Two experienced brand-protection lawyers can look at the same marketplace listing and reach opposite, defensible conclusions. One sees a counterfeit. The other sees a gray market problem: genuine goods moved through an unauthorised channel. That disagreement is not a knowledge gap — it is the shape of the underlying law. At volume it becomes the largest single source of wasted enforcement capacity: rejected notices, duplicated investigations, and legal spend consumed by matters that should never have entered the queue. Getting counterfeit vs gray market goods right is the first legal decision in online enforcement, and it is still made on instinct.
The problem: enforcement breaks at the first legal call
Programmes are built around detection because detection is measurable: listings scanned, alerts raised, links found. The decision that follows each alert — what is this, and what can we assert — has not scaled with the scanning. It runs through a person, one listing at a time, at the speed of reading and judgement. When a programme grows from fifty alerts a month to five thousand, the scanner keeps up and the classification does not.
The two answers lead to legally opposite conclusions. If the item is counterfeit, the rights holder has a federal claim and a strong platform remedy. If it is genuine stock sold without authorisation, trademark rights are generally exhausted by first sale and the dispute is commercial, not counterfeiting.
Both errors are expensive, and they are not symmetrical. A counterfeiting complaint against genuine goods gets rejected — and the rejection is recorded against the brand's accuracy standing, the same standing that governs access to faster removal tools. A genuine counterfeit treated as a channel dispute stays live, keeps selling, and in categories touching consumer safety, keeps shipping. Neither outcome comes from weak detection. Both come from counterfeit vs gray market goods classification made too late, too thin, or not at all.
Why counterfeit vs gray market classification is so hard
The legal tests differ in kind, not degree
Counterfeiting is a narrow concept. The federal definition requires a spurious mark identical with, or substantially indistinguishable from a mark registered on the principal register and in use, applied to the goods or services for which it is registered, and likely to cause confusion, mistake, or deception. It deliberately excludes marks merely reminiscent of a protected mark, and does not extend to trade dress such as colour, shape, or packaging unless separately registered.
Genuine goods sit outside that definition. Under first sale, a trademark owner that authorised a product's first sale generally cannot use trademark law to control its later resale. An unauthorised reseller of authentic stock is normally a distribution problem — unless the goods are materially different from what the brand authorises for that market, which triggers a third regime: restricted gray market articles, where the remedy is disclosure and labelling rather than seizure, and where the rights holder must prove physical and material difference.
Three regimes, different evidence, different remedies — and all three wearing the same product photograph.
The listing page does not disclose the answer
A lookalike listing can be a copied mark on a fake product, a genuine unit rerouted from another market, a legitimate unit listed by a reseller who broke a channel agreement, or a hybrid offer that shows an authentic image and ships something else. Price is not a reliable tell. Seller opacity is not a reliable tell. Sellers have every commercial reason to keep the question ambiguous, because ambiguity delays action and delay is profitable.
The evidence burden differs by theory
Counterfeiting allegations generally want physical proof: a purchased sample, comparative analysis of labels, packaging, materials and markings, and a record of how the sample was obtained and handled. Gray market and channel matters want documentary proof: invoices, authorisation chains, distributor agreements, serial or batch data, and evidence of what the brand authorised for that territory. These are not interchangeable. Running the heaviest evidence path on every alert is unaffordable; running none produces notices that cannot be verified, which is how accuracy records degrade.
Enforcement systems have no memory and no baseline
Rights scope lives in one place, the authorised distributor list in another, seller history in a spreadsheet, prior outcomes in a shared inbox. There is no maintained reference for what the genuine product looks like in each market, so every reviewer rebuilds the comparison. Triage defaults to arrival order or platform, and a seller who reappears under a new storefront is assessed as a first-time incident.
How AI resolves counterfeit vs gray market classification at scale
AI is not required to make the hard call in a genuinely contested matter. It is required to make the repeatable part of the call consistent, complete, and traceable, so judgement is spent only where it changes the outcome.
Maintain the reference set once. Classification is a comparison, and comparisons need a fixed reference. Registrations with their classes and territories, the authorised channel structure, and a genuine-product baseline per market can live as structured data rather than being rediscovered per case. That removes most inter-reviewer variation and is the precondition for everything downstream.
Extract the signals that distinguish the theories. A listing holds more legally relevant data than a screenshot: how the mark is used and where, whether imagery matches known brand assets, claimed product identifiers, category overlap with the registration, price and velocity against the authorised range, and the seller's prior appearances. Normalising those attributes into one matter record makes comparison possible and makes missing fields visible — often the honest answer to why a claim cannot yet be filed.
Price the evidence to the theory. Rather than ordering the heaviest evidence for everything, the workflow asks what theory the current facts support, then spends accordingly. Preservation comes first, because listing state changes fastest. A targeted test purchase is committed only where it will change the outcome, converting an unbounded investigative cost into a targeted one.
Route by legal theory, not by habit. Once classification is explicit, the route follows: a counterfeit report, a channel or contract action, a material-difference review, or monitor-and-close with a recorded reason. Filing everything through the complaint type a team knows best accumulates rejections in categories it never needed to enter.
Keep the disputed call with counsel. Ambiguity is not a failure state to automate away. Where evidence supports more than one reading, the matter should reach a lawyer with the record complete and the open questions named. Evidence discipline is what makes that handoff usable, as set out in How to Preserve Online Trademark Evidence Before Takedowns.
What value correct classification delivers
It converts an individual judgment into an organisational capability. When the reference set, evidence standard, and routing rules are defined once, similar facts reach reviewers in comparable form and produce comparable decisions. Discretion survives where it matters; accidental variation does not.
It protects the standing that buys speed. Platforms assess brands over trailing windows, and accuracy records govern access to self-service removal and machine-learning protections that act before a human files anything. Precision compounds quietly; so does carelessness.
It reduces exposure, not just effort. Asserting counterfeiting without a basis invites seller appeals, platform scrutiny, and, in copyright-adjacent matters, misrepresentation risk. Classification-first workflows catch those filings before they leave the building.
It reallocates legal capacity. Hours spent reconstructing provenance and comparing label variants are hours not spent on repeat offenders, negotiation, or escalation strategy.
It makes enforcement memory permanent. A seller treated as an isolated listing and one recognised as a recurring actor are different business problems, and only the second justifies escalation — the argument behind How to Identify Repeat Counterfeit Sellers Across Marketplaces: Turn Takedowns Into a Defensible Enforcement System.
It makes the decision explainable after the fact. When the call is questioned, the team can show what it observed, what it compared against, which theory it applied, and who approved it. That is the difference between a programme leadership can budget confidently and one that must be defended defensively.
How CourtifyAI solves the same class of problem
CourtifyAI treats counterfeit vs gray market goods classification as infrastructure, not a preliminary step squeezed in before the real work. CourtifyAI's AI legal assistant for evidence-backed legal work — the Copilot product — is built for the analysis around ambiguous matters: assembling case material, extracting the factual elements that determine which theory fits, mapping facts against rights and rules, and producing reviewable drafts whose citations a lawyer can check rather than trust.
Auto Pilot carries that discipline into automated IP enforcement. It monitors across channels, preserves the observable record with timestamps and hashes so it survives later scrutiny, maintains the rights and reference information the comparison depends on, organises matters into reviewable units, and tracks what happened next — platform responses, relists, repeat actors. Suspected listings surface with their distinguishing attributes attached, so the first legal call is made once, on a complete record, instead of being re-litigated case by case. Strategy and final decisions remain with the legal team; what changes is that the foundation is never rebuilt from scratch.
Frequently Asked Questions
Is it counterfeit or gray market if the seller is unauthorized?
Authorisation alone does not decide it. Genuine goods sold without authorisation are usually a channel or contract issue, because first sale generally exhausts trademark rights over that resale. It becomes counterfeiting only if the goods are not genuinely the brand's, or if materially different goods carry the same mark.
How do you prove a product is counterfeit and not gray market?
Compare the item against the brand's genuine product for that market using a documented sample — labels, packaging, markings, materials — plus provenance evidence where available. Documenting how the sample was obtained and handled matters as much as the comparison.
What is the difference between counterfeit and gray market goods?
Counterfeit goods are not the brand's products and bear a spurious mark identical or substantially indistinguishable from a registered one. Gray market goods are genuine products sold through unauthorised channels, often parallel imports, and become actionable only in specific circumstances such as material differences from the authorised version.
References
Sources: the federal definition of a counterfeit mark under 18 U.S.C. § 2320 and U.S. Customs and Border Protection's rules on restricted gray market articles under 19 CFR 133.23.
This article provides operational information, not legal advice. Enforcement standards and platform procedures vary by jurisdiction and marketplace.