Sentiment

Being named is not the same as being recommended. An answer can list you as the budget option, warn that your support is slow, or call you the best tool for small teams. Sentiment measures that tone, so you can tell a mention that helps from one that hurts.

Sentiment vs brand perception. Sentiment is how warmly AI talks about each brand, scored from 0 to 10 and split by topic. Brand perception is why: the attributes AI describes the brand on, the quotes behind them and the sources it cited.

---

How sentiment is scored

Every time an answer names a brand, GEOflux reads how the answer talks about it and scores that mention from 0 to 10:

A brand's sentiment is the average of its mention scores in the current view. A score of 7.4 means the answers that name the brand lean mildly positive overall; it is not the grade of any single answer.

• Zones. Below 4 is negative (red, frowning face), 4 up to 7 is neutral (amber), 7 and above is positive (green, smiling face). Charts shade these zones behind the line, with a legend beside them.
• Every brand is scored. Each brand an answer names gets a score, yours and every competitor's, tracked or not. Any brand's sentiment is on the Brands tab and its brand profile.
• What tracking changes. The Sentiment tab compares your brand with the competitors you track. Tracking a brand adds it to the Brand filter, the Per-brand bars and the All brands average, with its earlier answers included, since those were already scored.
• All brands. With All brands selected, sentiment is the average across your brand and your tracked competitors, weighted by how often each is named, so the most mentioned brands move it most.
• Per answer. On the Answers page each answer shows the score for the brand in view: yours, or the one brand you filtered to. Hover the face for the exact value, for example "Positive sentiment, 8/10". The Answers Sentiment filter uses this same score.
• No mentions, no score. A brand with no mentions in the current view has no sentiment. It shows as No sentiment recorded (a grey face) or "--", never as a number, and is left off the sentiment bars.

A brand that is only listed, with nothing said about it, scores 5. So does a mention whose tone could not be read. A cluster of 5s usually means "named in a list", not "described as average": open the answers to tell which.

---

Where you read it

• The Sentiment tab on the Overview:
  • The lead shows your sentiment, your rank and the best rated brand. Under All brands it shows the best rated brand, the market sentiment, and your lead or your gap to the best.
  • Sentiment Score compares every brand (Per-brand bars) or shows your score over time (Sentiment trend).
  • What's driving this lists the attributes pulling your sentiment down and holding it up.
  • Sentiment by topic shows which topics read well or poorly, as a treemap or a list. It is built from the topic on each prompt, so how you set topics decides what it can show. See Why topics matter for sentiment.
  • Brand Perception breaks your sentiment down attribute by attribute. See Brand perception.
• Rankings on the Visibility and Share of Voice tabs, the prompts table, prompt detail pages, brand profiles and source pages all carry a sentiment column or score.
• Alerts can email you when sentiment falls below a floor. See Alerts.

---

Why it matters

A brand can have high visibility and still lose: if AI names you mostly to say you are expensive, every mention pushes buyers elsewhere. Sentiment is also where AI is most often out of date. An answer may still describe the pricing you changed a year ago, or a problem you fixed, because that is what the sources it reads still say.

---

Acting on sentiment

1. Find the cause. Start with What's driving this on the Sentiment tab, then Sentiment by topic to see where it happens.
2. Read the answers. On the Answers page, filter Sentiment to Negative and read what is actually said, on which engine, for which prompts and personas.
3. Find the source. Each answer lists the pages it cited. An outdated review, a forum thread or a comparison page is usually behind a persistent negative description.
4. Fix what you can reach: update your own pages, correct third-party listings, publish the answer the engines are missing, and work on the sites that keep getting cited.
5. Watch it recover over weeks, with a chart note on the day you made the change.

Compare yourself with competitors rather than with an absolute target. In many categories AI is mostly neutral about everyone, and a 6.5 that leads the field is a good score.