What is cost per result in Meta Ads?
Cost per result in Meta Ads is spend divided by whatever event the ad set optimizes for. Here is why that makes two rows in one column hard to compare.
Cost per result in Meta Ads is spend divided by whatever event the ad set optimizes for. Here is why that makes two rows in one column hard to compare.
A client once forwarded me a screenshot with one line of text: "Why are we paying €38 here when this one is €4?" Both rows sat in the same Ads Manager column, both labelled cost per result. The €4 ad set was optimizing for landing page views. The €38 one was optimizing for purchases.
Same column header. Two different questions.
Short answer: Cost per result in Meta Ads is spend divided by the number of results, and a result is whatever event that ad set optimizes for: a purchase, a lead, a landing page view, a ThruPlay. Two ad sets with different optimization events report cost per result in different units, so the column only compares like with like.
The takeaways
A result in Meta Ads is the event the ad set was told to optimize for. Meta's help center describes results as the number of times your ad achieved an outcome based on the objective and settings you selected. So the result on a sales ad set set to purchases is a purchase. On a traffic ad set it is a link click or a landing page view. On a video views ad set it is a ThruPlay.
Cost per result is then plain division: amount spent over that count.
The trouble is the label. Ads Manager leads with "Results" and "Cost per result" in its default views, and the header never tells you which event sits underneath. The small grey text under each number does, and that line is the first thing I read on any account I inherit. A cost per result with no event named next to it is half a number.
Cost per result changes meaning whenever someone edits the ad set's optimization event, because the column follows the event and keeps the same name. Move an ad set from optimizing for add to cart to optimizing for purchases, and its cost per result will jump. Nothing got worse. The metric now counts a rarer, later event.
I have seen this cause a panic. Someone switched the conversion event on a Friday, Monday's report showed cost per result up sharply, and the first instinct was to kill the creative.
The fix is boring. Note every optimization change in the report itself, and never draw one trend line across the edit. The numbers before and after are two series that happen to share a column. Expect a learning-phase reset too: Meta's documentation lists changing the optimization event as a significant edit, so early numbers after the switch are noisy anyway.
You can compare cost per result across ad sets only when every row optimizes for the same event under the same attribution setting. That condition fails more often than buyers assume, because an account grows ad sets one at a time and nobody audits the mix.
Mixed rows are where the €4 versus €38 conversation comes from. Scanning down the column, the eye does arithmetic the column was never built for. The page-view ad set looks ten times cheaper while possibly producing zero sales.
Before comparing, filter or group by optimization event. Check your summary row too: a total across mismatched events would mean nothing, so treat any blended figure there with suspicion (that row's label has changed over the years, so look at your own account).
If two ad sets must be ranked against each other, rank them on a shared column you added yourself.
Replace cost per result with an explicit column for the event that pays the bills: cost per purchase for ecommerce, cost per lead for lead gen, cost per install for apps. You add these through the column customizer, and they mean the same thing on every row regardless of what the ad set optimizes for.
That explicit column also catches a quiet failure. An ad set optimizing for landing page views can show a lovely cost per result while its cost per purchase is empty, because the people Meta found to click are not the people who buy.
So the working rule I use: cost per result is fine for glancing at delivery health inside one ad set. Any scale, kill or budget move gets made on the explicit column. The scale decision is where a wrong unit costs real money, which I covered from the other side in the Meta ads metrics that matter.
A cost per result inherits every caveat of Meta's conversion counting, so even on a purchase ad set it is a modeled count. The result is credited under the account's attribution window, it can still move for days as delayed conversions arrive, and it will not match your store's order count exactly.
That makes cost per purchase in Ads Manager a strong signal and a weak ledger. For a scale call, trust its direction over a few days. For a finance conversation, pull the orders from your checkout. The gap between the two is its own topic, and I wrote it up in why Meta conversions don't match your sales.
This mismatch is one reason Adscalr does not rank creatives on cost per result at all. Its composite score is built from 6 named metrics (hook rate, CTR, CPI, ROAS, share rate, revenue per install), and the weights are adjustable per project and funnel stage, so a top-of-funnel test and a purchase test are not scored on one hidden unit. The rest of how tests get read lives on the ad intelligence page.
This is the thinking behind Adscalr.
See the product →