Published Growing Search case
AI mentions and sentiment for a luxury brand
Track inclusion and the description attached to a brand as separate observations.
Results reported by Growing Search
- AI Overviews mentions
- +9%
- ChatGPT mentions
- +11%
- Sentiment score
- 67 to 74/100
Three months
Three months
Published scale, over three months
The question
Does the answer include the brand, and what kind of judgment does it make?
Inclusion and description as separate decisions
Classify a saved answer in two dimensions. Brand presence alone does not determine the reason for a recommendation.
Check inclusion first
Record whether the brand is named in the saved response. Leave the description field empty when the brand is absent.
Identify the stated reason
Keep the product distinction or price context given in the answer. A favorable adjective without a reason needs a different category from a qualified recommendation.
Verify the underlying distinction
Ask the business to confirm the relevant product facts and comparison scope. Retain an accurate limitation rather than rewriting it as praise.
Apply the agreed description rule
Classify the saved explanation with a repeatable rule. Keep the underlying words so another reviewer can assess the same decision.
| Published action | Recommended check | Record to retain |
|---|---|---|
| Competitor questions | Keep the compared options and the reason for each answer together. | Question, compared set, answer and product distinction. |
| External coverage | Approve the factual difference the proposed piece should explain. | Product evidence, price context and editorial purpose. |
| Answer state | Suggested record | Review decision |
|---|---|---|
| Brand absent | Inclusion absent, description empty. | Do not assign sentiment to an omitted brand. |
| Brand named without evaluation | Inclusion present, descriptive mention. | Retain the surrounding wording. |
| Recommendation with a reason | Reason and conditions saved. | Verify the product distinction. |
| Criticism or qualification | Objection and scope saved. | Assess the factual basis before a response. |
Category and comparison coverage
Category and price answers omitted an unnamed online luxury brand. Growing Search reviewed competitors and produced external articles. It reports the three-month results above, with sentiment scored out of 100.
Separate records for inclusion and description
Record inclusion before classifying the description in a new observation plan. A brand may appear more frequently while the reasons given for recommending it remain unchanged. Keeping the two observations separate makes the proposed editorial work easier to evaluate.
Write down how each description will be classified. Save the underlying answer and the product question, rather than retaining only a score. A score change is difficult to review if another team member cannot identify which language or claim justified the classification.
Evidence for price and product fit
Ask which product distinctions a comparison needs to explain. Confirm the price context and relevant product facts with the business, then give the writer that evidence. A positive adjective without an explanation of product fit adds little for a person choosing between brands.
- Keep category-selection questions distinct from direct competitor comparisons.
- Preserve the question wording and dated answers before changing the observation set.
- Use the same classification rule when reviewing later answers.
Application across markets
The related Spanish-search project concerns the same brand and a different measurement. Keep that market-specific search record separate from AI mentions. For a new project, the AI Overviews guide can establish Google's page checks before coverage work is commissioned.
Prepare a check for your own site
Choose the checks that fit your site and add the baseline you will use. The download includes this analysis, related guides and the original Growing Search publication.
Findings
Use the rubric to explain a classification, without treating it as a purchase forecast. If a reviewer changes the rule, preserve that version and recheck the earlier sample before comparing totals.