Measure the First Three Seconds Without Losing Sight of Sales
Hook rate answers an early, narrow question: how often did an impression become a qualifying early video view?
That makes it a practical diagnostic for UGC creators, paid-social teams, brands, affiliate marketers, and creative strategists. It does not make hook rate a verdict on creative quality. An ad can earn plenty of early views and still lose people before the offer, attract the wrong audience, generate weak clicks, or fail commercially. Another ad can produce a lower hook rate while qualifying a smaller, more profitable audience.
The right approach is to define the metric precisely, compare like with like, test the opening without changing the entire ad, and validate any apparent winner against retention, clicks, conversions, CPA, revenue, and ROAS.
What hook rate measures—and what it cannot tell you
Hook rate commonly means the percentage of impressions that produce a qualifying early video view. For Meta video ads, practitioners usually calculate it as:
Meta 3-second hook rate (%) = 3-second video plays ÷ impressions × 100
If an ad receives 10,000 impressions and generates 2,500 qualifying three-second plays, its hook rate is 25%. This formula is an operator-created ratio built from the impression and three-second-play events reported by Meta, rather than a standardized default platform metric (Gino Gagliardi’s custom-metric guide).
That distinction matters. Two dashboards can both display a column called “hook rate” while using different viewing thresholds or denominators. A useful report makes the formula visible instead of assuming everyone means the same thing.
Hook rate is evidence of early viewing behavior. It indicates how many delivered impressions reached a defined viewing threshold. It does not establish:
- Whether the viewer was deliberately paying attention
- Whether the person understood or remembered the message
- Whether the opening attracted the intended audience
- Whether the rest of the video retained interest
- Whether the viewer clicked
- Whether the traffic converted
- Whether the campaign was profitable
“Scroll-stopping” remains useful shorthand for the purpose of a hook, but it should not be treated as a proven explanation for every qualifying play. The metric reports what crossed the threshold, not why.
That limitation does not make hook rate useless. It defines the metric’s role. Hook rate is the first line in a creative diagnosis: it helps determine whether the opening deserves investigation, preservation, or further testing. By itself, it says nothing about what happened after the threshold.
It also helps to distinguish the creative hook from the hook rate:
- The hook is the opening visual, spoken line, text overlay, demonstration, action, sound, or combination of elements intended to earn continued viewing.
- The hook rate is a measurement applied to early views generated by the delivered ad.
A creator makes the hook. A reporting system calculates the rate. Conflating the two encourages teams to treat a numerical result as a complete judgment on the creative rather than one observation about its opening.
How to calculate hook rate on Meta and TikTok
The correct formula depends on the platform and qualifying-view event. Always put the platform and threshold in the metric name.
Calculating Meta hook rate
For Meta video ads, use:
Meta 3-second hook rate (%) = 3-second video plays ÷ impressions × 100
A worked example:
- Impressions: 10,000
- Three-second video plays: 2,500
- Calculation: 2,500 ÷ 10,000 × 100
- Meta 3-second hook rate: 25%
Meta counts the underlying event when at least three seconds of the video play—or nearly the full video plays when the asset is shorter than three seconds—and excludes replays, according to the platform definition summarized in this Meta hook-rate explainer.
That creates an important edge case. For a video shorter than three seconds, the qualifying event represents nearly the entire asset. For a video only slightly longer than three seconds, it represents a large share of the asset. A nominal “three-second hook rate” therefore describes materially different viewing depths across a two-second clip, a six-second ad, and a 30-second video.
Do not casually place those assets into one creative benchmark. Separate very short videos or, at minimum, include video length beside the rate.
To create the metric in Meta Ads Manager:
- Open Columns.
- Select Customize columns.
- Choose Create custom metric.
- Use 3-second video plays as the numerator.
- Use impressions as the denominator.
- Format the result as a percentage.
- Save it in the relevant custom column set.
This setup path and percentage formatting are documented in a practitioner guide to Meta custom metrics.
Name the column Meta 3-second hook rate, not merely “hook rate.” The longer label prevents someone from later combining it with a two-second metric, an organic video measure, or a ratio based on video starts.
Calculating TikTok hook rate
For TikTok, a common paid-social calculation is:
TikTok 2-second hook rate (%) = 2-second video views ÷ impressions × 100
For example, 3,100 two-second views from 10,000 impressions produce a TikTok two-second hook rate of 31%. BMG360 uses this two-second-views-over-impressions calculation and also refers to the result as thumb-stop rate in its TikTok hook-rate analysis.
Do not compare a TikTok two-second result directly with a Meta three-second result. The qualifying events differ, as do the feeds, placements, user behavior, delivery systems, and creative conventions. The numbers may sit beside one another in an executive dashboard, but they should remain separate metric families.
Keep the denominator explicit
Impressions are the usual denominator in the formulas above. Replacing impressions with video starts would create a different ratio:
Qualifying early views ÷ video starts × 100
That ratio could be useful for a particular analysis, but it no longer answers the same question. It measures progression from a recorded start to the early-view threshold rather than progression from all impressions.
If a team uses that alternative, give the metric an unambiguous name such as three-second views per video start. Calling both calculations “hook rate” makes historical and cross-team comparisons unreliable.
The same discipline applies to organic reporting. A platform may provide views, plays, reached viewers, or another exposure measure instead of paid impressions. That can support an organic early-retention ratio, but it should not be silently merged with the paid-ad formula.
What is a good hook rate? Use benchmarks carefully
The reviewed evidence does not establish an authoritative universal hook-rate benchmark.
Some commercial and practitioner sources treat Meta results around 20%–25% as a workable baseline and results of 30% or more as a strong early-view signal. One benchmark summary, for example, describes 20%–25% as workable and provides overlapping higher-performance bands (AdManage’s Meta benchmark guide).
Another vendor guide similarly uses 20%–25% as a working range and above 30% as a strong attention signal, while explicitly recommending account history as the primary benchmark (Zeely’s Meta hook-rate guide). These ranges are directional practitioner guidance, not platform standards.
The published bands vary materially and generally do not disclose enough about sample size, geography, industry, placement, audience, objective, offer, video length, selection criteria, or statistical methodology to support universal conclusions. Platform-event definitions, practitioner workflows, and benchmark opinions should therefore be evaluated separately.
Use third-party percentages, at most, as orientation. Do not turn them into rules such as:
- “Anything below 20% is a failed ad.”
- “A 30% result is always good.”
- “A 40% result should automatically be scaled.”
- “Every creator must deliver at least the account-wide average.”
A safer primary benchmark is the advertiser’s own matched history. Compare an ad with previous ads delivered under similar conditions:
- Same platform
- Same placement
- Same audience type or temperature
- Same creative format
- Same campaign objective
- Same or comparable offer
- Similar video length
- Same reporting window
- Similar attribution and optimization settings
Build internal baselines using historical medians or account percentiles for matched cohorts. For example, calculate the median Meta three-second hook rate for prospecting UGC ads in Reels that promote the same offer and fall within the same length band. That is more actionable than comparing every ad with one percentage found online.
Medians are often preferable to simple averages because a few unusually high or low results can distort an account average. Percentiles add context: a result can be described as being in the upper quarter of comparable account creatives without pretending the same threshold applies elsewhere.
Placement mix can change the aggregate result
An aggregate Meta rate can blend delivery across Feed, Reels, Stories, Audience Network, and other placements. If two variants receive different placement mixes, their overall rates can differ even when their within-placement results are similar.
Suppose Variant A receives most of its impressions in a placement that historically produces higher early-view rates for the account, while Variant B receives more delivery in a lower-rate placement. The aggregate comparison may favor A without proving that A has the stronger opening. A practitioner discussion of placement-mix distortion in Meta hook rates warns against overinterpreting account-level differences for this reason.
Before naming a winner:
- Break out results by placement.
- Check whether each variant had comparable delivery.
- Compare within placements where event counts are adequate.
- If necessary, describe the aggregate result as inconclusive.
A percentage is not enough
A difference between 30% and 35% may be operationally important, noise, or an artifact of delivery. The percentages alone cannot tell you which.
Record the underlying counts:
| Variant | Impressions | Qualifying views | Hook rate |
|---|---|---|---|
| A | 200 | 60 | 30% |
| B | 200 | 70 | 35% |
| C | 20,000 | 6,000 | 30% |
| D | 20,000 | 7,000 | 35% |
The percentages are identical across the two comparisons, but the amount of information behind them is not. A defensible decision also considers delivery comparability and uncertainty rather than assuming any visible difference is meaningful.
There is no universal impression count, budget, or test duration that guarantees a sound decision. Published fixed windows, spend levels, and impression thresholds are generally presented as practitioner rules of thumb without enough methodology to make them universal.
Treat an unusually low matched result as a reason to investigate the opening, placement, format, crop, and audience. Do not treat it as proof that the hook alone caused poor campaign performance.
Hook rate vs hold rate, CTR, completion, and ROAS
Hook rate is one stage in a sequence. Nearby metrics answer different questions.
| Metric | Typical definition | Question it answers | Main limitation |
|---|---|---|---|
| Hook rate | Qualifying early views ÷ impressions | Did impressions become early views? | Does not show what happened after the threshold |
| Hold or retention rate | A named later-view event ÷ a named earlier-view event, or average percentage watched | Did viewers continue through the body? | The term is not standardized |
| Completion rate | Full-video views ÷ the platform’s specified eligible-view or start event | How often was the entire video watched? | Denominators vary and must be named; video length strongly affects the result |
| Average watch percentage | Average amount watched ÷ video duration | What average share of the asset was watched? | Can hide where drop-offs occurred |
| CTR | A specified click event ÷ impressions | Did the delivered message produce that type of click? | “All clicks” and link clicks are not equivalent; click quality remains unknown |
| CPA | Spend ÷ defined acquisitions | How much did each measured acquisition cost? | Depends on the conversion definition and measurement setup |
| ROAS | Attributed revenue ÷ ad spend | How much attributed revenue was produced per unit of spend? | Does not by itself describe margin or incrementality |
ROAS is a different economic question from early viewing: the ratio compares attributed revenue with spend rather than three-second plays with impressions (Eonik’s metric comparison).
The sequence of questions is straightforward:
- Hook rate: Did the opening earn a qualifying early view?
- Retention or hold: Did the body keep viewers?
- CTR: Did the message produce the specified click?
- Conversion rate and CPA: Did the traffic convert efficiently?
- Revenue and ROAS: Did the campaign generate commercially useful returns?
Reports should name the exact click event used for CTR—such as link clicks—rather than assuming every click-based ratio is comparable. The same precision required for hook rate should apply to the rest of the measurement chain.
“Hold rate” needs a formula, not just a label
Hold rate is not standardized. Sources variously define it as:
- ThruPlays divided by three-second plays
- Fifteen-second views divided by three-second plays
- Six-second views divided by two-second views
- Video completion rate
- Average percentage watched
These are not interchangeable. A report should say ThruPlay-from-3-second hold rate, 15-second-from-3-second retention, or average watch percentage, depending on the actual calculation.
ThruPlay also should not automatically be described as completion. On Meta, it can represent at least 15 seconds watched for longer assets, so a viewer can qualify without finishing the video. A detailed comparison of hook rate, hold rate, and ThruPlay explains this distinction and presents more than one commonly used hold-rate formula.
Video length further complicates retention comparisons. Completing a six-second ad is not the same viewing task as completing a 45-second ad. Average watch percentage, completion rate, and later-view retention should therefore be segmented by materially different length bands.
Whenever possible, inspect the retention curve as well as the aggregate. A single average cannot show whether viewers left steadily, abandoned the video at a weak transition, or remained until a confusing offer appeared.
A diagnostic matrix for hook rate and downstream performance
Metrics indicate where to investigate. They rarely prove why a result occurred. Treat each diagnosis below as a testable hypothesis.
| Pattern | What it may indicate | What to investigate next |
|---|---|---|
| Low hook rate, low retention | Opening weakness plus additional body problems—or poor delivery fit | First frame, message speed, crop, audience, placement, format, then retention drop-offs |
| High hook rate, low retention | Hook-body mismatch or weak transition | Overpromise, irrelevant curiosity, clickbait, pacing, proof, body fulfillment |
| Strong hook and retention, weak CTR | Message is watched but not prompting action | Relevance, product understanding, offer clarity, proof, CTA |
| Healthy CTR, weak conversion efficiency | The ad earns clicks but the traffic or conversion path underperforms | Audience quality, offer, landing page, checkout, destination continuity |
| Low hook rate, strong CPA or ROAS | Smaller viewing group may be highly qualified | Preserve the control; test broader hooks without replacing it |
| High hook rate, weak CPA or ROAS | Early attention is not becoming commercial value | Audience qualification, message continuity, offer, post-click experience |
Low hook rate and low retention
Inspect the opening first because fewer impressions are reaching the early-view threshold. Audit the first frame, initial statement, product visibility, crop, sound dependence, and placement fit.
Then examine the retention curve for later failure points. A weak opening and weak body can coexist. Audience mismatch or unfavorable placement delivery can also suppress both numbers, so do not assign all causation to the first few seconds.
High hook rate and low retention
The opening is generating early views, but the body is not sustaining them. Test whether:
- The opening promises something the body does not deliver
- Curiosity is resolved in an irrelevant or disappointing way
- The result is exaggerated or insufficiently supported
- The transition into the body feels abrupt
- The body becomes generic after a specific opening
- The opening attracts people outside the intended buyer group
A high hook rate can be produced through spectacle or ambiguity. That attention has limited value if the following message cannot fulfill or commercially use it.
Strong early viewing and retention, weak CTR
Viewers appear willing to watch, yet they are not clicking. Investigate:
- Is the product identifiable?
- Is the problem connected clearly to the product?
- Does the proof support the claim?
- Is the offer understandable?
- Does the viewer know what to do next?
- Is the message relevant to the delivered audience?
- Does the CTA arrive naturally and visibly?
The answer is not necessarily a louder CTA. The video may be entertaining without making the product’s role clear.
Healthy CTR, weak conversions, CPA, or ROAS
Once the ad is generating clicks, move scrutiny downstream. Review audience quality, offer strength, price expectations, landing-page speed and clarity, checkout friction, conversion tracking, and continuity between the ad and destination.
A visitor who clicks for one implied promise but reaches a page centered on another offer may abandon even when both hook rate and CTR look healthy.
Low hook rate but strong economics
Do not replace a profitable ad merely because its hook rate trails an internet benchmark. Its opening may repel casual viewers while qualifying the right buyers.
Keep the profitable version as the control. Test alternatives designed to broaden early attention, but require them to preserve conversion efficiency. The goal is useful attention, not the largest possible early-view audience.
Two hypothetical examples
High-hook, weak-commercial ad: A skincare video opens with an unusual visual transformation that earns a high early-view rate. The reveal is mostly a camera trick, the product’s role remains vague, and curiosity-driven viewers rarely click or buy. The hook succeeds at attracting attention but fails at qualification and promise fulfillment.
Lower-hook, stronger-commercial ad: A specialist software ad opens by naming a narrow workflow problem. Many general viewers leave immediately, producing a lower hook rate. The relevant professionals who remain understand the use case, click at a healthy rate, and convert efficiently. The smaller early-view audience produces better economics.
These examples illustrate hypotheses, not fixed laws. Similar metric patterns can have different causes, which is why the next step should be a focused test rather than a confident story.
How to run a controlled hook test
A controlled hook test asks whether changing the opening improves performance while the rest of the ad remains meaningfully stable.
Start with a clear hypothesis:
An immediate product demonstration will produce more qualifying early views than a spoken problem statement, while maintaining downstream conversion efficiency.
That is more useful than “test three hooks.” It identifies the variable, expected result, and commercial guardrail.
Build modular variants
Use one unchanged video body and attach several distinct two-to-four-second openings that transition naturally into it. This modular method is described in a Meta hook-testing workflow, although its fixed volume and timing suggestions should be treated as rules of thumb rather than universal requirements.
Keep these variables constant where practical:
- Creator
- Product
- Core claim
- Offer
- Body sequence
- Proof
- Call to action
- Audience
- Placement
- Campaign objective
- Video length
- Reporting window
Perfect control is not always possible. A demonstration may require a slightly different transition or total duration. Record those exceptions rather than pretending the variants are identical.
Either change one interpretable hook variable at a time or make each concept structurally distinct enough that the learning can be named. For example:
- Problem statement versus immediate demonstration
- Outcome-first versus process-first
- Product visible versus product initially off-screen
- Text-led versus action-led
- Direct audience callout versus broad curiosity
Record counts, not percentages alone
For every variant, save:
- Impressions
- Qualifying views
- Calculated hook rate
- Placement distribution
- Retention endpoint and formula
- Clicks and the click definition
- Conversions and the conversion definition
- Spend
- Revenue where available
Break results out by placement before concluding that one opening won. An aggregate leader may simply have received a more favorable delivery mix.
Do not assume that a fixed seven-day window, fixed budget, or fixed number of impressions is sufficient for every test. Ask instead:
- Are there enough impressions and qualifying views to distinguish the variants?
- Is delivery reasonably comparable?
- Is the difference large enough to matter operationally?
- Are downstream event counts adequate for a commercial decision?
- Would choosing the wrong variant carry substantial cost?
An opening can advance provisionally on early-view evidence, but it should not become a commercial winner until retention, CTR, conversion rate, CPA, ROAS, and revenue have been checked where available.
If the current ad is profitable, keep it as a control. Do not pause it merely to force spend into an unproven broader hook.
Reusable hook-test log
| Field | What to record |
|---|---|
| Variant name | Short, consistent identifier |
| Platform threshold | Meta three-second, TikTok two-second, or another explicit event |
| Hook hypothesis | The predicted effect and reason for the test |
| Fixed variables | Body, creator, offer, CTA, length, and other controls |
| Placement | Feed, Reels, Stories, or other placement |
| Audience | Prospecting, retargeting, interest, broad, or relevant cohort |
| Impressions | Delivered impression count |
| Qualifying views | Numerator event count |
| Hook rate | Labeled calculation |
| Retention endpoint | Exact event or average-watch definition |
| Clicks | Count and click definition |
| Conversions | Count and conversion definition |
| CPA | Cost per defined acquisition |
| ROAS | Attributed revenue divided by spend |
| Decision | Keep, reject, iterate, or gather more data |
| Next iteration | The specific follow-up hypothesis |
What to test in the opening of a UGC video
Opening ideas are hypotheses, not universal formulas. A style that works for one audience or placement can fail elsewhere—or raise early views while reducing buyer quality.
First-frame audit
Review the first frame before rewriting the spoken script:
- Immediate subject clarity: Can viewers tell what they are looking at?
- Product visibility: Is the product present when it needs to be?
- Motion: Would purposeful action clarify the scene or create contrast?
- Text readability: Can the overlay be read quickly on a phone?
- Placement-safe cropping: Are faces, products, captions, and demonstrations visible across intended placements?
- Visual contrast: Does the focal subject separate clearly from the background?
- Sound-off comprehension: Can the basic premise be understood without audio?
These variables—including product visibility, motion, text, crop, and audience fit—are also emphasized in Zeely’s diagnostic checklist.
Motion is a variable to test, not a guaranteed cure. A static result image might outperform purposeless movement if it communicates the value more quickly. Likewise, product visibility may help a demonstration but undermine a story built around a legitimate reveal.
Message audit
Ask whether the opening establishes:
- Audience identification: Who is this for?
- Problem specificity: Is the pain concrete rather than generic?
- Outcome framing: Is the desired result understandable?
- Proof: Is there a credible reason to continue watching?
- Relevance: Does the message connect with the intended viewer?
- Curiosity: Is there a useful unanswered question?
- Offer preview: Does the opening honestly set up what follows?
Curiosity can earn early views, but it can also attract poorly qualified attention. The reveal must matter to the intended buyer and connect to the product, body, and offer.
Hook families to test
| Hook family | Variable being tested | Main risk |
|---|---|---|
| Problem-led | Whether a specific pain creates self-recognition | The problem may feel generic, exaggerated, or irrelevant |
| Outcome-first | Whether showing the desired result earns attention | The outcome may seem implausible or obscure the product’s role |
| Question | Whether direct inquiry prompts mental participation | Broad questions can attract curiosity without buying intent |
| Contrarian | Whether challenging an assumption creates interest | It can feel forced or undermine trust |
| Demonstration | Whether immediate product use improves clarity | The action may lack context or differentiation |
| Transformation | Whether before-and-after contrast communicates value | The change may appear unrepresentative or insufficiently explained |
| Social proof | Whether evidence or other users increase credibility | Weak, vague, or irrelevant proof can distract |
| Storytelling | Whether a narrative setup sustains interest | Slow exposition can delay the product or value |
| Unusual visual | Whether visual novelty earns early viewing | Spectacle can attract the wrong audience |
| List-based | Whether a structured promise sets expectations | The list can feel generic or postpone the key point |
Treat visual, spoken-copy, text-overlay, and audio elements as separable variables where production capacity allows. A promising spoken line may fail because its text overlay is unreadable; an effective demonstration may be weakened by unrelated audio.
Openings that depend entirely on sound can struggle in sound-off contexts, but that does not mean silent-first treatment always wins. Test captions, visual demonstration, spoken delivery, sound design, and combinations of them within the actual placement environment.
Common candidates for revision include slow buildup, generic greetings, logo-first introductions, late product appearance, vague claims, and damaging placement crops. These are practitioner hypotheses, not proven universal failures. A logo may be useful when brand recognition is the intended qualifier; a deliberate buildup may work when the audience already has context.
Finish every opening review with a promise-fulfillment check:
- Does the body deliver what the hook implied?
- Does the proof support the opening claim?
- Does the offer match the problem or outcome introduced?
- Does the CTA follow logically?
- Does the landing page continue the same promise?
Improving hook rate by widening the gap between the opening and the actual offer is not an improvement.
Turn hook rate into a repeatable creator-and-brand workflow
A single winning hook is less valuable than a system that explains where, when, and why it appeared to work.
Build a tagged hook library organized by:
- Platform
- View threshold
- Placement
- Audience
- Offer
- Funnel stage
- Creator
- Format
- Hook concept
- Visual treatment
- Copy treatment
- Audio treatment
- Launch date
Store the exact metric definition beside every result. A Meta three-second rate should never be averaged with a TikTok two-second rate. If a formula or platform definition changes, preserve the historical definition rather than rewriting old records.
Every library entry should include the associated body, offer, retention endpoint, CTR, conversions, CPA, revenue, and ROAS. Otherwise, a team may repeatedly reuse an opening because it earned attention even though the associated ads failed commercially.
Translate findings into creator briefs as multiple hypotheses:
- “Test immediate demonstration against a spoken problem callout.”
- “Create one version where the product is visible in the first frame and one where the outcome appears first.”
- “Try a direct audience identifier and a proof-led alternative.”
Avoid asking creators to reproduce one supposedly universal winning line. The original result may depend on the creator’s delivery, audience, placement, offer, or moment in time.
Keep instructions specific enough to support a test while leaving room for natural delivery. Specify the intended hypothesis, required claim, first-frame information, crop constraints, and transition point. Do not script every gesture unless that gesture is the variable being tested.
For affiliate and commerce content, connect creative observations with commercial outcomes. Larping Agency’s TikTok Shop affiliate guide recommends recording hooks and formats while tracking clicks, add-to-cart activity, orders, and returns rather than relying only on views or likes. The guide does not provide a proprietary hook-rate benchmark, but it reflects the appropriate measurement principle: attention must be read in commercial context.
Separate observations from conclusions in every report.
Observation:
Variant B generated a higher Meta three-second hook rate than Variant A in Reels, while Feed results were similar. Variant B’s CTR and conversion count were lower.
Hypothesis:
The unusual visual may have attracted broader but less qualified curiosity in Reels.
The observation describes the data. The hypothesis proposes an explanation to test. Keeping those distinct prevents a plausible narrative from hardening into an unsupported fact.
Use this compact reporting checklist:
- [ ] Formula labeled with platform and threshold
- [ ] Comparison matched by audience, placement, offer, format, and window
- [ ] Placement breakout reviewed
- [ ] Impressions and qualifying-view counts recorded
- [ ] Retention or hold definition stated
- [ ] CTR click event identified
- [ ] Conversion definition recorded
- [ ] CPA, revenue, and ROAS reviewed where available
- [ ] Observation separated from causal hypothesis
- [ ] Decision recorded
- [ ] Next test specified
Frequently asked questions about hook rate
Is hook rate the same as thumb-stop rate?
Often, but not reliably.
Practitioners commonly use “hook rate,” “thumb-stop rate,” “thumb-stop ratio,” and “three-second view rate” for the same general idea: qualifying early views divided by impressions. Some TikTok reporting uses a two-second event, while Meta reporting commonly uses a three-second event. Focal, for example, uses hook rate and thumb-stop rate as alternative names for a three-second-views-over-impressions calculation in its UGC hook explainer.
Never rely on the label alone. Check the platform, numerator, denominator, and threshold. Two teams can both report “thumb-stop rate” while calculating different ratios.
What is a good hook rate for Meta ads?
There is no authoritative universal percentage. Some third-party sources treat approximately 20%–25% as a workable Meta baseline and 30% or more as a strong early-view signal, but the ranges vary and generally lack enough disclosed methodology to serve as industry standards.
Use matched account history as the primary benchmark. Compare the same placement, audience type, objective, offer, format, video-length band, and reporting window. Then check retention, CTR, conversions, CPA, and ROAS before deciding whether the creative is genuinely strong.
Why can an ad have a high hook rate but poor ROAS?
The opening may attract irrelevant curiosity, overpromise the result, or fail to transition into the body. The message may retain viewers without communicating the product or offer clearly. Alternatively, the ad may generate qualified clicks but lose them through weak audience fit, pricing, landing-page continuity, checkout friction, or another conversion-path issue.
Hook rate measures early viewing, while ROAS measures attributed revenue relative to ad spend. They concern different stages of the journey, so strong performance in the first does not guarantee strong performance in the second.
How many impressions are needed before comparing hook variants?
There is no universal minimum that makes every comparison valid. The required volume depends on the baseline rate, difference between variants, placement mix, event variability, and consequence of making the wrong decision.
Record both impressions and qualifying views, compare matched delivery, and assess uncertainty. If the commercial decision depends on CPA or ROAS, early-view volume alone is insufficient; the variants also need enough downstream conversion evidence for that decision.
Can I compare Meta and TikTok hook rates?
Not as equivalent measurements. Meta hook rate commonly uses three-second video plays divided by impressions, while TikTok commonly uses two-second video views divided by impressions. A cross-platform explainer from rule1 documents these separate thresholds while also warning that results vary by industry, audience, and format (rule1’s hook-rate guide).
Report them separately as Meta 3-second hook rate and TikTok 2-second hook rate. You can compare whether each creative improves relative to its own platform baseline, but the raw percentages should not be treated as directly interchangeable.
Hook rate is the first line in a creative diagnosis, not the final score. Label the platform and view threshold, compare matched delivery conditions, and test the opening without rebuilding the entire ad. Preserve downstream guardrails throughout.
The practical goal is not the highest possible three-second percentage. It is an opening that attracts the right viewer, accurately sets up the message, and contributes to commercially useful outcomes.