CPM Benchmark Ranges

Directional ranges to sanity-check your own CPM/fill-rate assumptions in the ad revenue calculator — not a precise lookup for your specific site.

These are order-of-magnitude ranges, not quotes, and cover DISPLAY ads specifically — video and in-app inventory command different rates. Your real CPM also depends on your specific advertisers, traffic quality, season, and viewability — none of which a static table can capture. If you have real numbers from your own ad network dashboard, trust those over anything here.

By content niche

TierExamplesTypical CPM (USD)Note
Premium / high advertiser LTVFinance & investing, B2B/SaaS, insurance, legal$8 – $25Advertisers in these verticals have a high customer lifetime value and bid aggressively for qualified traffic — consistently reported as the highest-paying content categories for open-exchange DISPLAY ads. A 2026-09-10 domain-expert review found this band's original $15-$50+ figure implied implausibly high page RPMs when run through this project's own formula, and was likely sourced from RPM/video-CPM/direct-deal quotes rather than per-impression display CPM — lowered accordingly. CPMs well above $25 do exist in this vertical (video, newsletter sponsorships, direct-sold placements) but aren't representative of ordinary open-exchange display inventory.
Mid-rangeTech, home & DIY, food, personal finance-adjacent lifestyle, health$5 – $15The broad middle of the market — decent advertiser demand, not a specialty vertical.
Broad / general interestGeneral news, entertainment, gaming, broad lifestyle/aggregator content$1 – $5High traffic volume but lower advertiser willingness-to-pay per impression is a well-known, long-standing pattern in this tier — large audiences do not automatically mean high CPM.

By traffic geography

TierExamplesRelative to Tier 1Note
Tier 1US, UK, Canada, Australia, and similar high-GDP English-speaking marketsBaselineThe great majority of premium display/programmatic ad budgets target these markets — use this tier's own niche CPM band above directly.
Tier 2Western/Northern Europe (non-UK), Japan, South Korea, and similar developed marketsRoughly a third to two-thirds of Tier 1, though this tier is the least clean-cutStill strong advertiser demand, but a smaller pool than Tier 1 and often split across more languages/currencies. Where a market actually falls is genuinely contested — some of the largest Western European markets (Germany, the Netherlands, Switzerland, the Nordics) sit closer to Tier 1 in many advertiser classifications, while smaller-language markets fall further below it. Treat this tier as a rough middle ground, not a precise band.
Tier 3Most of the rest of the world by ad-buyer volumeOften a tenth or less of Tier 1, sometimes far lowerThe single biggest lever on realized CPM for most sites is traffic geography, not content niche — a Tier-1-majority audience can be worth several times a Tier-3-majority audience for otherwise identical content. Ranges here are intentionally loose: the actual gap reported across sources varies enormously by ad network and vertical.

Typical fill rate by setup

SetupTypical fill rateNote
Header bidding (multiple demand sources competing)85% – 98%Generally the highest and most reliable fill rate setup for an established site — some sources report the top end reaching 97-100%.
Single ad network / single-SSP open exchange60% – 100%Sources genuinely disagree here — a 2026-09-10 domain-expert review found some ad-tech sources call 60-85% "good" (partly because deliberately-set price floors leave some inventory unfilled on purpose), while others say a healthy AdSense/AdX setup should reach 95-100% and that anything under 90% signals a problem. Check your own network's dashboard for your real number (AdSense calls this "coverage") rather than trusting either end of this range.
New or very low-traffic site40% – 70%Fill rate is generally lower before an ad network has enough data/history on a site's traffic to bid confidently — expect it to improve over the first few months.

Full sourcing notes, confidence levels, and citations are in this project’s lib/cpm-benchmarks.ts and docs/domain-reference.md.

Frequently asked questions

Why are these ranges so wide?
Real CPM varies enormously by exact advertiser demand at the moment, seasonality, ad format, device, viewability, and dozens of other factors no static table can capture — and precise-sounding per-niche numbers circulating online are often unsourced or inflated. These ranges are deliberately wide and rounded so they stay honest about that uncertainty rather than implying false precision.
Which matters more for my CPM: my content niche or my traffic's geography?
Geography is usually the bigger lever. A Tier-1-majority audience (US/UK/Canada/Australia and similar) can be worth several times a Tier-3-majority audience for otherwise identical content — often a larger swing than moving between niche tiers. If your analytics show most of your traffic comes from outside Tier 1, expect your realized CPM to sit well below a niche band's stated range even in a high-value niche.
Where do these numbers come from?
Directional pattern (which niches and geographies pay more, and roughly why) is well-established, uncontested ad-tech knowledge. Specific dollar figures were retrieved via WebSearch synthesis (WebFetch was unavailable this session) and deliberately widened/rounded rather than trusted at face value — several sources found during research returned suspiciously precise, unsourced numbers that this project chose not to present as fact. A domain-expert review also checked these figures against the project's own revenue formula and adjusted one band that implied unrealistic results — see this project's own docs/domain-reference.md for the full sourcing note.
Do these ranges apply to video or in-app ads too?
No — these are display-ad ranges, matching what this project's calculators assume. Video ad CPMs commonly run several times higher than display for a comparable audience, and native ad CPMs sit in between. If most of your inventory is video or in-app, treat these bands as a poor fit rather than adjusting your display assumptions to match them.