Hook Rate, Hold Rate, and Thumbstop for Facebook Ad Videos

Report-ready definitions and formulas

Hook rate: the first attention gate

Hook rate measures the share of served impressions that produce a short video view. A practical reporting formula is:

Hook Rate = 3-second views / impressions x 100

For example, 24,000 3-second views from 120,000 impressions equals a 20% hook rate. That does not prove the ad will convert. It only says the opener is relevant enough, surprising enough, or visually clear enough to earn a first look.

Keep the denominator stable. Do not compare one ad using impressions, another using reach, and a third using video starts. That creates false winners and makes creative decisions look more precise than they are.

Hold rate: the bridge from curiosity to substance

Hold rate measures how many early viewers continue to a deeper checkpoint. A practical formula is:

Hold Rate = checkpoint views / 3-second views x 100

For 15- to 30-second videos, many teams use a 15-second checkpoint. For longer VSL openings or demo cuts, use 50% completion or 20 seconds, whichever comes first. The exact checkpoint matters less than using the same rule across comparable assets.

Hold rate is where weak message architecture usually shows up. If the hook works but viewers disappear before proof, the creative probably has delayed evidence, slow pacing, unclear benefit sequencing, or a trust gap.

Thumbstop: useful proxy, not universal truth

Thumbstop is not a single universal native metric across every report setup. Many teams calculate a proxy close to hook rate:

Thumbstop Ratio = 3-second views / eligible video impressions x 100

Use this only for apples-to-apples comparison. If one export includes placements where video autoplay behavior differs, separate the placement groups before ranking creatives.

Average watch time belongs next to thumbstop, not inside it:

Average Watch Time = total watch seconds / video plays

A strong thumbstop with weak watch time usually means the opener creates curiosity but the next five to ten seconds do not pay it off.

Benchmarks that help without pretending to be universal

The ranges below are estimates for directional diagnosis, not guarantees. They should be rebuilt from your own account data by objective, placement mix, market, and creative length.

A 22% hook rate can be excellent in a cold, broad audience and mediocre in a warm retargeting pool. A 40% hold rate can be healthy for a 30-second problem-solution ad and weak for a short testimonial clip. Benchmarks are useful only when the comparison set is clean.

Before scaling, use the same discipline described in the Facebook ads scaling playbook: isolate the variable, confirm the signal across enough spend, and check whether attention metrics connect to downstream intent.

How to calculate the metrics from one export

Pull one clean reporting window first. Seven days is usually a better minimum than one or two days because day-of-week behavior can swing results, especially for consumer offers and affiliate funnels.

Export impressions, 3-second views, checkpoint views, average watch time, spend, CTR, landing page views, purchases or leads, placement, and creative length. Then bucket ads by objective and length before ranking them.

Worked example

Assume one 20-second video has:

• 120,000 impressions
• 24,480 3-second views
• 8,880 checkpoint views
• 7.2 seconds average watch time
• 1,120 landing page views

Hook rate is 20.4%. Hold rate is 36.3%. The opener is doing real work, but average watch time suggests many viewers leave before the full argument lands.

The next edit should not be a louder first line. It should probably move proof earlier, cut setup, tighten subtitles, or show the result before explaining the mechanism.

Data quality checks before declaring a winner

Use one denominator set for every creative in a test group. Compare videos of similar length and format. Separate Reels, Stories, feed, and in-stream placements when the mix is large enough to matter.

Do not overread low-spend tests. As a practical rule, rates from a few hundred impressions are unstable, and even a few thousand impressions can mislead when the audience is narrow. Treat early reads as triage, then require downstream confirmation before budget moves.

Diagnosing common creative failure patterns

Weak hook, strong hold

This pattern often means the video works for the right viewer but fails to signal relevance quickly enough. The people who stay are interested, yet too few viewers recognize themselves in the first seconds.

Fix the first frame, opening line, and visual specificity. Show the product, result, problem, or audience marker immediately instead of starting with a generic question.

Strong hook, weak hold

This is the classic curiosity gap problem. The opening earns the stop, but the body does not deliver enough proof, clarity, or forward motion.

Move evidence earlier. Replace one abstract claim with a concrete demonstration, customer moment, comparison, or objection answer. If the ad is for a VSL or affiliate offer, the first proof beat should arrive before the viewer feels tricked by the hook.

High thumbstop, low watch time

This usually points to a format issue. The thumbnail, motion, or first frame is attractive, but the edit becomes hard to follow.

Check subtitle timing, scene changes, audio clarity, and whether the first claim requires too much background knowledge. The cure is often simplification, not more intensity.

Decent video metrics, weak conversions

Attention is not persuasion. If hook, hold, and watch time are all healthy while conversion remains weak, audit offer fit, landing page continuity, load speed, price framing, compliance, and post-click proof.

This is where live competitive context can help. Daily Intel Service is useful when you need to compare your metrics against active scaling creatives and current funnel states, but your own account data still decides whether a creative can scale in your audience.

Watch-time targets by creative type

Average watch time is most useful when grouped by creative format. A six-second UGC clip and a 45-second VSL opener should not share one benchmark.

For a more transparent view of how Daily Intel Service evaluates live signals, review the Daily Intel Service methodology. The important principle is simple: judge each asset against its job, not against a universal video average.

Quality gates before you scale spend

A creative is not ready to scale just because hook rate improves. It needs clean measurement, a stable audience read, and evidence that attention is helping the funnel.

Use these checks before increasing budget:

• The same metric formulas were used across every compared ad.
• Placements and video lengths were grouped or controlled.
• Watch-time improvement also shows in CTR, landing page views, leads, purchases, or another business-relevant signal.
• Claims in the ad match the landing page and comply with Meta ad standards.
• The page and article experience follow Google helpful content guidance by serving the user first, not a search formula.
• Public ad libraries are treated as reference points, not proof that a creative is currently scaling.

Public snapshots can be stale or incomplete, so pair them with live funnel checks and your own test data. If you are evaluating external ad examples, understand the limits of Facebook ad library research before copying a hook.

Practical workflow for media buyers

Start by ranking ads inside one clean cohort: same objective, similar length, similar placement mix, and enough impressions to compare. Flag weak hook, weak hold, and weak watch-time patterns separately.

Next, assign one edit hypothesis per asset. Do not change the opening, proof order, subtitles, offer framing, and landing page all at once. A cleaner test may feel slower, but it gives you a decision you can reuse.

Finally, validate the winner against business outcomes. Daily Intel Service can add a live-signal layer for teams comparing active scaling ads, but it should support, not replace, disciplined testing.

Frequently Asked Questions

Q: What is a good hook rate for Facebook ads?
A: A practical starting range is about 15% to 35% when using 3-second views divided by impressions, but the real benchmark depends on audience temperature, placement, objective, and creative format.

Q: What is the difference between hook rate and hold rate?
A: Hook rate measures whether the ad earns early attention. Hold rate measures whether those early viewers continue to a deeper checkpoint where the argument, proof, or offer begins carrying the message.

Q: How do I calculate thumbstop if Meta does not show a thumbstop metric?
A: Use a consistent proxy such as 3-second views divided by eligible video impressions. Keep the same denominator across every creative in the comparison.

Q: Why can hook rate be high while conversions stay low?
A: Hook rate only proves attention capture. Conversion also depends on proof quality, offer fit, landing page continuity, pricing, trust signals, and post-click relevance.

Q: Should I compare one watch-time benchmark across all video lengths?
A: No. Compare short UGC clips with short UGC clips, 30-second VSL openings with similar VSL openings, and long demos with other long demos.

Q: Does Daily Intel Service replace my own Facebook ad testing?
A: No. It can provide active market references and funnel context, but your own tests are still required to validate audience fit, economics, and creative variants.