Answerability: A score (0-100) measuring how cleanly an AI can extract an answer to a specific question from your page. High information density with low answerability means the content is all there but the page's structure prevents the AI from confidently using it. Answerability carries the second-highest weight (22.9%) in the overall optimization score.
Mixed Identity Page: A page that tries to serve two purposes at once, such as being both a fair comparison article and a product showcase. The structural inconsistency (one entry dramatically longer, headers shifting from editorial to promotional language, tone changes) signals to AI engines that the page isn't a clean, trustworthy source for the question type it should be answering.
The Briefing Test: Before running the optimizer, open your page, read the prompt you want to win, and ask: "If I found this page while writing a briefing, would I cite it as a trusted source for answering this specific question?" This editorial instinct is the skill; the tool makes it precise.
The optimizer scrapes your page, pulls the pages currently winning citations for that prompt, and compares them head-to-head. You can view the workflow to see exactly what it's doing.
The target zone is 65-82%. Below the overall score, four sub-scores tell the real story:
Look for the combination of high weight + low score first. That's your priority.
Each recommendation maps to a score category with a predicted impact. Do not treat them as a checklist. For each one, weigh:
Shows the pages you were compared against, the ones currently winning citations for your target prompt. These are the same pages from your citation supply chain data in the Analyze section. The optimizer is doing automatically what you learned to do manually: comparing your page against the competition that's beating you.
The most common reason a page has high information density but low answerability:
The fix: commit to the page's primary identity. A comparison article that over-indexes on one product doesn't feel like a trustworthy comparison, to readers or to AI engines.
