Your average retention number doesn't show where the drop happens — the curve does.
Definition
A watch-time bottleneck is the point on your retention curve where viewer drop-off is sharpest — not just where the number is low, but where it falls fastest. This is what AlgoLens looks for to show you where in the video viewers stop watching the most.
Drop-off analysis finds and shows the section of your retention curve with the biggest viewer drop, expressed as a percentage-point decline at a given point in the video — for example, "a 12-point drop around the 40% mark."
Segment retention is the share of viewers still watching at a specific point in the video, shown relative to your own channel's baseline (1.0 = your average). A segment reading below 1.0 means that part is losing viewers faster than the rest of your videos typically do.
Why the average number hides where viewers leave
A single retention percentage — say, 42% average retention — tells you how much of the video people watched on average. It doesn't tell you whether viewers left gradually and evenly, or whether most of them dropped off in one sharp moment. Two videos can have the same average retention and completely different bottlenecks — one loses viewers evenly throughout, the other loses half its audience in a single 10-second stretch. Only the curve itself shows you which one you have.
What to do
Where AlgoLens helps
AlgoLens's Video Analysis tab locates the sharpest drop-off point on each video's retention curve automatically and shows the percentage-point decline at that spot, so you don't have to eyeball the whole curve manually.
Related terms
FAQ