Alert quality
How to reduce false positives in trademark watching
Reducing noise does not mean raising a threshold until the inbox is empty. The goal is to remove repetitive, irrelevant hits while retaining controls that reveal what exclusions might hide.
Measure before filtering
For a defined period, label results as relevant, uncertain or dismissed and record why. Without that sample, adjustments rely on impressions and cannot show whether precision improved or volume merely fell.
Group noise by rule, term, class, source and cause. A small number of queries or expansions often creates most false positives and offers the safest starting point.
Make exclusions specific and reversible
Exclude known owners, descriptive terms or exact combinations only after confirming a pattern. Retain the reason, author, date and version of every exclusion so it can be reviewed.
Avoid portfolio-wide negative lists. A term that is irrelevant for one mark may be central to another, and a shared exclusion can create a blind spot that is hard to discover.
Use goods and services carefully
Classes help narrow results but do not replace goods-and-services analysis. Maintain reasoned relationships between nearby terms and periodically confirm that they still reflect protected activity.
A literal filter misses equivalent wording; linking whole classes may flood review. Test the balance against historical results and a recent sample.
Control for false negatives
Sample automatically dismissed hits and run reference queries against known signs. If an expected rule stops retrieving them, investigate before trusting the lower volume.
Track precision, volume per rule, review time and recovered exceptions. Reopen rules after changes to the portfolio, languages, goods, sources or team criteria.
Official sources
Frequently asked questions
What is a false positive in trademark watching?
It is a hit retrieved by a rule that, after review, does not warrant monitoring or action for the defined objective.
Does raising the threshold always save work?
It reduces results but may hide useful hits. Validate it with samples from both sides of the cut-off.
How often should exclusions be reviewed?
Review them when portfolio or coverage changes and on a regular cycle informed by volume, dismissals and detected exceptions.