Creative Metrics: Hook Rate, Hold Rate, and IC to Decide What to Scale
Learn to read hook rate, hold rate, and initiate checkout to decide which creative to scale even with a small test budget.
Which creative metric do you look at when there's no sale yet?
Look at initiate checkout. It's the signal that gets closest to a sale before the sale actually happens. If a creative has already generated IC, especially when the cost per IC lands around a third of the product's price point, it has a real shot at converting. Until that happens, you read hook rate and hold rate to figure out whether the creative holds attention or not.
The problem with deciding on sales alone is time. You launch 10 creatives and none of them have sold in the first few hours. Now what? Wait until you burn double the budget to be sure? No. You read the intermediate signals and decide before that.
Initiate checkout: the near-sale signal
Initiate checkout is the person who clicked the ad, saw the page, got convinced enough to start buying, and then stalled at the finish line. They got close.
The math is simple. If your cost per IC is coming in around a third of the price point, the creative has enough room to make the numbers work. A $297 ticket with IC coming in at $90 or $100 is workable. If your IC is coming in at the full price of the product, the numbers don't add up: the creative pulls clicks but not buyers.
IC isn't a sale. But it's the best guess you have before the sale lands. Use it as a filter to decide what keeps running and what gets cut.
Hook rate and hold rate: the attention metrics
These two tell you whether the video holds people or not. And they're different.
Hook rate is the percentage of people who stopped to watch the start of your creative. They scrolled past the feed, the video showed up, and they held their thumb instead of scrolling on. If it's at 75% or 80%, you nailed the opening. The hook landed.
Hold rate is the percentage that stayed from start to finish. They watched the whole body and didn't bail in the middle. This number shows whether the meat of the creative holds attention after the hook has done its job.
The two measure different things about the same video:
- High hook, low hold: the opening grabs them, but the body loses them along the way.
- Low hook, high hold: few people stop, but those who do watch to the end.
- Both high: a rock-solid creative, this is a candidate to scale.
- Both low: it didn't land. Cut it.
How to blend the opening and body of winning creatives
Here's the move most people don't make. When you have one creative with a high hook rate and a garbage hold rate, and another with a low hook and a high hold, you don't scrap either one. You combine them.
Take the opening from the creative with a 75% hook rate and stitch it onto the body of the creative that holds attention to the end. The opening that makes people stop, glued to the middle that makes people stay. A Frankenstein creative, and it works.
This is pure trial and error. You won't figure it out in the first week. You figure it out by studying the numbers across dozens of creatives and realizing each one nailed a different part. The job is combining the pieces that already proved they work.
Once you start blending, the number of variations explodes. One good opening tested against five different bodies is already five new creatives. Doing that one by one in Ads Manager turns into a lost night. That's the scenario for a platform like DirectAds to handle batch uploads of variations, pushing all the combinations at once instead of redoing the setup manually for every test.
Test volume and hit rate
The question everyone asks: if I launch 10 creatives, how many will hit?
There's no crystal ball. Nobody has one. What you can say is the average practice shows: out of 10 creatives, three hit. A 30% hit rate. The other seven don't land and you cut them.
But the picture changes once you understand what to do with the three that hit. You replicate them. You build variations on top of what already works. And then that same 30% hit rate starts turning into 80%, because you stop testing in the dark and start multiplying a validated pattern.
The creatives that tend to surprise you are the ones that don't look like ads. A question box, a recipe format, something disguised as organic content. Plenty of people launch the pretty, high-production video and it dies, while the ugly one that looks like a friend's post is the one generating IC.
How to buy data little by little without burning cash
You don't need a thousand bucks per creative to know if it's any good. Once the VSL has proven it sells, the bottleneck isn't the offer, it's the creative. And you test creatives by buying data little by little.
Here's how it works: launch 10 creatives with small variations, a controlled budget on each, and let the test volume print. You're not betting everything on one video. You're buying information in small doses until the numbers point to the winner.
The logic is to print volume. The more variations you run on a low budget, the faster hook rate, hold rate, and IC tell you where the gold is. Then you pull budget from the losers and dump it on the ones that showed a signal.
Launching 10 small variations at once, spread across accounts, is where the manual operation grinds to a halt. People running multiple BMs at this pace tend to lean on parallel duplication across accounts in an automation layer, because the work of setting up each variation by hand eats the time that should go into reading the numbers.
Takeaways
- Read initiate checkout before the sale: if your cost per IC is around a third of the ticket, the creative has a shot at closing.
- Separate hook rate (who stopped at the start) from hold rate (who stayed to the end). Different problems.
- Blend the opening from a high-hook creative with the body from a high-hold one. Frankenstein sells.
- Launch 10 variations on a low budget, cut the losers fast, and replicate the three that hit. A 30% hit rate turns into 80% when you multiply what works.
Frequently asked questions
What's a good hook rate on Meta Ads?
Above 70% already shows the opening grabs people. Below 30% the hook didn't land and there's no point pushing it. But the absolute number varies by niche; what matters is comparing your creatives against each other.
Can initiate checkout help me decide without a sale?
It works as the closest signal to a sale. If your cost per IC is coming in around a third of the ticket, the creative has room to convert. If IC comes in expensive, as much as the full product, the numbers don't add up.
Do I need to spend a lot per creative to test?
No. Once the VSL is already selling, the test is about the creative and you buy data little by little. A low budget across many variations prints volume faster than a high budget on a few videos.
How many creatives usually hit out of every 10?
The practical average is around three out of ten, a 30% hit rate. The trick is replicating those three until the hit rate climbs, because you stop testing in the dark and start multiplying a validated pattern.
