Pre-Bid vs Post-Bid Fraud Detection: Why Timing Decides Whether Your Budget Survives

In short: post-bid detection tells you how much of your budget went to fraud after you have spent it. Pre-bid detection declines to buy the impression in the first place. The fraud found is often the same — the difference is whether you are reading a report or keeping the money.

Almost every brand running programmatic media already has fraud detection. Very few have fraud prevention. The gap between those two words is the whole subject of this article, and it is worth real money.

What post-bid detection actually does

Post-bid verification sits downstream of the purchase. The bid is placed, the impression is served, the invoice is raised — and then a tag or log-level feed analyses what was delivered and flags the share that was invalid traffic.

It is genuinely useful. It is how you learn your real invalid-traffic rate, which placements are rotten, and which supply paths to cut next quarter. It produces the number you take to your agency.

What it does not do is stop the purchase. By the time the report exists, the impression has been bought, served and paid for.

What pre-bid filtering does instead

Pre-bid filtering runs before the bid. Each bid request is scored on the signals available at that moment — device and ad ID history, IP and data-centre reputation, app or domain profile, traffic patterns on that supply path — and requests that fail are simply not bid on.

No impression. No invoice. No credit to chase. The budget stays in the account and gets spent on the next request that passes.

The cost of the gap, worked through

Take a £100,000 monthly programmatic budget running at a 20% invalid-traffic rate — unremarkable for open-web display, given the web-wide bad-bot share of 37% (Imperva, 2025).

With post-bid detection only: all £100,000 is spent. The report arrives and flags £20,000 as invalid. You raise it with your supplier, and some portion comes back — but only what the vendor can prove, typically as media credit with the same supplier rather than cash, and typically weeks or months later. You have paid full price for 80% delivery, and your remedy is a voucher.

With pre-bid filtering: the bids behind that £20,000 are never placed. The money is still in the budget, in the same month, and buys inventory that can actually reach a person. You get roughly 20% more reach for the same spend — not as a rebate, but as working media.

The compounding matters more than the single month. Credits recover a fraction of a fraction, late. Impressions never bought are 100% recovered, immediately, every month.

The honest limitation

Pre-bid is not a complete answer. It works on the signals visible before the auction, and the more sophisticated attacks — the ones built to look human — are designed precisely to survive that moment. Some fraud will always get through and some good inventory will be declined.

That is exactly why post-bid measurement still matters: it is the audit that tells you how well the filter worked. The mistake is not using post-bid verification. The mistake is using it as the defence — treating the monthly fraud report as protection when it is really a receipt.

Ask your agency one question: are we filtering before the bid, or reporting after it? If the answer is only the second, your fraud strategy is documentation.

Want to see what pre-bid filtering keeps in your budget? Run a live campaign with us.