Social media · Lesson 2.2
Measuring Hook Performance: Retention, Swipe Away, Rewatch
Most people check views and likes, then guess why a post did well or badly. Those numbers come late and mix many causes together. If you want to know whether your hook worked, look at the numbers that sit closest to it.
Three are worth a weekly check: early retention, swipe away rate and rewatch. This guide explains each, where to find it, and what to change when it is weak. Names and layouts differ by platform and change over time. For example, YouTube's Shorts analytics describe a report of the percentage of times viewers viewed your Short versus swiped away, and a separate key moments for audience retention report. Other platforms use other names, so look for the closest equivalent in your own analytics.
Views tell you what happened. Early retention, swipe away rate and rewatch tell you why. Check them every Monday and change one thing per week.
The three numbers
| Number | What it shows | Where to look | Weak result points to |
|---|---|---|---|
| Early retention | Share of viewers still watching after the first few seconds | Retention graph in the video analytics | The opener |
| Swipe away rate | Share who left almost at once | Often shown as "stayed to watch" or drop off in the first seconds | The first frame and first line |
| Rewatch | How often people watch again or loop | Average watch time compared with video length | Strong content, or a satisfying ending |
Early retention
On the retention graph, look at the first few seconds. A steep fall at the start means the hook failed to hold people. A gentle slope means it worked and the body is now doing the job. Compare each post with your own recent posts rather than with a figure from the internet. Your normal is the only fair baseline, because topics, audiences and platforms vary.
Swipe away rate
This is the share of viewers who left right away. "Swipe away rate" is a working name used in this post, not an official metric name. On YouTube Shorts the nearest report is viewed versus swiped away, also described as stayed to watch. Where it is not shown, the first point on the retention graph or a viewed versus swiped away figure serves the same purpose. A high swipe away rate suggests the first frame, first words or text line did not give a reason to stay.
Rewatch
When viewers watch more than once, average watch time can exceed the length of the video on some platforms. Check your platform's own definition of the metric. A high figure can mean a short, loopable post, a satisfying ending or content dense enough to need a second look. Rewatch is a good sign, though it does not by itself tell you the hook was strong. Read it with the other two.
A Monday routine
- List last week's postsPut them in a simple table with date, hook style and topic.
- Record the three numbersEarly retention, swipe away and rewatch, copied from your analytics.
- Spot the best and worstCompare the hook of each. What did the strong ones share?
- Pick one changeChange one thing next week, such as a faster opener or a new style, so you can tell what caused the shift.
- Save the winnerAdd the best hook to your swipe file with a note on why it worked.
A spreadsheet is enough. If you want a faster way to gather the figures, the data lesson 4.4 and the agents lesson show how to pull numbers into one view with AI help.
Read the numbers together
- Weak early retention, strong later retention: the hook is the problem, the body is fine. Rewrite the opener.
- Strong early retention, steep fall in the middle: the hook works but the body drifts or does not match. See From Hook to Post.
- Good retention, low views: the post held people but was not shown widely. Look at topic, timing and how it was packaged.
- High rewatch, low follows: the post entertains but does not give a reason to follow. Add a clear next step.
Avoid over reading small samples
A single post with a few hundred viewers can be noise. Wait for several posts before drawing a conclusion, and treat one big hit or miss as a lead to test, not a rule.
Test openers properly
Change only the hook between two versions and keep the topic and length similar. Post them on different days at similar times and compare the three numbers. This is the testing step in Lesson 2.2: write ten openers, choose two, keep the one people finish. To find openers worth testing, see 15 Hook Formulas or the list of common faults in Hook Mistakes That Kill Reach.
A worked example of a Monday review
Say you posted five videos last week. This is an illustration with made up figures, only to show the process.
You list them. Video A and D have the steepest early drop. Video C holds well all the way. You look at the openers: A and D began with greetings, C began with the finished result on screen. The pattern points at the opener, so next week you cut the greeting from every video and compare again. You write down: "Start on the result."
You did not need a perfect number to learn that. You needed a comparison between your own posts.
A simple tracking sheet
Make a sheet with columns for date, topic, hook style, early retention, swipe away, rewatch and one line of notes. Fill it once a week. After a month, sort by early retention and read the top five and bottom five hooks. The differences are your next experiments.
What to do when numbers look fine but the business does not move
Hooks get attention, but attention is not income. If your posts hold viewers and nothing follows, look at what comes after the hook: a clear next step, a welcome message and a way to follow up. See the welcome email lesson for one way to catch the people who do respond.
Questions people ask
What is a good retention number?
There is no universal figure. Compare each post with your own recent average and aim to beat it slowly.
My platform does not show swipe away rate. What do I use?
Use the first point of the retention graph or the share of viewers who stayed past the opening seconds, if shown.
How often should I check these?
Once a week is enough. Daily checks tempt you to react to noise.
Do likes still matter?
They show approval, but they come after the hook has already done its job. Use them as a secondary signal.
Where to go next
This Monday, record the three numbers for your last five posts and pick one change. For the full process of writing and testing hooks, see Lesson 2.2.