Here is the uncomfortable thing most affiliate managers learn too late: by the time you discover an affiliate has been bidding on your brand, the money is usually already gone — and your own contract is the reason you can't get it back.
Not because the network is hiding anything. Because the system is built to pay out by default. Affiliate networks auto-approve commissions unless you actively decline them inside a short window. Miss the window — and most brands miss it, because they aren't scanning their own brand terms — and the commission locks. Once it locks, your agreement with the network typically says, in writing, that you have no right to recover it.
So brand-bidding fraud is not really a detection problem, and it's not really a legal problem. It's a race against a clock most brands don't know is running. This piece is about that clock: where it's written down, how long you actually have, and why it turns detection speed from "good hygiene" into the entire ballgame.
I run AdCrime and I sell software that wins this race for people, so weigh that as you read. But the clauses below are in your contracts whether you ever talk to me or not — and you should read them.
The clause nobody reads until it's too late
Pull up your affiliate network's advertiser agreement and find the section on validated transactions. I'll use Awin because their language is the bluntest on record — and because Awin is one of the biggest networks in the verticals brand bidding hits hardest.
Awin's Access advertiser agreement (the April 2023 terms) says, in substance:
"The Advertiser… has no right whatsoever to recover from the Company or any Participating Publisher any Fees or Commissions paid… in respect of Approved Transactions or Approved Leads… The Advertiser hereby waives and releases the Company and any Participating Publisher from any claims the Advertiser may have for recovery of paid Fees or Commissions."
Read that twice. Once a transaction is approved, you have "no right whatsoever" to recover the commission, and you have waived any claim to it. Not "limited rights." Not "subject to review." None.
And Awin's own help documentation (updated March 2026) closes the door behind it:
"Approved transactions cannot be reversed… Transactions and the associated commission already paid to partners cannot be reversed."
I'm a founder, not a lawyer, and clause numbering varies between Awin's UK, US, and German entities and between contract versions — so go read your copy, not mine. But the structure is not unique to Awin. It is the standard shape of the deal across the industry. The network's position is consistent and legally tidy: we facilitate the relationship, you police it, and once we've approved a transaction, the commission is final.
Doing nothing is a decision — and the decision is "pay the fraudster"
Here's the part that turns the screw. You don't approve these transactions by clicking "approve." They approve themselves.
Awin again, on auto-validation:
"Each Transaction and Lead will auto-validate and be deemed to have been approved by the Advertiser at the end of the respective Validation Period, unless declined…"
The Auto-Validation Period can't even be switched off. It runs on a timer, and when the timer hits zero, the commission is "deemed approved" — by you — whether you looked at it or not. The burden is entirely yours: you have to catch the bad transaction and actively decline it before the clock runs out. Silence is consent. Inaction pays.
This is true across the networks, not just Awin:
- CJ Affiliate: transactions lock on the 10th of the following month, automatically. Once locked, a transaction "can no longer be edited or voided."
- ShareASale: locks on the 20th of the following month; once locked, the commission is "officially released to the credited Affiliate… [and] can no longer be edited/voided."
- Impact: the action-locking period runs from the tracked date; "once this period has passed," the brand "can no longer modify or reverse the action."
- Rakuten: once an order locks, the commission "cannot be adjusted or cancelled."
In affiliate marketing, doing nothing is not neutral. It is an instruction to the network: pay them.
The window is shorter than you think
So how long do you actually have? Less time than the size of the problem deserves. These are typical windows — every one is configurable per advertiser agreement, so verify your own contract — but the shape is consistent:
| Network | Typical window before the commission locks | After it locks, can you reverse it? |
|---|---|---|
| Awin | Up to ~67 days (Access; advertiser-set, auto-validates) | No. Access: "approved transactions cannot be reversed" (support, exceptional cases only). Already paid: never. |
| CJ Affiliate | Locks the 10th of the following month (or custom 7–60 days) | No — "can no longer be edited or voided." |
| ShareASale | Locks the 20th of the following month | No — commission "officially released… can no longer be edited/voided." |
| Impact | Advertiser-set, commonly ~27–46 days | No — "can no longer modify or reverse the action." |
| Rakuten | Monthly invoice close, or per-order auto-lock | No — once locked, "cannot be adjusted or cancelled." |
| Partnerize | Contractual "Force Approval Period" | Contract-specific; not a self-serve reversal. |
(Windows are typical and configurable. Confirm the exact terms in your own advertiser agreement before you rely on a number.)
Notice what these have in common. None of them gives you months of comfortable leeway. A transaction that happened on the 2nd of the month might lock on the 10th or 20th of the next — call it three to seven weeks. If your "monitoring" is a quarterly audit, or a SERP check you run when you happen to think of it, the commission has locked and paid long before you ever look. You are not late by a little. You are late by design.
The recovery curve falls off a cliff
In the Playbook I modeled this out, and the shape is worth seeing because it explains why speed beats everything else. Take a single infringing affiliate generating ~$2,000/week in fraudulent commissions, where ~90% of those sales would have converted organically anyway. Your recoverable loss isn't linear — it's a cliff:
- Week 1 — transaction recorded, window wide open. Decline it and recovery is ~95%. Net loss: pocket change.
- Week 3 — commissions approaching validation on the faster networks. Recovery ~75%.
- Week 5 — most have locked and paid. Recovery now needs a direct demand to the affiliate. ~30%.
- Week 9 — locked, paid, affiliate moved on. Recovery ~15% if they're a U.S. entity, ~0% if they're offshore.
(That's an illustrative model, not audited data — but every affiliate manager who has run a real clawback recognizes the curve.) Multiply it across the eight-to-twelve infringing affiliates that sit on a typical mid-market program at any given time, and a few weeks of delay is the difference between a rounding error and a six-figure write-off.
The cheapest fraudulent commission to recover is the one you decline before it ever validates. Every day after that, your odds get worse and your contract gets less friendly.
What this actually changes about how you fight fraud
Once you internalize the clock, the whole problem reorders itself.
Detection speed stops being a "nice to have" and becomes the entire strategy. A brilliant evidence package assembled in week six is a write-off with footnotes. A mediocre one filed in week one gets your money back. Catching the fraud fast enough to decline the transaction inside the validation window is worth more than catching it perfectly.
"Dispute" is the wrong verb. "Decline" is the right one. Disputing a paid commission means asking a network to do something its own terms say it doesn't have to. Declining a pending transaction inside the window is something the platform is built to let you do. Same fraud, two completely different success rates, separated only by which side of the lock date you're on.
The unit you're racing to find is the publisher ID. You can't decline what you can't identify. The static publisher ID — pulled out of the redirect chain, not the per-click hash — is what lets you find the transactions and file the decline. (That's its own discipline: how to read an affiliate redirect chain.) Then it's the network-specific decline-and-terminate workflow, which is a piece of its own and is laid out in the Playbook.
None of this is exotic. It's just a reordering of priorities around a clock that the contracts have been quietly running the whole time.
Everything here is in contracts you already signed and dashboards you already have — no software required to understand it, and the decline buttons are already there. The hard part isn't filing. It's finding the fraud, across every brand term and every region, fast enough to file before the clock runs out — which is the part I built AdCrime to do. Either way, the goal is the same: fewer fraudsters getting away with it.
Mario Vaher is the founder of AdCrime, which scans Google Ads across multiple regions for affiliates bidding on brands they aren't authorized to touch, captures the redirect chain and publisher ID behind each one, and packages it as evidence affiliate managers can file — fast enough to matter. If you want to see what's bidding on your brand right now, the first scan is free — or read the full, ungated Affiliate Brand-Bidding Fraud Playbook.