Skip to content
All terms

Glossary

Risk Scoring

Risk Scoring Explained: Prioritising Brand Threats

Risk scoring is the practice of giving each suspected infringement, seller or website a numerical score that reflects how likely it is to be harmful and how much damage it could cause. In brand protection, scores let teams sort thousands of detections and deal with the most serious threats first.

Without scoring, a team facing a long list of flagged items has to read them in the order they arrived. With scoring, effort goes where the impact is largest.

What Is Risk Scoring?

A detection system finds many suspect items, but they are not equal. A single listing for a used genuine product is very different from a fake shop that collects card payments and sells nothing. Risk scoring combines several facts about an item into one value, often a number from 0 to 100 or a label such as low, medium and high.

There are two ideas folded into one score:

  • Likelihood: how likely is it that the item really is an infringement or scam?
  • Impact: if it is, how much harm does it do, based on reach, sales volume, safety and the victims involved?

Scores can come from fixed rules (for example, add points if the price is far below normal), from a machine learning model trained on past reviewed cases, or from a mix of both. A score is a prioritisation aid. It is not a legal finding, and it does not replace review.

Risk scoring is related to infringement detection, which finds suspect items, and to seller clustering, which groups related sellers. Scoring tells you what to look at first.

Typical Signals

Common inputs to a score include:

  • Price gap: how far the price is below the usual market or recommended price.
  • Image and text match: closeness to official photos, logos and descriptions (see reverse image search and keyword monitoring).
  • Seller profile: account age, number of listings, ratings, location, whether the seller is known as an authorized reseller.
  • Seller history: earlier confirmed violations or links to known bad actors.
  • Volume and reach: units sold, views, followers, ad spend.
  • Risk category: products with safety risks, such as cosmetics or electronics, may rank higher.
  • Website signals: domain age, hosting, lookalike domain patterns, missing legal information.

Example

A sportswear brand receives 2,000 new detections in a week. Most are second-hand resale or small accounts. Risk scoring places at the top a new seller with copied product photos, prices well under half the normal price, a thousand units sold and links to a network of similar accounts. The team reviews those first and sends notices within hours, rather than days.

Why It Matters for Brands

  • Efficiency. Reviewers focus on the cases with the highest impact.
  • Consistency. Clear criteria make decisions repeatable between team members and over time.
  • Fewer errors. Low-scoring items can be monitored instead of reported, which reduces false positives.
  • Better reporting. Scores help show management where the largest threats are, and whether they are shrinking.
  • Cost control. Enforcement budget is spent where it makes a difference.

How to Use Risk Scoring Well

  1. Define what matters. Decide with legal, sales and customer service which harms count most: lost revenue, safety, brand image or fraud against shoppers.
  2. Pick transparent signals. You should be able to explain why an item scored high.
  3. Set thresholds and actions. For example: high scores go to immediate review and notice, medium scores to review within a week, low scores to monitoring only.
  4. Keep humans in the loop. Review before enforcement, especially near thresholds.
  5. Recalibrate. Compare scores with real outcomes and adjust. Sellers change their behaviour, so signals age.
  6. Document decisions. Keep evidence for every case that moves to action.

How Remove.tech Helps

Remove.tech detects suspect listings, sites and accounts across marketplaces, social media, app stores and search engines, and ranks them so the most harmful cases are handled first. Confirmed infringements are taken down through notices we send on your behalf. See our brand protection overview or get started.

Get started

Join Remove.tech,like other leading brands.

Or contact us by email: hello@remove.tech