Home/Benchmarks

Real advertising costs, from accounts I actually managed.

First-party Meta Ads and SEO figures with the reporting behind them, not an aggregate of other people's surveys.

Most published advertising benchmarks are aggregates of aggregates. These are not. Every figure below comes from an advertising or search account I personally managed, and every one links to a case study showing the reporting it was read from.

The benchmark sets

Other documented results

These don't yet have enough comparable accounts behind them to form a benchmark set, so they live as individual case studies rather than as data pages. Listed here so the full record is in one place.

ResultChannelClient & marketSource
121 purchases at $0.71 average, best product $0.32Facebook conversion adsFashion e-commerce, BangladeshCase study
333,110 Reel views, 96% non-followers, 40,000 Tk extra outlet sales in 2 daysOrganic creativeFashion e-commerce, BangladeshCase study
Page 1 in three months from zero organic traffic, +75% traffic, 516 clicksMedical SEOClinic, SingaporeCase study
#1 for "takaful insurance in canada" and "shariah insurance in canada"SEOIslamic insurance fintech, CanadaCase study
Page 1 for "AI in home farming" from a standing startSEOAgriTech platform, TorontoCase study

How these figures are produced

  • Read directly from Meta Ads Manager, Google Search Console or Google Analytics on accounts I managed.
  • Nothing is modelled, estimated or extrapolated. Where a campaign is still running and a final figure would mislead, the table says so rather than dividing two numbers.
  • Each figure links to the case study carrying the screenshot it came from. Two clients are anonymous at their request; that is noted on their pages.
  • Markets are never blended. US lead-generation costs and Bangladeshi conversational-commerce costs live on separate pages because averaging them would produce a number that is true of nothing.

If you find a figure on this site you can't trace to a screenshot, tell me and I'll either produce it or remove the claim.

Last reviewed . Published under CC BY 4.0: reuse with attribution.