice to watch was above or below your usual, and whether the retention chart had one obvious drop. After a month you will have a picture no benchmark can give you, because it is built from your own audience. Patterns appear quickly when they are written down: perhaps Shorts that open on a place outperform Shorts that open on a date, or Shorts about people hold better than Shorts about objects.

§4Give it time, and do not panic

The urge to check a Short every few minutes after upload is understandable and almost never useful. Early numbers are small and noisy. We suggest you leave a new Short alone for at least a couple of days before you draw conclusions from it, and that you read it again a week later alongside the Shorts published around it. That is our advice, not a rule YouTube has stated.

Two official statements should take the fear out of a weak video. First, the recommendation system uses fresh performance data for each individual video rather than relying on past results⁵. Your next Short starts with its own chance. Second, YouTube states that an individual video's underperformance does not penalise a channel overall⁴. A Short that falls flat is information, not damage.

YouTube also says that breaks are not penalised and that publish time is not known to affect a video's long-term performance⁵. So you do not need to chase a perfect hour or post in a panic after a quiet week. Keep publishing good work at a rhythm you can sustain, and let the numbers accumulate into patterns.

If several Shorts in a row fall below your baseline, that is worth a closer look, but look at the shared cause rather than at the channel. Did they all open the same way, cover a subject your audience never asked for, or run longer than they needed to? The answer is almost always in the videos themselves.

§5A twenty-minute weekly review

The point of reading analytics is to decide what to make next. The routine below is our suggestion for doing that once a week, in about twenty minutes, without drowning in numbers.

1. List the week's Shorts

   Open YouTube Studio and line up the Shorts you published in the last seven days, plus the ten or so before them as your baseline.

2. Mark the choice to watch

   For each new Short, note whether viewed vs. swiped away was above or below your own recent typical value². Write one word for why: the opening, the topic, or the cover moment.

3. Find one drop

   Open the retention chart of the weakest and the strongest Short. Find the steepest fall in each³, watch that second, and write down what happened on screen.

4. Check where viewers came from

   Look at traffic sources and at subscribers gained. Note any Short that brought in subscribers or search traffic out of proportion to its views.

5. Decide three things

   End with three decisions: one angle to repeat, one habit to drop, and the topics for next week. Write them down. A review that ends without a decision was only reading.

This is where the reading meets ViewMade. Take the angles that worked to Ideas, steer a batch of ideas with that theme, and pick next week's subjects from the cards, favouring the ones that follow the same thread. Then go to Studio and run the winning angle again on a fresh subject: if a Short about a forgotten engineering disaster held people well, try another disaster with the same structure of mystery, cause and consequence. Everything you made stays in Library, so you can open last month's best Short, compare it with this week's, and see what changed.