How should I define website sections to measure?
Website-section measurement starts with a fixed map of meaningful page groups, not a single domain-wide visibility score. Group pages by the job they perform for buyers, such as product pages, solution pages, comparison pages, documentation, pricing, case studies and the company blog. Use URL folders, page templates or a manual page list, depending on how consistently the site is structured.
Keep sections separate when they serve different search intents or have different owners. A product section should not be blended with documentation merely because both support the same product. A domain average can hide the fact that one section is frequently cited while another is absent from answers.
Record the rules for each group before collecting results. Note included URL patterns, excluded pages, redirects and pages that belong to more than one commercial area. If a page could fit two sections, assign one primary section for reporting and record any secondary role separately. Consistent classification matters more than creating a large number of small groups.
Choose the buyer questions for each section
Each website section needs a prompt set based on the questions its pages should answer, rather than a shared list copied across the whole domain. Product pages might receive questions about use cases, alternatives and limitations. Documentation might receive setup and troubleshooting questions. Comparison pages might receive evaluation questions where competitors appear naturally.
Use the same intent categories for every section so results remain comparable. Include discovery, evaluation, implementation and follow-up questions where the section genuinely supports them. Keep brand names out of some prompts when you want to test whether an engine introduces the company without being asked.
A prompt belongs to a section when a reasonable answer should cite or mention at least one page in that group. Do not assign a question to a section merely because the section owner wants visibility there. The page must have a legitimate role in answering the question. Cituna's separate article on multi-website AI visibility tracking is useful when the same measurement framework must cover more than one domain.
Set the expected answer page for every prompt
Every tracked prompt should have an expected answer page or an explicitly accepted page set before results are reviewed. The expected page is the URL that best answers the question today, while the accepted set covers cases where several pages could reasonably be cited. This distinction separates a section problem from an internal linking or content-routing problem.
Record the expected page, section, intent, product area and language for each prompt. Mark prompts that should produce a direct recommendation, a comparison, a definition or a procedural answer. A page can be visible for a prompt and still be the wrong page if an engine cites a weaker article while ignoring the intended source.
Do not force one page to answer every question in a section. If the expected page changes after a content revision, preserve the old assignment and date the new one. Otherwise, an apparent visibility improvement may only reflect moving the goalposts. Website redesigns deserve special care because URL changes can make section comparisons look like content gains or losses.
Capture a baseline across all seven engines
A useful baseline records whether ChatGPT, Perplexity, Gemini, Claude, Grok, Google AI Overviews and Google AI Mode mention the brand, cite a page and cite the intended section. The seven engines can produce different answers for the same prompt, so a single engine cannot represent overall AI visibility.
Capture the prompt, date, engine, answer text, cited URLs, brand mention, cited section and expected page. Preserve the answer or a reliable record of its citations because outputs change. Treat a brand mention without a source citation differently from a cited page, and treat a citation to another section differently from no citation.
Cituna tracks whether those seven engines mention and cite a brand for buyer questions every day, and shows which competitors and pages they cite instead. That gives a recurring baseline, but the measurement logic still depends on the section map and expected-page assignments. Record engine-specific results before combining them into a summary, so a strong result in one engine does not conceal a complete absence in another.
Separate mentions, citations and correct-section citations
Section-level AI visibility needs three separate outcomes: the brand was mentioned, a website page was cited, and the cited page belonged to the intended section. Combining those outcomes into one score hides the reason a section is underperforming.
A mention shows that the engine knows the brand in the context of the answer, but it does not prove that the website influenced the answer. A citation shows that a page was used or presented as support, but it may point to a page with a weaker commercial or informational purpose. A correct-section citation shows that the site is routing attention to the page group designed for that question.
Report results by section and engine before calculating an overall view. A section with many mentions but few citations may need clearer evidence, accessible source material or stronger page alignment. A section with citations to the wrong page may need internal linking, canonical checks or a clearer division between overlapping content. Cituna shows which competitors and pages engines cite instead, making the alternative source part of the diagnosis rather than an afterthought.
Join AI answers to Search Console evidence
Google Search Console adds search demand and page performance context to AI visibility results, but it should explain a section rather than replace direct answer checks. Join each expected or cited URL to impressions, clicks, queries and landing-page data for the same reporting period.
A section with strong traditional search activity but weak AI citations may already contain useful demand signals and need clearer answer extraction or authority cues. A section with weak search activity and weak AI visibility may have a broader discovery or content-fit problem. High AI citation activity with little Search Console activity is not automatically a failure, because answer engines and Google web search do not expose identical demand.
Check whether cited pages are indexed, canonical, internally linked and accessible to crawlers. Compare the cited URL with the page receiving search impressions. When an engine cites an old article instead of the intended section, inspect duplication and topical overlap before creating more content. Cituna joins AI answers to Google Search Console data and provides SEO, AEO and GEO fixes, while the underlying decision remains which evidence supports the next change.
Compare sections before choosing the first fix
Prioritize the section with a clear buyer need, a credible expected page and a measurable failure, rather than the section with the lowest raw score. Rank problems by the gap between expected and observed outcomes: no mention, mention without citation, citation to the wrong section, or citation to a competitor instead.
A low-visibility section is not always the best first target. A small section with a few high-value questions may deserve attention before a large blog section with many low-intent prompts. Conversely, a section with broad demand but scattered citations may benefit more from consolidation than from publishing another page.
Use a decision rule that combines business importance, prompt coverage, page readiness and the likely size of the fix. Repair missing or unsuitable destination pages before judging the section on citation performance. Improve existing pages when engines cite nearby material but not the intended URL. Create new content only when the prompt has no suitable answer on the site. The most useful comparison is between sections facing the same prompt intent, not between unrelated page types.
Repeat the measurement and report the change
Repeat the same section-level measurement after a defined content or technical change, keeping prompts, engine names, section rules and expected pages stable. Stable inputs make it possible to distinguish a real change from a different test set.
Review movement by engine and outcome, not only by a combined domain score. Ask whether more engines cite the intended section, whether competitors are still cited, whether citations moved to a better page and whether Search Console signals changed for the affected URLs. Record publication dates, redirects, internal-link changes and major technical releases beside the results.
Report one baseline, one change and one next decision for each important section. Avoid claiming that a page caused an answer change when several edits happened together or when the engine output changed without a site change. A section dashboard should preserve raw observations as well as summaries. Cituna includes Search Console and an MCP server from its entry plan, while API access is available on its Max plan, so teams should choose an export or review process that preserves comparable records over time.
Related reading
Sources consulted
- Google Search Console Help (support.google.com)
- OpenAI Platform Documentation (platform.openai.com)
Drafted with AI assistance from our own research and Search Console data, and reviewed by Rahul A before publishing. Rules and prices change; check the linked official source before you act.