Glossary
False Positives
False Positives in Brand Protection: Causes and Fixes
A false positive is a result that a detection system flags as an infringement but that turns out to be legitimate, such as a listing from an authorised reseller, a genuine product sold by a private person, or a fan page. In brand protection, too many false positives waste review time and risk wrongful takedown notices.
Every detection method produces some false positives. The goal is not to remove them entirely but to keep them low enough that enforcement stays fast, accurate and defensible.
What Are False Positives?
Detection tools work by matching patterns: a brand name in a title, a similar image, a price below the usual range. A pattern match is only a hint. When the match is wrong, the result is a false positive. The opposite error, a real infringement the system fails to flag, is a false negative.
The two errors pull against each other. Tuning a system to catch more infringements (higher recall) usually lets in more false positives. Tuning it to be stricter (higher precision) can miss real cases. Good brand protection finds a balance and uses human review where the cost of a mistake is high.
False positives differ from borderline cases. A dupe or a look-alike product may be legal and still worth watching. A false positive is simply wrong: the item is not what the system said it was.
Common Causes and Examples
- Authorised sellers flagged as unauthorised. A distributor lists genuine goods under the brand name. If the system does not know the brand's authorized resellers, it may flag them.
- Genuine resale. A shopper sells a used original item. Under rules such as the first sale doctrine in the United States and exhaustion of rights in the EU, resale of genuine goods is often lawful.
- Descriptive or fair use of a name. A repair shop writes "fits Brand X watches" in its listing. That mention is not necessarily trademark infringement.
- Image reuse by partners. An affiliate uses official product photos with permission, and reverse image search flags it.
- Name collisions. A different company with a similar name, in an unrelated field, appears in keyword results (see likelihood of confusion).
- Outdated lists. A reseller's contract was renewed, but the allow list was never updated.
Why It Matters for Brands
- Wasted effort. Analysts who review hundreds of harmless results have less time for real threats.
- Damaged partnerships. A takedown notice sent to an authorised distributor can strain a business relationship.
- Legal risk. Sending notices without a good-faith basis can create exposure, for example a false DMCA claim in copyright cases. Platforms may also restrict accounts that submit inaccurate reports.
- Lower credibility. Marketplaces and hosts take reports from accurate, well-documented senders more seriously.
- Wrong metrics. Counting false positives as "found infringements" makes results look larger than they are.
How to Reduce False Positives
- Maintain allow lists. Keep a current list of authorised sellers, distributors, affiliates and your own channels.
- Combine signals. A brand name alone is weak. Name plus unusual price plus new seller plus stolen images is stronger. Risk scoring does this systematically.
- Add human review. Have a trained person check cases before action, especially for high-value sellers or unclear cases.
- Check the legal basis. Confirm the brand holds the right being claimed, for example a registered trademark, in the relevant country.
- Use test purchases. For high-impact cases, buying the item removes doubt about authenticity.
- Feed back decisions. Use reviewer decisions to tune rules and models so the same mistakes are not repeated.
- Track precision. Measure what share of flagged items turned out to be real, and watch the trend.
Related Terms
- Risk Scoring: ranking flagged items by how likely they are to be infringing.
- Infringement Detection: finding violations online.
- Keyword Monitoring: searching for brand-related words.
- Authorized Resellers: partners allowed to sell the brand's goods.
- Evidence Collection: documenting a case before enforcement.
How Remove.tech Helps
Remove.tech uses AI to detect suspect listings, accounts and sites, and a review step to filter out legitimate ones before any notice is sent. You can share your authorised sellers so that partners are not targeted, and confirmed cases are documented before a takedown. To see how it works for your brand, get started.

