Publishers host a plain text file listing the companies authorized to sell their inventory. It exists because the programmatic supply chain had no built-in way to check who was selling what.
Bid requests carry claims, not proof
A bid request states the domain the impression comes from. Before the file existed, nothing in the pipeline verified that the claim was true.
Fraudulent sellers exploited this by listing worthless inventory under the names of well-known publishers, and buyers paid premium prices for impressions that were never there.
Because the deception happened inside automated systems, the affected publishers often had no visibility into it and no way to disprove it.
The file inverts the burden of proof
The publisher publishes, at a fixed path on its own domain, a list of the exchanges and resellers permitted to sell its inventory, with each one's account identifier.
Buyers crawl those files and compare the seller identity in every bid request against the authoritative list for the claimed domain.
An unlisted seller offering that domain is now demonstrably unauthorized, and the buying platform can discard the request without needing to investigate.
Adoption worked because it was cheap on both sides
Publishing the file requires no integration, no vendor and no ongoing cost. It is a text file served from the root of a domain.
Buyers had a clear incentive to enforce it, since unauthorized inventory was money leaving their campaigns for nothing.
That combination of near-zero implementation cost and direct buyer benefit is why the standard spread across the American market faster than most ad tech proposals.
Companion files extended the same idea
The publisher-side file only covers the first hop. Inventory frequently passes through several intermediaries before reaching a buyer.
Sellers therefore publish their own file declaring the entities behind the identifiers they use, and bid requests can carry the chain of intermediaries involved.
Together these let a buyer reconstruct the path an impression took and see how many parties stood between the publisher and the purchase.
Maintenance failures create real revenue loss
A publisher that adds a new demand partner without updating the file will find that buyers reject the resulting requests as unauthorized.
Stale entries cause the opposite problem, leaving former partners nominally authorized long after the relationship ended.
Because nothing breaks visibly, these errors persist for months, which makes periodic review of the file part of routine revenue operations rather than a one-time setup task.