The Brand Protection Stack for Scaling Multi-Marketplace Brands

8 min read

TL;DR

• Amazon says its Counterfeit Crimes Unit pursued more than 24,000 bad actors and seized 15 million counterfeit products worldwide using AI and multimodal models back in 2024.
• That same number is a two-sided signal: strong enforcement volume, but also evidence of counterfeit activity produced with the same generative tools brands use to move faster.
• eBay’s Verified Rights Owner programme, running since 1998 with more than 40,000 registered rights owners, operates as report-and-remove only, it doesn’t verify units before shipment the way Amazon’s Transparency does.

There’s a point in a brand’s growth where counterfeit activity stops being an occasional annoyance and becomes a recurring operational cost. As the catalogues expand, new marketplaces get added, sometimes across borders, and the same duplicated listing starts appearing in multiple places at once.

And the scale of the underlying problem is widely known:

• Global trade in counterfeit and pirated goods reached an estimated USD 467 billion in 2021, or 2.3 percent of total global trade, according to the OECD and EUIPO’s 2025 joint report, Mapping Global Trade in Fakes.
• Clothing, footwear, and leather goods accounted for 62 percent of all counterfeit goods seized globally, per the same report.
• Around 65 percent of seizures now involve small parcels and mail, a shift the report attributes to counterfeiters favouring shipping methods that are faster and less scrutinised.

For a brand still relying on one takedown process across every marketplace and every country, the real issue isn’t effort, it’s that listing-level enforcement was never built to solve this at the volume, or the geographic spread, counterfeiters now operate across.

Why Listing-Level Enforcement Breaks at Scale

While takedown removes a listing, it does not remove the seller, their funds, their other channels, or their reason to try again. That gap barely matters for a single incident. It becomes the whole problem once a brand is fielding the same pattern across multiple marketplaces and multiple countries weekly, because every removed listing is a fresh account, sometimes in a different jurisdiction entirely, away from reappearing.

Amazon has said its detection systems use AI and machine learning, including multimodal models that analyse listing images, text, pricing patterns, and seller behaviour together rather than reviewing listings in isolation, according to its March 2025 Brand Protection Report. Its Counterfeit Crimes Unit, which extends enforcement into civil litigation and law enforcement referrals, says it pursued more than 24,000 bad actors and seized 15 million counterfeit products worldwide in 2024.

Read one way, that’s a genuine testament to how far automated enforcement has come. Read another way, it’s a baseline metric of just how fast the counterfeit economy is industrialising its own use of AI, since 15 million seizures in a single year only happens if there’s a comparable volume being produced and listed in the first place, much of it with the same category of generative tools brands themselves are starting to rely on. Amazon’s AI is simultaneously the strongest evidence that automated enforcement works, and the clearest sign that the problem it’s enforcing against is scaling just as fast underneath it.

That has two practical implications for a scaling brand:

• It raises the floor. Automated, AI-assisted detection catches far more volume than manual review ever could, which is part of why enrolling in a marketplace’s own brand protection programme matters more than building a parallel manual process from scratch.
• It doesn’t close the loop. The tools that speed up a brand’s own documentation and monitoring are the same category of tools speeding up convincing counterfeit listings, cloned storefronts, and fake seller profiles, which is why the OECD’s 2025 report frames this as an evolving, not settled, enforcement challenge.

Amazon’s Three-Layer System

For brands selling on Amazon, Brand Registry is the entry point, requiring a registered or pending trademark, in whichever country or region the brand is enforcing in, and it unlocks two further tools worth using together:

• Project Zero lets enrolled brands remove counterfeit listings directly, without waiting on Amazon’s review queue, feeding that activity back into Amazon’s detection systems.
• Transparency operates further upstream: each physical unit gets a unique scannable code, and Amazon will not ship a unit lacking one, which stops counterfeit stock before it reaches a customer rather than reacting to a listing after the fact.

For a scaling brand, Transparency is usually the tool worth prioritising for high-risk categories specifically, electronics, supplements, beauty, anything where counterfeiting is common, since it’s the only layer in this stack that intercepts fakes before they’re ever listed.

Border Enforcement, by Region

Marketplace tools, including Amazon’s, only ever reach that one platform. Border-level enforcement reaches further, but it works differently depending on where a brand sells.

In the United States:

• Trademark and copyright owners can record their registration with US Customs and Border Protection for $190 per class of goods.
• CBP reported seizing more than 78 million counterfeit items in fiscal year 2025, with a combined manufacturer’s suggested retail price of over $7.3 billion had the goods been genuine.

In the United Kingdom:

• Rights holders file an Application for Action (AFA) with HMRC and Border Force, a service that is free to use, according to GOV.UK’s own guidance.
• An AFA is valid for one year and can be renewed, and rights holders are advised to submit it around 30 working days before goods are expected to arrive.
• UK trademark registration through the UK Intellectual Property Office costs £205 for the first class online, plus £60 for each additional class, as of fees updated in April 2026, with no annual maintenance fee across the ten-year term.

In the European Union:

• The equivalent Application for Action is also free to submit, filed through the EUIPO’s IP Enforcement Portal (IPEP), and can be filed as a national AFA for one member state or a Union AFA covering multiple member states at once.
• An EU AFA is valid for one year.
• A single EU trademark, filed through EUIPO, costs €850 for the first class online and covers protection across all 27 member states with one registration, rather than filing separately in each country.

The practical takeaway: US brands pay for border recordation but gain a well-established, high-volume enforcement track record. UK and EU brands get the equivalent border protection for free, which makes it one of the more underused tools available to brands already selling into those markets, since cost isn’t the barrier there that it is in the US.

eBay’s Verified Rights Owner programme remains relevant across all three regions for brands cross-listing there, but it’s worth understanding its limit: it’s a report-and-remove system, not a verification-before-shipment one like Transparency, and it depends on active monitoring rather than intercepting fakes automatically.

Building the Case to Escalate

Some counterfeit sellers are worth pursuing beyond a standard takedown request, specifically the repeat, identifiable ones who relaunch under new accounts after each removal, in whichever country they’re operating from. Legal specialists draw a clear distinction here: takedowns remove content, legal action targets conduct.

Building that case requires documentation most brands don’t collect as a matter of routine:

• Confirmed test purchases
• Records of lost sales or refund spikes tied to the activity
• A dated paper trail establishing the pattern isn’t isolated, across whichever marketplaces and regions it’s appeared in

FAQ