How this comparison is put together

Last reviewed 29 July 2026

Every figure in the popunder ad network comparison falls into one of two buckets before it is written down: verifiable against a document, or claimed by the party that benefits from the claim. Both appear on the site. They are never presented the same way.

Source hierarchy

Sources are ranked, and a lower rank never overrides a higher one.

1. A company's own documentation about itself. Its pricing page, terms, help centre, or a product dossier it supplied. For policy questions, the policy text itself. This rank never applies to what one company says about a competitor.
2. Dated third-party datasets. Measurement firms and research houses that publish a sample size and a collection period. Fraudlogix invalid-traffic data and GWI ad-blocker figures both qualify; each is cited on the comparison with the period it covers.
3. Independent testing published by someone else. Usable when the tester says what they spent and when. Cited as testing, not as a rate card.
4. Everything else. Round-up posts, affiliate reviews, undated blog claims. These do not establish a fact. At most they point at something worth checking properly.

Figures that only ever appear at rank four do not go on the site. That rule alone removes most of what circulates in this category, including several statistics examined on the about page.

Contractual figures versus marketing figures

Deposit minimums, payout floors and payment schedules are contractual. A publisher sees them on a withdrawal screen, an advertiser sees them on a funding screen, and neither can be quietly restated. These are the numbers worth planning a budget around, and they carry the most weight in the comparison table.

Impression counts, click counts, GEO counts and advertiser counts are marketing. No network in this category submits to MRC accreditation for its volume claims, so there is nothing to check them against. They appear on the site because readers ask about scale, and they are labelled self-reported every time.

The arithmetic check on volume claims

A network that publishes bid requests, impressions and clicks has also published two derived figures whether it meant to or not. Divide impressions by bid requests and you have a fill rate. Divide clicks by impressions and you have a click-through rate. Both land in a knowable range for real exchange traffic.

Worked through on the comparison page for one network: 30 billion daily bid requests producing 900 million impressions implies roughly 3% fill, which is ordinary for an exchange bidding into open supply. Those 900 million impressions producing 30 million clicks implies about 3.3% CTR, plausible for pop and absurd for display banners. The claims are internally consistent, which is not proof they are true but does mean nobody fabricated them carelessly.

Run the same two divisions on any network you are considering. A published fill rate above 30%, or a click-through rate that would embarrass a push campaign, tells you what the rest of that page is worth. The check costs thirty seconds. It is the single most useful thing on this page, and it is not original — it is just arithmetic nobody bothers to do.

When sources disagree

They do, more often than the tidy tables suggest, and the current comparison carries one unresolved case worth spelling out. A product dossier supplied to this site lists the HilltopAds advertiser minimum at $50. HilltopAds' own pricing page states $100.

The dossier came from a competing network. Under the hierarchy above that puts it at rank four for any claim about HilltopAds rather than rank one, because rank one only covers what a company documents about itself. The comparison therefore carries $100 and flags the disagreement.

The general rule: a party's own current page wins, the disagreement gets disclosed, and the reader is told which figure to verify before committing money.

What this site does not do

No campaigns have been run through any of these networks. Every CPM figure comes from published benchmarks or from testing done and published by others, always attributed. There is no first-person performance data here, and inventing some would be the fastest route to making the rest untrustworthy. The author page states the same rule from the other direction.

Nobody pays for placement or for a favourable position on this site. Which links are commercial, and what those links do not buy, is set out on the advertising disclosure page.

Updates and corrections

Terms in this category move roughly twice a year. Minimums shift, payment schedules change, GEO counts get rounded upward, and networks rarely announce any of it. Every table on the comparison carries the month it was checked, and the page-level modified date changes when a figure changes rather than on a schedule designed to look fresh.

Corrections are welcome and get handled properly — the figure is fixed, the change is dated, and where the error affected a conclusion the conclusion changes too. Send them through the contact page with the source you are relying on.