--- title: "What 18,920 Meta ads say about cost per lead" description: "We graded every Meta ad we ran for gyms, studios, med spas and salons. The popular formats are not the cheap ones, and Meta's favorite ad is often not yours." section: Benchmarks author: David Steel published: 2026-10-10T18:16:58.562Z url: https://one.sneeze.it/blog/what-18920-meta-ads-say-about-cost-per-lead tags: ["meta-ads", "creative", "cost-per-lead", "creative-benchmarks"] --- # What 18,920 Meta ads say about cost per lead _We graded every Meta ad we ran for gyms, studios, med spas and salons. The popular formats are not the cheap ones, and Meta's favorite ad is often not yours._ Most ads you run will not work. That is not a sign your ads are bad. It is how Meta advertising works, and the data shows it plainly. We pulled every Meta ad Sneeze It has run for gyms, studios, med spas, salons, wellness clinics and franchise brands, and asked one question of each: did it do the work? :::stats $3.95M | Meta spend analyzed 18,920 | ads graded 121,333 | leads tracked 6.2% | became winners source: [Creative Benchmarks](https://benchmark.sneeze.it), Meta insights across 137 accounts and 41 brands ::: ## Most ads barely run Two out of three ads spent less than $100 in their whole life. Only 6.2% became winners, which we define as an ad that spent at least ten times its account's median ad and at least $500. Meta tries an ad, decides quickly, and moves the money elsewhere. So the job is not to predict which ad will win. Nobody does that reliably. The job is to find winners fast, understand why they worked, and make more like them. ## Meta's favorite ad is often not yours Of the 1,175 ads Meta put the most money behind, fewer than half brought in leads cheaper than their own account's average. :::chart dots "Out of every 100 winning ads, how many beat their account's average cost per lead" 47 | 100 source: Creative Benchmarks ::: :::pullquote Spend tells you what Meta likes. Leads tell you what works. ::: That is why we grade every ad twice: once on how much Meta backed it, and again on what a lead from it actually cost. An ad has to pass both before we call it a winner. ## The popular formats are not the cheap ones We labeled 2,126 ads by how they look, how they open and how they were made, then compared what a lead cost from each. The average across all of them was $26.62. :::chart bar "Cost per lead by how the ad looks" unit=$ highlight="Before & after,Product demo,Problem, then fix" Before & after | 18.73 | $18.73 Product demo | 20.58 | $20.58 Problem, then fix | 21.73 | $21.73 Cinematic footage | 24.51 | $24.51 Offer-first banner | 25.52 | $25.52 Photo collage | 27.75 | $27.75 Feature and benefit list | 32.22 | $32.22 Headline only | 34.35 | $34.35 source: Creative Benchmarks, 2,126 labeled ads across 22 brands ::: Headline-only designs are one of the most common ads we see, and they cost 83% more per lead than before-and-after ads. How an ad opens matters even more. Ads that opened with a contrarian take or a bold claim beat the average by $6 to $12 a lead. Announcements cost almost three times the average. :::chart bar "Cost per lead by how the ad opens" unit=$ highlight="Contrarian take,Bold claim" Contrarian take | 14.30 | $14.30 Bold claim | 20.16 | $20.16 Curiosity | 23.43 | $23.43 Offer only | 26.30 | $26.30 Talking to the viewer | 27.01 | $27.01 Authority | 33.63 | $33.63 Relatable moment | 37.93 | $37.93 Announcement | 73.48 | $73.48 source: Creative Benchmarks ::: ## Polish is not the point Phone-shot UGC brought the cheapest leads of any production style, $22.21, and became a winner 15.5% of the time. That is more than twice the rate of polished, high production ads, which won 7.0% of the time. :::compare "High production" "Phone-shot UGC" **$25.50** per lead. Became a winner **7.0%** of the time. 784 ads. --- **$22.21** per lead. Became a winner **15.5%** of the time. 64 ads. ::: :::callout note "A fair caution" Some of these groups are small. 64 UGC ads is enough to notice a pattern, not enough to bet a whole budget on it. Treat these as the places to test first, not as rules. ::: :::takeaways - Expect most ads to fail. Launch more of them, and judge them fast. - Do not trust spend alone. Check every winner against its account's cost per lead. - Test before-and-after, product demos and problem-then-fix before another headline-only ad. - Open with a contrarian take or a bold claim. Avoid announcements. - Shoot some on a phone. ::: :::cta button="See the full benchmarks" href="https://benchmark.sneeze.it" How do your ads stack up? We line your current ads up against these benchmarks and email you the report. No login, no call, no cost. :::