From Screenshot to Terminated Publisher: How Affiliate Brand-Bidding Enforcement Actually Works (and the Three Places It Breaks)

Let me start with the most honest account of brand-bidding enforcement in the public record, because it tells you what you're actually up against.

Jamie Birch, who runs the affiliate agency JEBCommerce, has written about an affiliate who bid on the brand terms of a retailer he was managing, Coldwater Creek. The affiliate geo-targeted the ads to exclude the state Birch worked in, so every manual check Birch ran came back clean. In his words: "I only found out when I traveled to a state I usually didn't travel to… I had no idea what geo targeting was at the time and he ended up swindling away tens of thousands in commissions from us that we weren't able to get back."

Caught the fraud. Lost the money anyway. That's the gap this article is about — not detection, which everyone sells, but enforcement, which almost nobody explains honestly. The clean story goes "find the ad, file with the network, they reverse the commission and terminate the publisher." The real story has a Step 0 nobody mentions, a dead end most people waste a week in, and a clock that's usually already run out. Here's the whole thing, with the contract clauses and the case outcomes, and the three places it breaks.

I run AdCrime, and our product exists to win this race for people. I'll be straight about where software helps and where it doesn't — because half of this is contract work you do once, with no tool involved.

The workflow, and where it breaks

1. Step 0 — Have the contract first. You can't claw back what your program terms don't prohibit. (Breaks here if your terms are vague.)
2. Step 1 — Build evidence that bridges the double-blind. The network can't see the affiliate's ad — you have to hand it to them, publisher ID and all.
3. Step 2 — Don't waste a week on Google. The trademark complaint is a dead end for most brand bidding. (Breaks here if you expect Google to fix it.)
4. Step 3 — Reverse the commission inside the validation window. It's a financial action, not a "policy complaint," and it has a deadline. (Breaks here if you're slow.)
5. Step 4 — File with the network the right way, then escalate. Account manager → demand letter → litigation, only as far as you need to go.

Step 0 — Have the contract before you have the evidence

This is the step that decides whether any of the rest is enforceable, and it's the one detection tools can't do for you.

Most affiliate networks do not prohibit brand bidding at the network level. In the research behind this piece, the only major network whose base publisher agreement names keyword bidding on brand terms as a violation was Partnerize. Everyone else — Awin, CJ, Impact, ShareASale, Rakuten — leaves it to the advertiser's program terms. Awin's framework is explicitly two layers: the network-level publisher terms cover general fraud and IP infringement, and the granular "you may not bid on our brand name" rules live in each advertiser's own PPC policy. Reversal authority flows from your program terms, not the network's.

Which means the enforcement lever is a contract you have to write. And the language matters enormously:

"Operators who lack explicit brand-bidding policy language in their affiliate contracts cannot enforce commission clawback: without a documented prohibition, affiliates can contest any recovery attempt, and courts or arbitrators will side with the ambiguity." (NewMotion IT enforcement guide, May 2026.)

What strong language looks like — this is a real affiliate program's terms, quoted by an operator in an Indie Hackers thread:

"We have a strict no tolerance policy on PPC trademark bidding. We will not enter a discussion about when the violation started and when it stopped; you will forfeit all Affiliate Fees for a minimum of the past 30 days, your Affiliate Fee balance will be set to $0 without warning, and we may terminate your participation in the Program at our discretion." (ConvertKit affiliate terms.)

Notice what that does: it makes the penalty mandatory, not discretionary, it pre-empts the "when did it start" debate, and it grants the clawback and termination right in advance. The enforcement guidance is blunt about why that phrasing wins — "'Violations may result in suspension' gives you discretion but implies the affiliate might receive only a warning… 'Second violation results in commission reversal for the violation period and account suspension' removes ambiguity and signals a process the affiliate cannot negotiate around."

The networks' own machinery backs this up if the contract is there. CJ's publisher agreement (the version filed publicly with the SEC) lets an advertiser charge back a payout for "Publisher failure to comply with Advertiser's Program terms," and states plainly that "Charge-backs may be applied to Your Account at any time, including previous payment cycles" — and that an advertiser may terminate a publisher for "diluting, tarnishing or blurring an Advertiser's trademarks." (That's an older filing and current terms differ in detail, but the charge-back-for-program-breach structure has held across the industry.)

So: before you scan for a single ad, make sure your program terms — on every network — explicitly prohibit bidding on your brand name and its variations, and grant you the right to reverse commissions and terminate for it. If they don't, your detection is unenforceable. Fix the contract first. The full clause language is in the Playbook.

Step 1 — Build evidence that bridges the double-blind

Here's a structural fact that explains why weak complaints get ignored. The network can't see what you saw. They see a click and a commission — not the affiliate's Google Ads account, not the SERP, not the keyword that was bid. As one network's position is fairly summarized: they cannot verify the ad, the keyword, or the campaign. This isn't always stonewalling; it's a genuine visibility gap. Your evidence package has to close it entirely from your side.

The package that actually triggers a reversal — assembled from what monitoring vendors and compliance guides converge on — is seven things:

1. A timestamped SERP screenshot with the keyword visible (in the search bar or URL) and the ad in context.
2. The ad copy — full headline and description.
3. The affiliate tracking URL — the real click destination, not the display URL.
4. The redirect chain — every hop from click to landing page.
5. The static publisher ID resolved out of that chain — awinaffid=111222, siteid=, the Impact partner ID. Not a per-click token (awc, irclickid, cjevent): those reset on every click and identify nothing. If that distinction isn't reflexive yet, it's the entire subject of how to read an affiliate redirect chain.
6. The capture location and time — because the single most common affiliate defense is geo-targeting.
7. The date range — to scope the commission reversal.

Two of these are where amateurs lose. The publisher ID is the linchpin — a screenshot without it names no one, and the network can't act on "someone is cheating us." And the location/time, because of the Coldwater Creek problem: a sophisticated affiliate runs ads only in regions you don't watch and only after hours, so a single check from your office at noon proves nothing. Monitoring vendors counter this by capturing "from a global network of servers at all hours of the day." A multi-region, timestamped capture is far harder for the affiliate to wave away as "we don't even advertise there."

And the hard case: cloaking. A cloaked affiliate shows a clean page to monitoring bots and the real redirect to actual users — which is why industry reporting attributes a large and rising share of affiliate fraud to it (mFilterIt's figure, circulated via Search Engine Land, is ~45% by 2024 — vendor data, so treat the precise number as directional). Beating it requires evidence captured the way a real user would see the page, not the way a scraper does — because if the cloak can tell your check from a real visitor, it serves your check the clean version and you walk away with nothing.

Step 2 — Don't waste a week on Google

Most brands' first instinct is to file a Google Ads trademark complaint. For affiliate brand bidding, it's usually a dead end, and it's worth knowing exactly why so you don't burn a week on it.

Straight from Google's trademark policy (current as of 2026): Google will not restrict "using trademarks as keywords." It only reviews the use of a mark in ad text, only against specific advertisers you identify by URL, and only where you hold registered trademark rights in that country. Since July 24, 2023, complaints are per-advertiser and per-ad, not industry-wide — so every new infringing affiliate account is a fresh complaint. And the courts back Google's stance: in 1-800 Contacts v. Warby Parker (2d Cir., Oct. 8, 2024), the Second Circuit held that bidding on a competitor's trademark as a keyword "standing alone — cannot support a claim of trademark infringement absent additional use" of the mark.

Here's the gap that creates. A brand-bidding affiliate bids on your name as a keyword, runs an ad with no brand name in the copy ("Today's Best Coupons →"), and redirects through an affiliate URL. Nothing in that ad's text uses your mark, so Google's complaint mechanism never touches it. The keyword bid — the actual harm — is explicitly outside Google's enforcement.

So: file the Google complaint only when the affiliate put your trademark in the ad copy or a confusing display URL. Then it can work. For a clean-bid affiliate, skip it and go to the network. (One exception worth knowing: in the EU/EFTA, Google's policy allows it to restrict trademark use as a keyword after a complaint — materially different from the U.S., and useful if you operate in Europe.)

Google will take your brand out of the affiliate's headline. It will not take the affiliate out of your auction. Those are different problems, and only the network can fix the second one.

Step 3 — Reverse the commission inside the validation window

This is the lever you actually have, and the verb matters: you're not lodging a "policy complaint," you're reversing a payment — and payments have a deadline.

Affiliate commissions auto-approve unless you decline them inside a validation window, and once they lock, recovery gets hard-to-impossible. The windows are short and they vary: CJ can lock as little as ~10 days from the transaction; others run 30–90. Whether you can still reverse after the lock depends on the network and your contract — CJ's terms (above) allow charge-backs for a program-terms breach "at any time, including previous payment cycles," whereas Awin's terms state advertisers have "no right whatsoever to recover" commissions on approved transactions. The reliable rule across all of them: decline inside the window. The full per-network breakdown of those clocks is its own piece — why your network won't claw back the money — and it's the reason detection speed, not detection perfection, is what gets your money back.

One more reality that quietly defeats people, specific to Impact: its monitoring "will not generate violations for non-contracted partners." If the fraudster isn't already inside your Impact program, Impact's paid-search monitoring is blind to them. The same is broadly true elsewhere — the network's enforcement reaches the publishers it has a relationship with. A fraudster operating entirely outside your programs is a trademark/legal problem (Step 4), not a network-reversal one.

Step 4 — File with the network the right way, then escalate

When you do have a contracted publisher and a complete package, file it as an individual case (batched complaints get bulk-dismissed), cite the exact program clause, and state the action you want: reverse commissions for the window, terminate the partnership, confirm in writing. The structured template is in the Playbook — fill the variables, don't rewrite the structure, because the structure is what makes it processable.

Then be realistic about the response, because this is the third place enforcement breaks: networks are slow and inconsistent, and almost none publish an SLA for routine brand-bidding cases. The clearest public evidence of how slow is the Honey saga. Public evidence of Honey's affiliate-attribution conduct landed in December 2024; Rakuten didn't terminate Honey until January 12, 2026, Impact suspended on the 16th, and Awin confirmed the breaches and suspended payments on the 21st — more than a year after the evidence, and only after a class action and millions of users had left. Three networks, the same confirmed conduct, three different responses (Rakuten terminated; Awin merely suspended). One consultant's public reaction captured the enforcement-consistency problem exactly:

"What's being discussed here is not a grey area… it is a credible allegation of fraud. Partners have been removed from platforms for far less in the past." (Benjamin Völk, on Awin's Honey response, Jan. 2026.)

If a network slow-walks or refuses, you climb a ladder — and only as far as you need to:

• Layer 1 — your account manager. Above a spend tier, every network assigns one. They can move a transaction from "under review" to "reversed" in days instead of weeks, and they respond differently to "systemic violation, terminate this partner" than to a queue ticket. Ask for any refusal in writing — most will reverse rather than create that paper trail.
• Layer 2 — a demand letter to the affiliate. When the network won't terminate a large partner, or the publisher is operating outside any program, you go direct. This is pre-litigation, so it goes out on counsel's letterhead, names the registered trademark, documents the ads/URLs/dates, demands cessation and disgorgement, and sets a 14–30 day deadline. I'm a founder, not a lawyer — this is the rung where you bring one in.
• Layer 3 — litigation. Rare, and post-Warby Parker you need more than keyword bidding: the Lanham Act claim attaches to the mark in the ad copy or a confusing landing page, not the bid itself. The credible threat is what makes Layer 2 work; the filing itself almost never comes. The full legal menu — Lanham Act, FTC, state unfair-competition, UDRP for typosquats — is in the Playbook.

What actually moves the odds

Strip away the steps and two variables decide whether you recover anything: speed and evidence quality.

Speed, because the validation clock is usually already running when you find the ad, and Coldwater Creek is what slow looks like — caught, but "tens of thousands we weren't able to get back." Evidence quality, because the network is blind to the affiliate's ad and will only act on a package that resolves a real publisher ID, captured where and when the ad actually ran, in a way the cloak couldn't dodge.

That's the honest case for systematizing detection — and it's where the vendor case studies, read with the right amount of salt, point. Adthena publishes a Sage case study claiming a 75% drop in branded CPCs within five days of acting on detection data (and a Go City study claiming an 86% reduction in partners bidding on brand terms in the first month); Bluepear documents a Rhino Affiliates engagement recovering €11,000 in disputed payments out of €131,000 in savings. Every one of those is the vendor reporting its own result, so don't bank the decimal — but the direction is consistent and it matches the mechanics: when you catch it fast, with evidence that names the publisher, the reversal and the CPC relief follow.

None of this requires our software. The contract you write once. The Google call you can reason through. What needs software is the part that actually breaks — catching the violation across every brand term, every region, and every hour, fast enough to decline inside the window, and clean enough to name the publisher even when the affiliate's link is cloaked. That's what I built AdCrime to do, and the evidence it produces is built to match each network's compliance form. 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, resolves the redirect chain and publisher ID behind each one, and packages it into the evidence file you'd send to the network — built to match each network's compliance form, fast enough to act inside the validation window. 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.