The Short Answer
We looked at 765 Meta ads from 18 brands, with $10.9 million in spend, to see which metrics actually move cost per purchase. Within each brand:
- The share of clicks that turned into purchases explained about 85% of the difference in cost per purchase between ads. The quarter of ads with the best post-click conversion rate had half the cost per purchase of the worst quarter, in 17 of 18 brands.
- CPM explained about 20%. Cheaper impressions helped, a bit.
- CTR explained close to none. High-CTR ads attracted clicks that converted worse and paid more to reach people. The two canceled out.
- Hook rate, hold rate and watch time didn't predict cheaper purchases. Ads with higher numbers on all three actually had lower ROAS in most brands.
- Add-to-cart rate did. The quarter of ads with the highest add-to-cart rate had about a third lower cost per purchase.
The takeaway: the cheapest purchases came from ads that sent people who were already close to buying. Most of that work happens in the ad, before anyone sees the landing page.
The Data
We pulled ad-level data from the US Meta accounts we have visibility into as a creative agency, covering July 21 to October 20, 2026. We kept every brand that spent at least $100,000 and tracked purchases or leads reliably, then looked at each brand's top 50 ads by spend.
To keep cost per result stable, we only included ads with at least $1,000 in spend and at least 10 purchases or leads. That left 765 ads from 18 brands, with $10.9 million in spend: consumer brands plus apps, subscriptions and lead-generation businesses.
Every comparison is made within a brand first, so a cheap-to-sell product can't make one kind of ad look better than it is. Then we took the median across brands. Brand names are anonymized. Part of our Meta Ads Statistics report.
Cost Per Purchase Has Three Parts
Every ad's cost per purchase comes from three numbers:
Cost per purchase = (CPM ÷ 1,000) ÷ CTR ÷ conversion rate
- CPM: what you pay for 1,000 impressions
- CTR: the share of impressions that become link clicks
- Conversion rate: the share of those clicks that become purchases or leads
For example, a $20 CPM is $0.02 per impression. At a 1% CTR, that's $2 per click. At a 2% conversion rate, that's $100 per purchase.
So which of the three explained the difference between a brand's cheap ads and its expensive ones?
| Part | Share of the difference in cost per purchase | Cheapest quarter vs. most expensive quarter |
|---|---|---|
| Conversion rate after the click | about 85% | 0.51x cost per purchase, in 17 of 18 brands |
| CPM | about 20% | High-CPM ads: 1.45x cost per purchase |
| CTR | close to 0% | High-CTR ads: 0.99x |
The shares overlap a little, which is why they add to more than 100%. But the picture is clear. The ads with the cheapest purchases weren't the ones people clicked most. They were the ones whose clicks bought.
The Click Isn't the Goal
CTR is the metric most people use to judge an ad. In this data it told you almost nothing about cost per purchase, for two reasons:
- High-CTR ads attracted clicks that converted worse. CTR and post-click conversion rate moved in opposite directions in 15 of 18 brands.
- High-CTR ads cost more to show. They paid a higher CPM in all 18 brands.
An ad that makes lots of people curious gets lots of clicks from people who weren't going to buy. Meta charges you for reaching them, and they leave. We cover this in more detail in our Facebook ads CTR benchmarks.
Hook Rate and Watch Time Can Work Against You
The same thing happened with video engagement. Comparing each brand's video ads:
| Metric | Top quarter vs. bottom quarter cost per purchase | Relationship with ROAS |
|---|---|---|
| Hook rate (3-second plays ÷ impressions) | 1.15x (cheaper in only 6 of 18 brands) | Negative in 12 of 14 brands |
| Hold rate (ThruPlays ÷ impressions) | 1.00x | Negative in 11 of 14 brands |
| Average watch time | 1.11x | Negative in 10 of 14 brands |
| 15-second retention | 1.02x | No clear pattern |
Ads that held more attention also paid more for it: hook rate and CPM rose together in 17 of 18 brands. And that attention didn't come from people more likely to buy.
None of this means you should make boring ads. A weak opening still loses people, and we use hook rate to spot that. It means attention is only useful when it's attention from the right people. A hook that stops everyone is less valuable than a hook that stops buyers.
It's Not Just the Landing Page's Job
When conversion rate is the problem, most teams look at the landing page. Faster load times, better copy, a cleaner checkout. Those matter. But this data says the ad does a lot of the converting before anyone arrives.
Look at what changes and what doesn't from one ad to the next within a brand. The product, the price, the brand and the website are mostly the same. Yet:
- Post-click conversion rate varied almost 4x between a brand's better and worse ads (90th percentile vs. 10th percentile).
- Whether clicks reached the page wasn't the issue. The share of clicks that turned into a landing page view had no relationship with cost per purchase (1.10x).
- Among people who did reach the page, the conversion rate still varied by ad, and it tracked cost per purchase closely. The best quarter of ads had about half the cost per purchase of the worst (0.53x).
Same store, same shelf, same price. The difference was who walked in, and how ready they were when they did. That's decided by the ad.
An ad that pre-sells does part of the landing page's job in advance. By the time someone clicks, they already know what the product is, who it's for, why it's better than what they use now, and roughly what it costs. The landing page just has to confirm it. An ad that only grabs attention sends people who still need to be sold from scratch, and most of them won't be.
What Pre-Selling Looks Like
To see which kinds of ads pre-sell best, we matched 353 of these ads to Atria's creative tags for theme, emotion, opening and offer. Comparing ads with and without each tag inside the same brand:
| Ads that... | Cost per purchase vs. the rest | Brands where it was cheaper |
|---|---|---|
| Explained why the product is different | 0.85x, with 1.6x the conversion rate | 6 of 8 |
| Compared it to the alternatives | 0.85x | 5 of 7 |
| Spelled out features and benefits | 0.87x | 9 of 13 |
| Spoke to a desire for certainty (proof, guarantees, "it works") | 0.82x | 7 of 9 |
| Led with a percentage-off offer | 0.89x, with 1.35x the conversion rate | 4 of 6 |
| Opened on a relatable statement | 1.23x | 3 of 12 |
These effects are smaller than the conversion rate gap itself, and they vary by brand. But they point the same way. The ads that converted cheaper gave people a reason to choose this product, not just a moment they related to.
In practice, an ad that pre-sells usually does most of these:
- Says who it's for. A specific person with a specific problem, so the wrong people scroll past instead of clicking.
- Explains why it's different. What it does that the thing they use now doesn't.
- Compares. Against the old way, the cheaper option or the competitor they're already considering.
- Shows proof. A demo, a result, a review, a guarantee. Something that makes the claim believable before they click.
- Is upfront about the offer. The price, the discount or the trial. Surprises on the landing page kill conversion.
- Matches the page. The landing page should continue the exact story the ad started, with the same claim and the same offer.
For example, in one of our Huel ads, the opening line is "Not another protein shake," followed by stat cards for protein, vitamins and fiber, then a new-customer offer. In about 14 seconds it says what's different, proves it and states the offer. Someone who clicks already knows why they're there. More examples in how to build a creative strategy.
What the Landing Page Still Has to Do
Pre-selling doesn't make the landing page unimportant. It changes its job. A good page:
- Continues the ad's story. Same headline claim, same offer, same visuals where you can.
- Answers the next question. The ad made the case. The page handles the details: sizes, shipping, ingredients, how it works.
- Makes buying easy. Fast load, clear price, few steps.
When the ad and the page work together, the page converts people the ad already convinced. When they don't, the page has to sell a stranger from scratch.
Which Metrics to Watch Instead
If CTR and hook rate don't pick winners, what should you look at?
- Cost per purchase or lead. Still the deciding number.
- Post-click conversion rate. Purchases ÷ link clicks, or ÷ landing page views. If an ad's clicks don't convert, the ad isn't pre-selling, or it's attracting the wrong people.
- Add-to-cart rate. The quarter of ads with the highest add-to-cart rate had 0.68x the cost per purchase, in 9 of 12 brands. Add-to-carts come in faster and in larger numbers than purchases, so they're a useful early read on a new ad.
- CPM. Expensive impressions raise cost per purchase. Watch for ads that win attention but pay too much for it.
- CTR, hook rate and hold rate as diagnostics. Useful for explaining why an ad underperforms. Not for picking winners.
What This Means for Your Ads
- Judge ads on what happens after the click. A high CTR with a low conversion rate is an expensive ad.
- Write ads that sell, not just ads that stop the scroll. The opening earns attention. The rest of the ad has to give people a reason to buy.
- Test messages, not just hooks. A new reason to choose the product, for a new kind of buyer, does more than a new first line. See creative diversity on Meta.
- Make the ad and the page one story. When conversion is low, check the ad before you redesign the page.
Get Ads That Sell Before the Click
Biddyco is a creative agency for Meta ads for brands spending $100K to $5M+/month. Our US-based senior team makes UGC, motion, hybrid and static ads built around real reasons to buy, and judges them on cost per purchase, not clicks or views. Book an intro call to talk through your account.
FAQ
What metrics predict cost per purchase on Meta?
In our data, post-click conversion rate explained about 85% of the difference in cost per purchase between ads in the same brand. CPM explained about 20%, and CTR close to none. Add-to-cart rate was also a strong early signal.
Does a higher CTR lower cost per purchase?
Not in our data. High-CTR ads attracted clicks that converted worse and paid a higher CPM, so their cost per purchase ended up about the same as low-CTR ads (0.99x).
Is hook rate a good predictor of ad performance?
No. Ads with the highest hook rates had 1.15x the cost per purchase of the lowest, and lower ROAS in 12 of 14 brands. Hook rate is useful for spotting weak openings, not for picking winners.
Why is my landing page conversion rate low?
Often it's the ad, not the page. Within the same brand and the same site, post-click conversion rate varied almost 4x between ads. Ads that explain what the product is, who it's for and why it's different send visitors who are ready to buy.
What does it mean for an ad to pre-sell?
A pre-selling ad does part of the landing page's job before the click. It says who the product is for, why it's different, shows proof and is upfront about the offer, so people arrive already convinced.
What should I measure instead of CTR?
Cost per purchase or lead first, then post-click conversion rate and add-to-cart rate. Use CTR, hook rate and hold rate to diagnose why an ad underperforms.
Where does this data come from?
Ad-level data from US Meta accounts we have visibility into as a creative agency, from July 21 to October 20, 2026: 765 ads from 18 brands spending at least $100,000, each with at least $1,000 in spend and 10 purchases or leads. Creative themes come from Atria's tags on 353 of those ads. Brand names are anonymized.
