Strong data
Large-sample industry reports and association data: WordStream/LocaliQ, Gartner, The CMO Survey, BrightLocal, SearchLight, NADA, NPMA and peers. Citable as-is.
/benchmarks/methodology · BENCHMARK LIBRARY
How every number is sourced and verified.
Marketing benchmarks are usually recycled blog numbers with no source attached. Ours work differently: a stat cannot enter this library without a named source, a data period, and a confidence label — and the whole library is re-verified every year. Every 2026 figure shows its last-verified date on the page.
No website yet? Book a 30-minute conversation
Confidence labels
Four tiers of trust, labeled on every stat.
Directional
Published by a single specialist agency or vendor. Useful for orientation, but not independently audited — treat as a range, not a target.
Limited data
No reliable vertical-specific public data exists. We say so plainly and show the closest category-level anchor instead of inventing a number.
Get X client data
Anonymized aggregates from real client accounts we manage — never individual clients, never named. Where public data is thin, this is the only real number on the internet.
Read the fine print
What these numbers can and can't tell you.
“Averages” are medians
The big Google Ads benchmark reports publish medians to limit outlier skew. A business can beat the “average” and still sit in the bottom half of performers.
Geography swings everything
Metro click costs run 2–4x rural. Figures are US-market data shown in CAD (converted from USD) — Canadian markets typically price below comparable US metros.
Recency varies by source
Google Ads data refreshes annually; association studies lag a year or more. Every stat shows its data period, and anything older than 13 months gets flagged at refresh time.
Guardrails, not targets
The decisive metric in every vertical is cost per sale and LTV-to-CAC — a $228 roofing lead against a $12,000 job is excellent; a $30 lead that never closes is not.
