How is media sentiment measured?
Each piece of coverage is classified by tone: positive, neutral or negative. Media monitoring tools automate this classification with sentiment analysis, a technique that extracts the opinions expressed in a text (Wikipedia, “Sentiment analysis”).
Automation has its limits. It struggles with irony, ambiguous headlines or a critical article about a competitor that cites your brand as a positive example. A human review of the most important articles remains necessary.
Why is a neutral article often worth more than a glowing one?
A factual article in a reference outlet carries more weight than an enthusiastic profile in a niche one. A neutral tone is the norm in journalism: a reader, an investor or an AI reads it as reliable information. Sentiment should therefore always be read together with the outlet's credibility and audience.
Do AI tools have an opinion about a brand?
AI tools do not form opinions, but their answers reflect the tone of the sources they read: articles, customer reviews, forums, comparisons. One highly visible negative article can colour hundreds of answers.
Tracking tools assign a sentiment score to answers. On the panel Linker tracks, its score was 68 out of 100 at the end of September 2026 (Peec AI), within the range the tool considers healthy, between 65 and 85.
What should you do if sentiment deteriorates?
- Identify the source: an article, a review, a discussion thread. The problem often comes from a limited number of pieces of content.
- Respond on the facts: correct an error, request a right of reply if the information is inaccurate.
- Create new sources: coverage, case studies and expert commentary that give more accurate information.
- Plan ahead: a crisis communication plan avoids improvising.
To go further
Media sentiment is a pillar of online reputation and a key metric in crisis communication. When information is wrong: right of reply, errors and false information in the media.
Last updated: October 2026
