Last updated: August 5, 2026

How Deposit Minimums and Bid Floors Are Checked

Every number on this site goes through the same routine before it is published, and the routine is written out here so a reader can repeat it rather than trust it. It covers four things: where platform figures are found, how payment-method exceptions surface, how prices are tested against each other, and how the remaining uncertainty is labelled. It is the routine behind every number on the Cheap Website Traffic page, and none of it requires access to anything private. The sources are operator pages, industry standards and arithmetic, all of them open. A figure you can re-derive is worth more than a figure you have to believe.

Step 1: find the operator's own statement

Deposit minimums live in terms and conditions or in help centre articles, not in marketing copy, and the two frequently disagree on the same domain. In the current comparison set, one platform states its minimum in its terms, two publish theirs in help centre articles, and two publish theirs on their own blogs. Whichever document carries the figure, it must sit on the operator's domain to count as primary. Marketing pages are read for context, and their claims are labelled as claims when they appear in the text.

Step 2: check the payment-method floor, not just the headline

A headline minimum is almost always the card minimum, and the interesting variation sits underneath it. Wire transfers routinely carry floors ten to twenty times higher, digital currencies often sit in between, and one platform in the current set denominates its wire minimum in euros rather than dollars, which quietly changes the number for a US buyer. The headline figure goes into the comparison table, since it determines whether a first campaign is possible at all. Every exception found goes into the body text. Confirm the floor for your own payment method before transferring: these are the figures that change most often, and they change quietly.

Step 3: separate bid floors from clearing prices

A floor is the lowest bid a platform's system will accept. A clearing price is what a campaign has to bid to win impressions in a given country and format. In competitive markets those numbers differ by an order of magnitude. Quoting one as the other is the most common error in this category. Published floors are recorded as floors, with the platform named. Clearing prices are given as ranges, since they move weekly and depend on vertical, device and time of day. Where a platform shows live bids only inside its dashboard, the page says so instead of substituting a stale public figure.

Step 4: cross-check directories to surface contradictions

Affiliate catalogues are read after the operator pages, and only to find disagreements worth reporting. The last pass produced two, running in opposite directions: a platform listed at $50 across catalogues whose own help centre says $100, and a platform listed at $200 in several 2026 reviews whose own terms say $100. Both go into the text with the sources named. For a buyer, the discrepancy is worth more than either figure alone. The same check runs against every platform in the set, including the one this site recommends. A method applied selectively measures nothing.

Step 5: run the arithmetic before trusting a price

Prices are checked against each other, not just against sources. The core test compares what a seller charges per thousand sessions with what a publisher is paid per thousand visits in the country the traffic is claimed to come from, then adds the network's margin, which no platform publishes. A seller quoting less than the publisher payout for the same country is either sourcing traffic somewhere other than where it advertises, or not sourcing humans at all. The test only works inside a single country. Cheap Tier 3 inventory and a suspiciously cheap "United States" session are different phenomena, and a global average flattens both into a meaningless middle.

Reported volume figures get the same treatment through ratios. A platform reporting billions of bid requests, hundreds of millions of impressions and tens of millions of clicks is really reporting a fill rate and a click-through rate; converting the headline into those two ratios shows immediately whether the numbers describe a plausible format mix or a press release. Company-reported figures are labelled as such wherever they appear.

Step 6: check quality claims only as far as they can be checked

Claims about traffic quality are the hardest category to verify from outside, so the page states what can be established and what cannot. Two checks are available to any buyer with server or CDN logs: autonomous system numbers, which expose datacentre sessions wearing a browser, and IP frequency distribution, which exposes residential proxy pools that pass the first check cleanly because they present consumer ISP addresses. Neither check runs in a standard analytics interface, which never exposes the addresses in the first place. Industry definitions for what counts as invalid come from the IAB and MRC invalid traffic standards, and the IAB Tech Lab spiders and bots list is the reference used at platform scale for declared crawlers.

Vendor-side filtering claims are reported as claims. A platform saying it excludes proxy connections has documented a feature. Whether the exclusion works is a separate question, and the text keeps the two apart.

Step 7: label what survives

Every published figure carries one of three statuses, visible in the sentence rather than in metadata. Confirmed means it appears on the operator's own page, and the document type is named. Company-reported means the operator publishes it about itself with no independent audit, which covers most volume and traffic figures in this industry. Unconfirmed means it appears only in secondary sources and could not be found on a primary page; the current example is a $0.0003 push floor that sits an order of magnitude below every comparable figure in the category and is flagged wherever it appears. A figure that fits none of the three labels stays out.

What a reader can repeat in ten minutes

Three of these steps need no tooling at all, which is the reason they are the backbone of the method. Open the platform's help centre and search for the word deposit: if the figure there differs from the one in a directory listing, the directory is stale and the operator page wins. Take any price quoted per thousand sessions and set it beside a publisher payout for the same country: if the seller is charging less than the publisher earns, the arithmetic has already answered the question about traffic quality. Convert any reported volume claim into a fill rate and a click-through rate: if either ratio is implausible for the formats involved, the headline was written for effect rather than for accuracy.

Review schedule and what is never published

Pages are re-verified quarterly and immediately on three triggers: a reader reports a change, a platform announces new terms, or a figure that feeds a calculation elsewhere on the site moves. Reports arrive by email and are kept for twelve months so a corrected figure can be traced back to its source; what else is stored, and for how long, is set out in the privacy policy. Verification dates shown on each page refer to the last full pass, not to the day the text was written. Four things are never published here: invented campaign results, screenshots of dashboards nobody opened, personal measurements that cannot be verified, and traffic-quality claims a reader could not check against their own logs. The reasoning behind those exclusions is in the editorial policy, the working method behind them is on the author page, and corrections go through the contact page.