Hook rate tells you whether a video ad stops the scroll; hold rate tells you whether the ad keeps the attention it earned. Together, these metrics can expose promising creative before purchases or qualified leads accumulate—but neither proves that an ad will be profitable.
A high hook rate with a low hold rate means the opening worked and the rest of the ad lost the viewer. A low hook rate with strong retention means the message may be valuable, but too few people reach it. That distinction lets a marketing system make specific creative decisions instead of killing an ad because one blended metric looked weak.
What Hook Rate and Hold Rate Actually Measure
Most ad dashboards make creative performance look simpler than it is. They put impressions, video views, clicks, and conversions into the same table, even though those events happen at different stages of attention.
A useful video funnel separates them:
- The impression creates an opportunity to be seen.
- The hook earns the first two or three seconds.
- The body sustains attention.
- The offer earns a click.
- The landing page converts the click.
- The sales process turns the conversion into revenue.
Hook rate and hold rate evaluate steps two and three. That makes them leading indicators: they appear earlier than purchases, booked calls, or closed revenue.
Hook rate: Did the opening stop the scroll?
Hook rate is the percentage of impressions that produce a qualifying initial view.
A common Meta calculation is:
Hook rate = three-second video plays ÷ impressions × 100
If a video receives 100,000 impressions and 28,000 three-second plays, its hook rate is 28%.
TikTok also reports two-second and six-second video-view metrics. That means a TikTok analysis may use two-second viewers as the initial attention event, while a Meta analysis uses three-second plays. Neither definition is inherently superior, but mixing them in one benchmark produces bad conclusions.
The opening earns the hook rate. That generally includes:
- The first spoken sentence
- The first on-screen text
- The initial image or movement
- The person, product, or problem shown
- The first implied promise
An ad can have excellent production quality and still fail here. If the opening looks like an advertisement before it creates curiosity, viewers leave before the message begins.
Hold rate: Did the ad keep the viewer?
Hold rate measures the percentage of initial viewers who reach a later viewing milestone.
Two useful calculations are:
Six-second hold rate = six-second views ÷ initial views × 100
ThruPlay hold rate = ThruPlays ÷ three-second views × 100
Meta defines a ThruPlay as watching a video to completion when it is shorter than 15 seconds, or watching at least 15 seconds when it is longer. That makes ThruPlay useful, but not identical across videos of different lengths. A completed 10-second video and 15 seconds of a 45-second video both qualify even though they represent different viewing behavior.
Suppose an ad produces 30,000 three-second views, 12,000 six-second views, and 4,500 ThruPlays. Its six-second hold rate is 40%, while its ThruPlay hold rate is 15%.
Those two percentages answer different questions. The first tests whether the transition out of the hook works. The second tests whether the argument holds together deeper into the video.
Why raw view counts are not enough
A video with 20,000 initial views is not automatically better than one with 8,000. The first may have received four times as many impressions.
Rates normalize the result:
| Ad | Impressions | Initial views | Hook rate | Six-second views | Six-second hold rate |
|---|---|---|---|---|---|
| Ad A | 100,000 | 20,000 | 20% | 9,000 | 45% |
| Ad B | 40,000 | 14,000 | 35% | 4,200 | 30% |
Ad A holds viewers better after the opening. Ad B earns initial attention more efficiently. The correct action is not simply to declare one winner: test Ad B’s opening against Ad A’s body.
That is the real advantage of separating hook rate from hold rate. The metrics reveal which component to preserve.
How to Read the Metrics Without Fooling Yourself
Benchmarks are useful as alarms, not verdicts. Performance changes with audience temperature, placement, creative length, optimization event, product category, and platform measurement rules.
For short-form ads shown to cold audiences, the following ranges are practical starting points when hook rate uses three-second views divided by impressions:
| Metric | Weak signal | Workable | Strong | Investigate as a potential winner |
|---|---|---|---|---|
| Three-second hook rate | Below 20% | 20%–30% | 30%–40% | Above 40% |
| Six-second hold rate | Below 25% | 25%–35% | 35%–45% | Above 45% |
| ThruPlay rate on impressions | Below 5% | 5%–10% | 10%–15% | Above 15% |
These are diagnostic ranges, not universal industry standards. An account selling a complex B2B service should not blindly compare itself with a consumer entertainment app. Build account-specific benchmarks after enough ads have run under comparable conditions.
The four diagnostic combinations
Hook rate and hold rate create a simple creative matrix:
| Hook rate | Hold rate | What it usually means | Next test |
|---|---|---|---|
| High | High | The opening and message both work | Test offer, CTA, format, and audience expansion |
| High | Low | The opening overpromises or the body loses momentum | Keep the opening; rewrite the transition and body |
| Low | High | The message works for people who reach it | Keep the body; test new first frames and opening lines |
| Low | Low | The concept is weak or poorly matched to the audience | Replace the angle before polishing execution |
The high-hook, low-hold combination causes the most false positives. A provocative opening can manufacture three-second views while attracting the wrong people. If the body does not deliver the promised payoff, retention collapses.
The opposite combination is more valuable than it looks. Low-hook, high-hold creative often contains a strong argument buried under a slow introduction. A new opening can unlock performance without rebuilding the entire ad.
Do not compare unequal cuts
A 12-second product demonstration and a 60-second founder story should not share one completion-rate target. Completion becomes harder as duration increases, while a longer video has more time to qualify the viewer and explain a complicated offer.
Compare creative in controlled groups:
- Same platform and placement
- Similar duration
- Same audience temperature
- Same campaign objective
- Same initial-view definition
- Similar delivery period
- Enough impressions to reduce random noise
A difference between 36% and 35% hook rate after 400 impressions is not a useful signal. A sustained 36% versus 22% after tens of thousands of comparable impressions deserves attention.
Build a Testing System, Not a Benchmark Spreadsheet
The fastest creative teams do not ask, “Is 31% good?” They ask, “Which variable caused 31%, and what should we test next?”
That requires structured naming and controlled variation.
Test hooks independently from bodies
Create multiple openings for the same core ad:
- Direct problem: “Your best video ad may be dying in the first three seconds.”
- Contrarian claim: “Clicks are too late to diagnose bad creative.”
- Specific result: “This metric tells us which opening to keep before sales data matures.”
- Demonstration: Open on the dashboard changing as viewers drop.
- Pattern interruption: Start with the rejected version, then explain why it failed.
Keep the body, offer, length, audience, and landing page stable. If one opening produces a 38% hook rate and another produces 19%, the system has learned something useful about framing.
Next, attach the winning hook to multiple bodies. One body may lead with proof, another with mechanism, and another with objection handling. Hold rate determines which sequence keeps attention.
Use a decision window
Conversion data often matures slowly. A B2B campaign may need weeks to generate enough qualified opportunities to evaluate revenue, while hook and hold signals begin forming within hours.
That does not justify making decisions after 100 impressions. It supports a staged process:
- Check delivery and tracking first.
- Use early attention metrics to eliminate obvious creative failures.
- Give promising ads enough spend to evaluate clicks and landing-page behavior.
- Let conversion and revenue data make the final decision.
An ad should advance through increasingly expensive tests. Attention buys it the right to compete for clicks; clicks buy it the right to compete for conversions.
Budget around decisions
A testing budget should fund comparable decisions rather than spread money evenly across every possible variation.
| Creative testing budget | Initial structure | Budget per concept | Best use |
|---|---|---|---|
| $1,500 | 3 concepts × 3 hooks | About $167 per hook | Eliminate weak openings and identify one or two viable concepts |
| $3,000 | 4 concepts × 3 hooks | $250 per hook | Test multiple angles, then move winners into body variations |
| $6,000 | 5 concepts × 4 hooks | $300 per hook | Run a continuous system with hook, body, and offer stages |
These figures are planning math, not promised media outcomes. The required spend depends on CPM, conversion value, audience size, and how much evidence the business needs before acting.
The expensive mistake is giving 20 unrelated ads tiny budgets and learning nothing from any of them. A disciplined system tests fewer variables at once and records what each result means.
Track a complete creative scorecard
Hook and hold metrics should sit beside business outcomes:
| Funnel stage | Primary metric | Diagnostic question |
|---|---|---|
| Impression to initial view | Hook rate | Did the opening earn attention? |
| Initial view to six seconds | Six-second hold rate | Did the body begin cleanly? |
| Initial view to deeper view | ThruPlay hold rate | Did the message sustain interest? |
| View to click | Outbound CTR | Did the ad create action? |
| Click to lead or sale | Conversion rate | Did the page fulfill the ad’s promise? |
| Conversion to revenue | CPA, ROAS, qualified pipeline | Did the campaign create economic value? |
A high-hook ad with weak conversion may be entertainment, not marketing. A lower-volume ad that produces qualified opportunities at a sustainable acquisition cost is the actual winner.
How Agentic Ad Systems Find Winners Earlier
Traditional campaign management turns these metrics into a weekly reporting task. Someone exports results, updates a spreadsheet, notices a pattern, requests another edit, and waits for the next reporting cycle.
An agentic system treats each ad as structured evidence.
BattleBridge operates 10 deployed AI agents across three servers with 46 registered skills. Those systems support production properties that include a senior-living directory spanning 977 cities, 51 states, and 4,757 communities, plus a CRM containing 8,442 contacts. The important point is not the raw agent count. It is that each agent has a defined job, data source, decision boundary, and handoff.
The same architecture applies to advertising:
- A collection agent retrieves placement-level performance.
- A measurement agent normalizes hook and hold definitions.
- A creative agent identifies the failing section of each ad.
- A testing agent proposes controlled variants.
- A performance agent waits for conversion and revenue confirmation.
- A human approves consequential budget or brand decisions.
That is agentic marketing: software does not merely generate copy; it observes results, chooses the next bounded action, and preserves what it learns.
Score components, not just ads
A conventional report labels an ad “winner” or “loser.” That throws away most of the information.
An agentic system can score individual components:
- Opening line
- First frame
- Visual format
- Spokesperson
- Problem statement
- Proof element
- Offer
- CTA
- Duration
- Audience and placement
If a specific first frame repeatedly produces hook rates above the account median, the system should reuse the pattern. If a body structure consistently loses half its initial viewers before six seconds, it should be retired or rewritten.
This turns creative testing into cumulative learning. The 50th ad benefits from the first 49 instead of starting from zero.
Add guardrails before automation
Early prediction can become early overreaction. An autonomous system needs minimum-impression thresholds, comparable cohorts, tracking checks, budget limits, and human approval rules.
It also needs a clear hierarchy of evidence:
Revenue beats conversions. Conversions beat clicks. Clicks beat deep views. Deep views beat initial views.
Earlier metrics accelerate decisions; they do not outrank later business outcomes.
Our architecture for an agentic marketing system follows that principle: automate observation and repeatable decisions while preserving explicit boundaries around consequential actions.
The goal is not to remove judgment. It is to stop wasting judgment on data collection, spreadsheet maintenance, and preventable delays.
Show Me How Ads Arsenal Finds Winning Creative
No platform migration or black-box media commitment is required to understand the system. Ads Arsenal applies the same multi-agent operating model BattleBridge uses across live content, CRM, and directory infrastructure.
Frequently Asked Questions
What is hook rate in video ads?
Hook rate is the percentage of impressions that become an initial video view, usually measured at two or three seconds. In a hook rate hold rate video ads report, it shows whether the opening stops enough people to earn attention.
How is hook rate calculated?
Divide qualifying initial views by impressions and multiply by 100. If an ad generates 28,000 three-second views from 100,000 impressions, its hook rate is 28%.
What is a good hook rate on TikTok or Reels?
For short-form cold traffic, a three-second hook rate of 25% to 35% is a useful starting range, while results above 35% are often strong. Compare ads using the same platform, audience, placement, and view definition.
What is hold rate?
Hold rate measures how many initial viewers continue watching to a later milestone, such as six seconds, 15 seconds, or completion. It separates ads that merely attract attention from ads that sustain it.
Can you predict a winning ad from hook rate?
A hook rate hold rate video ads analysis can identify promising creative early, but hook rate alone cannot prove profitability. Confirm the prediction with hold rate, click-through rate, conversion rate, acquisition cost, and downstream customer quality.
Get Your Free Hook Rate Hold Rate Video Ads Audit
BattleBridge runs autonomous AI agents that handle this end to end — research, content, distribution, and reporting — for a flat monthly rate instead of an agency retainer. We'll audit your current setup, show you exactly where agents outperform your existing stack, and hand you the findings whether you hire us or not.
Get your free audit — 30 minutes, no pitch deck, real numbers.