Hold Rate Calculator
A great hook gets the thumb to stop. Hold rate tells you whether the rest of the video earns the attention the hook borrowed. Enter how many viewers reached your chosen point — halfway, the end, wherever the story turns — against the total who started.
Video hold rate = viewers who watched to a chosen point ÷ total viewers × 100%. Where hook rate grades the first three seconds, hold rate grades everything after: whether the creative keeps the attention it captured. It is the paid-social companion to hook rate, and the two together diagnose a video far better than view count alone — a strong hook with a weak hold means the opening overpromised, while a weak hook with a strong hold means the few who stayed were the right audience.
Hold Rate Calculator inputs and result
| Hold rate band | What it suggests |
|---|
How to use this calculator
- Choose one retention milestoneDecide whether you are measuring hold to 50%, 75%, or the full video. Hold rate is only comparable when the milestone is fixed, so pick one and keep it across creatives.
- Pull viewers who reached itFrom your platform’s retention graph, take the count that reached your chosen point. Meta, TikTok and YouTube all expose this in their video or audience-retention reports.
- Pull total viewers who startedUse the count of viewers who began the video, defined consistently with your platform — a 3-second view, a thruplay, or a play, as long as it matches across tests.
- Read the rate and the bandThe big number is the share that stayed to your milestone. Low hold points at pacing or payoff; strong hold means the story is earning the attention the hook won.
- Export and actCopy a share link, download the CSV into your creative report, or print a PDF for the production review where you decide what to scale and what to recut.
RGM Expert Says
On paid social we split a video’s job into two questions, and hold rate owns the second one. The hook borrows attention in the first three seconds; the hold rate tells us whether the rest of the edit pays that loan back. A client will sometimes celebrate a viral-looking view count, and the retention graph quietly reveals that almost everyone bailed the moment the hook ended.
The number we trust is hold to a fixed milestone, not the platform’s headline average watch time, because average watch time hides the shape of the drop-off. A cliff at four seconds and a gentle slope to the end can produce the same average and demand completely different fixes. We read the curve, anchor on one milestone — usually 50% — and use this tool to make that single number portable across every creative in the test.
The most common pattern we see is a strong hook paired with a collapsing hold, which almost always means the opening wrote a check the body could not cash. The remedy is rarely a flashier hook; it is moving the payoff forward, tightening the cut, and adding captions so the sound-off majority can follow. Hold rate is how we prove the recut actually held more people.
How it works
Hold rate is a survival metric: of everyone who started, what fraction was still watching at the milestone you chose. It depends entirely on where you place that milestone, so fix it before you compare.
- Viewers reaching the point — count that watched to your chosen milestone (50%, 75%, completion).
- Total viewers — everyone who started the video, defined consistently with your platform.
- Worked example: 2,200 reached 50% ÷ 10,000 total = 22.0% hold rate to the halfway mark.
Hold rate complements hook rate (3-second views ÷ impressions). Platforms define a ‘view’ differently, so fix your definition before comparing; see Meta’s advertising help for video metric definitions. Bands shown here are rules of thumb, not platform guarantees.
Why hold rate, not view count, grades a video
A view count rewards the hook and ignores everything after it. Hold rate measures the part of the video that actually delivers the message, which is why it is the more honest scoreboard for creative. Two videos can earn identical views while one keeps a quarter of its audience to the payoff and the other loses almost everyone at second four.
Read hold rate as a curve, not a single point. The shape of the drop-off is the diagnosis: a sharp cliff early means the hook overpromised or the second beat is weak, while a steady decline means pacing fatigue. Fixing a cliff and fixing a slope call for different edits, so anchor on a milestone and study the whole retention graph behind it.
Hold rate also feeds the rest of your paid-social machine. The viewers who stay to your milestone are your warmest video audience — the right seed for remarketing and lookalikes — so a strong hold does double duty: it proves the creative works and it builds a better audience for the next campaign.
Hold rate in context
Hold rate depends on the milestone, the platform, the placement and the video length, so any range is orientation rather than a target. Compare creatives against each other at the same milestone, and against your own history.
| Hold band | Typical read | First lever |
|---|---|---|
| Below 10% | Steep early drop | Move payoff forward, tighten cut |
| 10% to 25% | Everyday mid-video hold | Pattern interrupts, captions |
| 25% to 40% | Strong retention | Study and reuse the structure |
| Above 40% | Exceptional | Scale spend, build lookalikes |
What creative teams say about hold
The hook buys you three seconds; the hold rate tells you whether you deserved the next thirty.
Build a repeatable creative engine — test the structure that retains, not the one-off that happened to go viral.