The Refund That Never Landed

Advertisers are reporting that Google Ads credits are being revoked after the money tied to them has already been spent. That is the detail that matters. A credit issued before spend and pulled before spend is a correction. A credit pulled after spend has already gone out the door is something else: it is a platform reaching back into a transaction it already allowed to complete.

This is happening in the same window that a new report projects AI buying platforms will manage 27% of U.S. ad spend by 2030. Put those two facts side by side and the argument writes itself. The industry is being asked to move more budget into systems with less line-by-line human control, at the exact moment the largest ad platform in that industry cannot get the basic mechanics of a credit right.

What "revoked after spending" actually means

A credit is supposed to be one of the simplest transactions in advertising. It is issued, it is applied, it reduces what an advertiser owes or increases what they can spend, and the account reflects that. Search Engine Land's reporting describes advertisers watching credits disappear only after the spend tied to them has already happened, which means the advertiser has already made decisions, allocated budget and run campaigns on the assumption that the credit was real and final.

There is no version of this that looks like ordinary platform housekeeping. Once spend has occurred, a credit is not a projection or an estimate that can be adjusted. It is a settled fact. Reversing it after the fact does not correct an error in the advertiser's favour; it corrects one in the platform's.

Advertisers do not have a seat at the table when Google decides how its credit systems are designed, audited or reversed. They find out what the policy actually is when the reversal shows up in the account, not before.

The automation pitch arrives at the worst possible time

The push toward AI buying platforms has been framed, consistently, as a story about efficiency: less manual bidding, less manual placement, more of the budget decision handed to a system that claims to optimise faster than a human team can. The 27% by 2030 figure is being used across the industry as evidence that this shift is not speculative, it is already underway and accelerating.

But efficiency and control are not the same thing. An automated buying platform that manages a growing share of ad spend is, by definition, a platform that advertisers are trusting with less oversight of individual decisions. That trust has to be earned somewhere. It cannot be earned by a track record of quietly reversing credits after the fact, because that is precisely the kind of decision an advertiser would have caught if they still had full visibility into every line.

The pitch for automation has always rested on the idea that the platform is a more reliable steward of the budget than manual oversight would be. Credit revocations after spend argue the opposite. They suggest a platform is willing to adjust the ledger in its own favour after the advertiser has already lost the ability to react.

Accountability is showing up elsewhere, not here

Regulators have not been silent on advertising practices this year. The FTC and the State of Connecticut secured a $4 million settlement with a vehicle dealership, a clear signal that misleading advertising claims carry real financial consequences when regulators choose to act. TikTok, for its part, rejected Meta ads that were pushing rivals to join a child safety settlement, an editorial decision made in public and reported as such.

Google's credit revocations have not drawn that kind of scrutiny yet. There is no reported settlement, no consumer protection action, no public accounting of how many advertisers were affected or how much money was involved. The reporting so far consists of advertisers describing what happened to their own accounts. That is not nothing, but it is a different category of accountability than a $4 million settlement with named regulators attached.

It is worth noting, too, that Google has recently documented a partner-only API for full-web search results, a move that tightens who gets direct access to search data rather than opening it further. Read together with the credit reports, the pattern is one of a platform pulling control inward: more gatekeeping over access, less transparency over billing outcomes advertisers thought were settled.

What advertisers should actually watch

None of this means automation itself is the problem. Google Analytics has just launched customisable dashboards, a genuinely useful reporting feature that gives advertisers more visibility into their own data, not less. The complaint here is not that platforms build tools. It is that the same company building those tools is also the one whose credit practices advertisers are now describing as unreliable after the spend has already happened.

Advertisers moving budget toward AI buying platforms should be asking a specific question, not a general one: what happens to a credit, a make-good or an adjustment after the money has already been spent, and is that policy written down anywhere an advertiser can point to before the fact. If the answer is that the policy only becomes clear when the reversal shows up in the account, that is not a detail. That is the whole risk.

The industry has spent the last several years selling automation as a trust upgrade. The credit reports are a reminder that trust is earned at the level of the smallest transaction, not the biggest projection. A platform that cannot reliably honour a credit after spend has not earned the right to manage a larger share of the budget with less oversight attached to it.