How PublisherAtlas Estimates Website Revenue (Methodology)

Every revenue figure on PublisherAtlas is an estimate produced by a model. This page explains where the data comes from, how the model is constructed at a high level, and — just as importantly — what the numbers can and cannot tell you. If you make a decision based on our data, make it with this page in mind.

Where the data comes from

The directory covers publishers from four premium ad networks: Mediavine, Raptive (formerly AdThrive), Ezoic and Journey by Mediavine. Three kinds of sources feed each site's profile:

1. Official seller lists. Each network publicly publishes a sellers.json file listing the publishers it works with. If a network lists a publisher publicly, that membership appears in our directory. We never access any network's private systems, and we make no claims of endorsement.
2. Public registration records. Domain registration and expiry dates come from RDAP — the modern, public successor to WHOIS. These records are public by design.
3. Third-party traffic intelligence. Monthly visits, pageviews, engagement metrics, traffic sources, top countries and top keywords come from a licensed traffic-intelligence provider. Like all traffic estimates (Similarweb-class data), they are themselves modeled, with real margins of error.

How the revenue estimate is built

At a high level, the model starts from a site's estimated monthly pageviews, applies an earnings rate typical of the network the site belongs to, and adjusts for the geographic mix of the audience — because ad rates differ enormously between, say, US visitors and global averages. The result is an estimate of monthly display-ad revenue; the "estimated worth" figure is derived from that monthly revenue.

We deliberately don't publish the exact rate tables. They are the product of the directory's core work, and publishing them would let anyone reproduce the numbers — including people who'd rather not do the collecting. What we will say plainly:

• The model is transparent in structure: network membership × traffic × audience geography.
• It is calibrated against the ecosystem, not against any single site's reported numbers.
• It is refreshed monthly along with the underlying traffic data, and every site displays the data month it reflects (e.g. "Data as of 2026-09").

What the estimates CAN tell you

• Scale. Whether a niche or site plays in the hundreds, thousands, or hundreds of thousands of dollars per month. Our 30,060-site study shows the full distribution.
• Comparison. Relative standing between two sites or two niches — which is more monetized, which is rising, which has unusual economics.
• Structure. Why a site earns: traffic level, audience geography, network tier. A DR 2 site with 3.4M US visits can out-earn a DR 50 site with 80k visits — the model captures that.

What the estimates CANNOT tell you

• Actual earnings. We see no one's bank account. Two sites with identical profiles can earn different amounts due to ad layout, viewability, seasonality, direct deals, or floor prices.
• Non-display revenue. Affiliate income, sponsorships, products, and newsletters are invisible to the model. Many premium publishers earn as much again outside display ads.
• Short-term noise. Monthly intelligence data lags reality by weeks, and model outputs shouldn't be read to single-digit precision. "$4,213" means "roughly four thousand."
• Small-site precision. Estimates for sites under ~10k monthly visits carry wide error bars. Treat them as ranges, not figures.

How we handle corrections

If you own a site listed in the directory and believe an entry is factually wrong, contact support@publisheratlass.com. We review every request: factual errors get corrected; public registration facts cannot be altered, but context can be added where warranted.

Why we publish estimates at all

Because the alternative is worse. Today, "how much do niche sites make?" is answered by cherry-picked income reports and survivorship bias. A model applied uniformly across 30,060 sites — with its limitations stated — gives a comparable baseline, even if each individual number is approximate. Used for direction and comparison, it works well. Used as a valuation for a specific acquisition without due diligence, it doesn't — and we say so everywhere we show a dollar figure.

Ready to use it for direction and comparison? Browse the full directory free, or check a single site with the site checker.