Hook Rate, Hold Rate and Thumbstop: Metrics That Predict Winners

Hook rate and hold rate are the two numbers that tell you whether a video ad is working before the purchase data can. Hook rate (often called thumbstop rate or thumbstop ratio) is the share of impressions that watched at least 3 seconds. Hold rate is the share of those 3-second viewers who kept watching to 15 seconds or ThruPlay. Together they show whether the opening earns attention and whether the body keeps it, which is exactly what a creative strategist needs to know on day two of a test. This guide covers the formulas, the Ads Manager setup, realistic ranges and how to act on them.

Quick answer: Hook rate = 3-second video plays divided by impressions. Hold rate = ThruPlays (or 15-second plays) divided by 3-second plays. Thumbstop is another name for hook rate. Set both up as custom metrics in Ads Manager. Practitioner ranges for cold traffic: hook rate 25 to 35% is healthy and under 20% is weak; hold rate of 30% or more is solid. Low hook means fix the first 2 seconds. Good hook but low hold means fix the middle. Good both but poor cost per purchase usually means the offer or landing page.

What Are Hook Rate, Hold Rate and Thumbstop?

Meta reports video plays at several thresholds: 3 seconds, ThruPlay (15 seconds or the full video if shorter), and 25, 50, 75, 95 and 100% of length. Hook rate uses the first of those. If an ad is served 10,000 times and 3,000 people watch at least 3 seconds, the hook rate is 30%. It measures the first frame, the first line and the first motion: did the ad stop the thumb. Thumbstop rate and thumbstop ratio are the same calculation under different names.

Hold rate picks up where hook rate stops. Of the 3,000 who watched 3 seconds, if 1,000 reached ThruPlay, the hold rate is 33%. It measures whether the body of the video delivered on the opening. Some teams define hold rate as ThruPlays divided by impressions instead, which is a stricter number; either works as long as you use one definition consistently and compare ads against each other, not against someone else's benchmark.

How to Set Up Hook Rate and Hold Rate in Ads Manager

Neither metric exists as a default column, but Ads Manager lets you build them as custom metrics in a few minutes. Once created, they appear as columns you can add to any report and save in a column preset, so the whole team reads the same numbers. Build them once at the account level and add them to your creative testing report alongside spend, impressions, outbound CTR and cost per purchase.

1. In Ads Manager, open the Columns dropdown and choose Customize Columns.
2. Click Create Custom Metric.
3. Name it Hook Rate. Set the format to percentage.
4. Formula: 3-Second Video Plays divided by Impressions. Save.
5. Create a second metric named Hold Rate. Formula: ThruPlays divided by 3-Second Video Plays. Format percentage. Save.
6. Optionally create ThruPlay Rate: ThruPlays divided by Impressions.
7. Add all three to a column preset named Creative Testing and save it.
8. Share the preset with anyone else on the account so reads are consistent.

If you prefer to see it done on screen, this short tutorial walks through creating both metrics step by step inside Ads Manager.

What Good Looks Like: Practitioner Ranges

There is no official benchmark, and ranges vary with placement, audience temperature and format. Reels placements tend to produce lower hook rates than Feed because the autoplay environment counts more passive views. Retargeting audiences hook more easily than cold ones. UGC-style video often hooks higher and holds lower than a produced demo. So treat the figures below as the rough consensus among media buyers for cold, broad DTC traffic, and then build your own account baseline from your top ten ads.

The most useful comparison is relative: your new ad against the median of your account's winners from the last 90 days. If your winners averaged a 28% hook rate and the new ad shows 19% after 3,000 impressions, you have a hook problem regardless of what any external range says. Build that baseline once a quarter and put it at the top of your testing report.

How to Diagnose a Creative With These Metrics

The power of these metrics is that they localize the problem. Each combination of hook, hold and CTR points to a specific part of the ad or funnel, which tells the creative team what to fix rather than just that something is wrong. Read them in order: hook first, then hold, then CTR, then cost per purchase, and stop at the first one that is off.

One trap: a very high hook rate with very low hold is not a win. It usually means the opening promised something the video does not deliver, and those viewers leave annoyed. Judge the pair together. A 24% hook with a 40% hold often beats a 38% hook with a 12% hold on cost per purchase.

Which Creative Metrics Predict Winners (and Which Do Not)

Across most DTC accounts, the metrics that correlate with eventual cost-per-purchase winners are, in rough order: outbound CTR, hold rate, then hook rate. Hook rate is the earliest signal and the least predictive on its own, because attention is necessary but not sufficient. Hold rate is a better predictor because it shows the message landed. Outbound CTR is the strongest early predictor because it shows intent. An ad that is above your account median on all three after 3,000 to 5,000 impressions is a strong candidate; one that is below on all three is almost always a loser.

Metrics that predict poorly on their own include link clicks (inflated by accidental taps on Reels), engagement (comments and shares reflect entertainment, not purchase intent), video average watch time without context (long videos always look better), and relevance diagnostics (too coarse to act on). Use them as supporting color, not verdicts. And never judge on cost per purchase under a minimum spend of 3 to 5 times your target; at low volume it is noise.

• Strong early predictors: outbound CTR, hold rate, hook rate (in that order).
• Supporting signals: ThruPlay rate, cost per 3-second view, saves and shares for UGC.
• Weak on their own: link clicks, total engagement, average watch time without length context.
• Lagging but final: cost per purchase after minimum spend, then 7-day trend of the same.

Static Ads: The Equivalent Metrics

Statics have no hook or hold rate because there is no play. The equivalent of hook rate is outbound CTR, since a static either earns a click or it does not, and the equivalent of hold rate is roughly the landing page view rate (landing page views divided by outbound clicks), which tells you whether the click was intentional and the page loaded. Cost per outbound click is a reasonable efficiency proxy at low volume, and the same minimum-spend rule applies before you read cost per purchase.

For statics, the diagnostic ladder is shorter. Low outbound CTR means the headline, image or offer is not stopping people, so change the headline first, the image second. Good CTR but low landing page view rate means accidental clicks or a slow page. Good CTR and good landing page views but high cost per purchase points at the page and offer, not the ad. Because statics are so cheap to iterate, the fix is usually to launch three new headlines on the same image the next day.

How Loraloop Fits

Loraloop's daily reporting is designed around the idea that creative problems should be caught on rate metrics, not weeks later on ROAS. Lora, the AI marketing lead, sends a morning briefing across your channels with recommendations, and Angie, the ads agent, runs daily optimisation on Meta such as shifting budget toward ads that are clearing their rate metrics and pausing those that are not, with every change waiting for your approval. When a diagnosis points to the opening, Angie can generate new hook variations and copy from your Brand DNA at 10 credits per ad so the fix is in next week's batch. The metrics setup in Ads Manager is still yours to do; this makes acting on them faster.

Frequently Asked Questions