Engagement Heat Map

TikTok video retention: find where viewers drop off before you post

TikTok's own analytics show a retention graph after a video has already spent its first test audience. Go Viral's engagement heat map estimates the curve from the clip beforehand, so the dip gets fixed in the edit instead of in the next video.

Find my drop-off

4.7 out of 5 from 321 ratings on the App Store

Last updated 2026-09-02

What a retention curve shows

A retention curve is a line that shows, for each second of a video, what share of the people who started it are still watching. It always slopes down, because nobody can join late and anyone can leave. What matters is the shape. Almost every short video shows the same four features: a cliff in the first three seconds where viewers decide, a gap a few seconds later where the promise of the hook is either kept or not, a sag in the middle where the pacing slackens, and a drop at the end that a clean loop can turn into a rise. The engagement heat map in Go Viral paints those four features onto the timeline of the clip before it is posted, and marks the moments viewers are likely to replay.

The four places viewers leave, and the edit for each

Retention drop-off points and fixes
Where	What the curve shows	Why it happens	The edit
Seconds 0 to 3	A steep cliff	No reason to stay was given before the viewer decided	Recut the opening. Score it with the hook score until the first word lands on the first frame.
Seconds 3 to 8	A second step down	The hook promised something and the next five seconds did not start delivering it	Cut the setup. The context can come after the first piece of the payoff, not before it.
The middle	A slow, steady slide	One shot held too long, a sentence that restates the last one, a list that runs past four	Find the second where the slope steepens and cut two seconds out on either side of it.
The last second	A sharp fall, or a bump	An outro, a "follow for more", or a fade; or, with a loop, a rewatch	End on the frame that leads back to the first one. Put the call to action in the caption.

Retention, completion and watch time are three different numbers

Retention is the curve: the share still watching at each second. Completion rate is one point on it, the share who reached the end. Average watch time is the area under it, expressed in seconds or as a share of the length, and it is the one that rewatches inflate. A 10-second video with a 60 percent completion rate and heavy rewatching can beat a 30-second video with 80 percent completion on average watch time, and the platform weighs watch time heavily.

This is why cutting a video shorter is not automatically a fix. A shorter video with the same sag has the same shape. A video of the same length with the sag removed has a better one.

How to read the heat map in the app

Warm colours on the timeline mark the seconds where viewers are most likely to leave.
Replay markers show the moments most likely to be rewatched. Those are the best seconds in the clip.
If the strongest replay moment is at second eighteen, ask whether it could be at second two. Often the best hook is already in the video, just late.
Read the heat map next to the view forecast: a low forecast with a clean opening means the middle is the problem, and the heat map says where.

Retention questions

What is a good average watch time on TikTok?

As a rule of thumb: for a clip under 15 seconds, most viewers should finish it and rewatches can push average watch time past 100 percent of the length. For 15 to 30 seconds, an average above half the length is healthy. For 30 to 60 seconds, above 40 percent. Over a minute, above a third. TikTok publishes no thresholds; these are the ranges creators and analytics vendors report as the point where distribution keeps widening.

Why do viewers drop off at 3 seconds?

Because that is when they decide. The first three seconds are the only part of the video every viewer sees, and the decision to keep watching is made on the first frame, the first word and whether a promise has been made. A drop at second three is a hook problem, and the hook score is built to catch it before posting.

Do rewatches count toward retention?

Yes. A viewer who watches a 12-second clip twice contributes 24 seconds of watch time, which is why a clean loop can lift average watch time above the clip length. The heat map marks the moments that get replayed so you can build the loop around them.

Can I see retention before I post?

Not from TikTok; its retention graph appears after the video has already been through its first test audience. The engagement heat map in Go Viral estimates the curve from the clip before it is posted, so a dip is fixed in this edit rather than diagnosed in the next one.

Fix the dip in this edit, not the next one

Import the clip, find the second where the curve bends, cut around it, and score the new cut. The heat map is the part of the app that changes what you do with the scissors.

Get the heat map