Help Understanding your results
Categories, findings and answers
Follow each category comparison back to the questions and individual answers that support it.
Compare categories
Categories compared shows the questions, usable answers and brand mentions for each category, together with the leading comparison brand when available. Read both the combined result and the breakdown by measured AI service. A result from one service does not establish the same pattern in another.
The denominator is usable answers, not all planned attempts. A brand can appear alongside several competitors, so their shares need not sum to 100%. The category labels distinguish competitors ahead, a close comparison, the brand ahead and little mention of either. They describe this sample, not market share or a product-quality ranking.
Below 30 usable answers the category is marked Few data. Such questions remain inspectable, but do not receive the same detailed interpretation as categories with enough data.
Read a finding
A finding compares mention counts, explains the observed pattern and shows an available action. A displayed excerpt comes from a measured answer and identifies its question and AI service. It is selected evidence, not the full answer corpus, and can be absent when no suitable excerpt fits.
Read the question before acting. An action may suggest an own-site format or an external domain. A suggested destination is not evidence of permission to publish, acceptance by that site or a future AI citation.
Read the answer matrix
Each row is a question; marks under each AI service represent individual samples. Distinguish an answer naming the brand, an answer not naming it and no answer. The last state is not a measured failure to mention the brand and is excluded from usable-answer shares.
An AI Overview not appearing for a query is separately explained in the reading guide. Do not assume that every empty cell is a negative mention result. Questions with insufficient evidence or changed historical wording can appear under Other questions instead of supporting a current category finding.
Keep questions naming a company separate
When the question names a company contains questions naming your brand or a tracked competitor. These are excluded from the open-question category comparison: explicitly supplying a name changes what is being tested.
Open Prompts for full answers and history. A saved PDF keeps its recorded wording; editing a prompt later does not make an older answer evidence for the new wording.