Odinstat
Methodologyvs YouTube StudioPricing
  1. Home
  2. Analytics
  3. YouTube audience retention analysis

Audience retention

How effective are my video hooks and retention?

Odinstat reads the real 101-point audience retention curve YouTube records for each video, scores the hook, finds the single largest drop-off, and classifies the shape of the curve.

A second analysis ranks long-form videos by how well they hold viewers compared with what is typical for their length on your channel, so a 25-minute video is not punished for being long.

Two questions, two analyses

Retention curves answer “where do people leave?”. Retention for length answers “which videos hold attention best once you account for how long they are?”. They use different data and different gates, and Odinstat keeps them separate rather than blending them into one number.

Retention curves: hook, cliff, profile

YouTube exposes each video’s audience retention as 101 points, one per percent of the video. In the default run over all your content, the hook score is the share of viewers still watching at the 5% point of the video. When you filter the analysis to Shorts only, the hook is read at the 10% point instead, because a Short’s opening is proportionally longer.

The largest single-point drop between consecutive points is the cliff: the exact moment the most viewers left at once. Each curve is then classified as, for example, steady, front-loaded, or a cliff, and the profile card and the hook verdict are computed from the same classification, so the page never praises your openings while a second section says most of your videos collapse.

This analysis includes Shorts.

Retention for length: a fair ranking

Raw average view percentage always favours short videos. Sixty-five percent of a three-minute video is two minutes of attention; forty-five percent of a twenty-minute video is nine. Odinstat fits a log-linear line of expected watch percentage against duration using your own videos, then scores each video by its residual: how far above or below the expectation for a video of that exact length it landed.

A positive residual means “held viewers better than a video this long usually does on this channel”. This analysis is long-form only; Shorts are excluded because their duration range is too narrow for the fit to mean anything.

What you need

Retention curves need at least 5 videos with retention data and include Shorts. Retention for length needs at least 8 long-form videos and excludes Shorts.

What this is not

  • Telling you what to say in the first fifteen seconds. Odinstat shows where viewers leave; the fix is yours.
  • Industry retention benchmarks. “Typical” always means typical for your channel.

Odinstat runs this on your own channel, read-only, with a 14-day free trial from $10/month.

See pricing

Related

  • YouTube outlier videos, by the numbers
  • Thumbnail CTR, adjusted for reach
  • YouTube traffic sources, explained
  • Video shelf life and half-life on YouTube
  • Shorts vs long-form analytics for your channel
  • Odinstat vs YouTube Studio
  • Odinstat vs vidIQ and TubeBuddy
  • Methodology: how every number is computed
AnalysesYouTube outlier videos, by the numbersYouTube audience retention analysisThumbnail CTR, adjusted for reachYouTube traffic sources, explainedVideo shelf life and half-life on YouTubeShorts vs long-form analytics for your channel
CompareOdinstat vs YouTube StudioOdinstat vs vidIQ and TubeBuddy
CompanyMethodologyPrivacy PolicyTerms of ServiceSecurity
← Back to Odinstat© 2026 Odinstat LLC. Not affiliated with YouTube or Google. YouTube is a trademark of Google LLC.