Counterfeit Risk Scoring: An Analytical Model for Counterfeit Products and Unauthorised Sales

Digital counterfeiting risk is not just a question of whether a product is fake. Assessing marketplace listings, seller behaviour, price anomalies, image similarity, social media and fraudulent domains together can help brand teams prioritise risks more meaningfully.

The Counterfeit Risk Scoring approach in this guide is not the result of a published global B10 dataset. It is an example analytical model for digital brand protection operations.

What Is Counterfeit Risk Scoring?

It prioritises listings and sellers for manual review by combining different digital signals into a common risk score.

Signals That Can Be Used in the Risk Model

Similarity between product and catalogue images,
Use of brand and model names,
Price anomalies,
seller history,
A newly created account or store,
Risky wording in the product description,
Links to previous infringements,
similarities in domains or social media accounts.

Price Anomalies

Listings priced well below authorised channels are not automatically counterfeit. Combined with other signals, however, a low price may increase review priority.

Seller Risk Profile

For each seller, assess:

store age,
Product variety,
Pricing behaviour,
Recurring trademark use,
Previous risk records

These factors can be assessed.

Visual Similarity

Perceptual hashing or embedding-based image models can help detect catalogue photos reused in different listings.

An image match is not a legal determination of infringement; it is only a risk signal.

Logos and Brand Elements

Logo detection, OCR and image analysis can provide extra signals when the brand name is absent from the listing text.

Monitor Marketplaces and Social Media Together

Counterfeiting or unauthorised sales may extend across several channels. A seller may direct traffic from a social media account to a marketplace or an independent online store.

Domain Risk

Typosquatting, fake campaign pages and domains resembling the brand name can be tracked as a separate risk category related to product infringements.

An Example Risk Score

The model can combine the following factors:

Visual similarity,
Price anomalies,
seller trust score,
Intensity of trademark use,
Previous infringement history,
cross-channel links.

Weights do not have to be identical for every brand and category.

Why Is the Score Not Universal?

Counterfeiting signals may differ across luxury fashion, electronics, cosmetics and spare parts. Calibration by brand and category is more appropriate than a single global threshold.

Managing False Positives

An authorised seller may use catalogue photos, and a genuine product may sell below market price. The system should not make the final decision; it should send high-risk cases for expert review.

Evidence Package

For the record under review, the following can be retained:

URL,
Seller,
listing title,
Price,
Image,
Screenshot,
Date and time,
Reasons for risk

These details can be retained.

How Is the Risk Map Used?

A risk map can be prepared by country, channel, category or seller where actual measurement data are available. Example scores produced without data must not be presented as real market statistics.

The B10 Approach

B10 Digital Agency aims to prioritise different digital signals through risk scoring in brand protection operations and use them together with expert verification.

Methodology Note

This model is an illustrative analytical framework. It is not presented as published empirical B10 research on a specific country, sector or marketplace.

Conclusion

A counterfeit monitoring system is valuable not for impressive maps, but for identifying risky records needing review earlier across high-volume digital channels.