Check this page with an assistantOpens a chat asking it to summarise this article and name the evidence behind each claim.
Claude opens with the prompt on your clipboard: Anthropic does not support prefilled prompts on the web, and we would rather copy it than ship a button that drops it.
Start With the Deliverable, Not the Crawl
A useful visibility audit ends with five to twelve actions ordered by dependency and expected consequence. Each action names the affected page or template, the observed evidence, the desired outcome, the person who can change it, and the date you will check again.
Most audits become unmanageable because they collect facts without making decisions. A missing meta description, a blocked template, and a service page that does not exist all become equal rows in a spreadsheet. They are not equal. The blocked template can suppress an entire section; the missing service page prevents relevance from existing; the description may not be used by Google at all because snippets are primarily selected from page content.
Open one working document with these columns: finding, evidence, affected URLs, dependency, next action, owner, and check date. Do not add a finding until you can fill the evidence and affected-URL columns. This rule removes most generic recommendations before they enter the queue.
You can explain why every item is in the audit and what observable state would let you close it.
Minutes 0–15: Can Search Systems Reach the Important Pages?
Check the homepage, one commercial page, one article, and one page created by a different template. Use Search Console URL Inspection for Google, Bing URL Inspection for Bing, and a normal browser with JavaScript disabled long enough to confirm that the primary text and links are still present in the HTML or rendered output.
- Read robots.txt as a policy file.
Confirm that important paths are not disallowed. A robots.txt block controls crawling; it is not a reliable method for removing an already known URL from search.
- Inspect the response, not just the screen.
Priority pages should return a successful status without a login, JavaScript challenge, redirect chain, or region rule that blocks search crawlers.
- Follow a real internal link.
Google documents standard anchor elements with an href as the reliable crawlable pattern. Button-like divs and script-only navigation create avoidable discovery risk.
- Compare templates.
One healthy homepage proves very little. Navigation, canonical tags, robots directives, and structured data often differ by CMS template.
If a priority template is inaccessible, stop the audit and record it as the first dependency. Google says pages eligible for AI Overviews or AI Mode must first be indexed and eligible to appear in normal Search with a snippet. Discovery is the gate before optimization.
Minutes 15–30: Which URL Is Each Page Asking Engines to Keep?
For every sampled page, verify four signals together: indexability, final status, canonical target, and sitemap membership. The desired state is usually a 200 response, no noindex directive, a self-referencing canonical, and inclusion in the XML sitemap.
Canonicals are hints, not commands, and conflicting hints create unnecessary ambiguity. A page that canonicals to URL A while the sitemap lists URL B and internal links point to URL C is asking the engine to resolve a disagreement the site created. Consolidate the signals before diagnosing rankings.
| Observed state | Interpretation | First check |
|---|---|---|
| Discovered, not indexed | Google knows the URL but has not selected it | Duplication, page usefulness, internal links, server reliability |
| Crawled, not indexed | The page was fetched but not retained | Near-duplicates, thin template output, soft-404 signals |
| Alternate page with canonical | Another URL was selected as the representative | Whether consolidation is intentional and signals agree |
| Blocked by robots.txt | The crawler cannot fetch content or read page-level directives | Whether the block is intentional for this exact path |
Do not turn indexed-page count into a vanity goal. The objective is for every page that earns a distinct job to be selectable, and for duplicate or low-value states not to compete with it.
Minutes 30–45: Where Is Existing Demand Meeting the Wrong Page?
Open Search Console’s Performance report and compare the last 28 complete days with the previous 28. Focus on clicks and impressions first; Google’s own documentation recommends reading trends more heavily than average position alone. Position is an average of the topmost result and changes with query, device, location, and search feature.
- Find high-impression queries in average positions 4–15. These are evidence that Google already associates the site with the demand.
- Find pages with falling clicks and impressions across comparable periods. Separate a demand decline from a page-specific decline.
- Filter one important query, then open the Pages dimension. If two pages trade visibility, inspect intent and duplication before calling it cannibalization.
- Separate branded from non-branded demand where the report supports it. Brand growth and discovery growth answer different questions.
Export the query view if the list is long, then use the free Search Console opportunity finder to create a first-pass triage. It compares CTR only with similar-position rows in your own export instead of pretending one universal click curve applies to every search result.
Minutes 45–60: Does the Result Make the Right Promise?
For the highest-impression pages, compare four things side by side: the search query, title element, visible H1, and opening answer. The promise should be consistent without becoming repetitive. Google may create title links from several page signals and usually creates snippets from visible content, so the page, not one metadata field, is the real unit of search appearance.
- Make the title unique, concise, and accurate. Remove repeated boilerplate that pushes the useful distinction out of view.
- Put a direct, qualified answer near the top when the query asks a question. Do not delay the answer with a generic market preamble.
- Check the mobile result and mobile page. Search journeys are not audited on a desktop screenshot alone.
- Use PageSpeed Insights to inspect field Core Web Vitals. Current “good” targets remain LCP within 2.5 seconds, INP under 200 milliseconds, and CLS at or below 0.1 at the 75th percentile.
A low-impression page for CTR before confirming that the page answers a real query and is eligible to appear. Appearance work cannot repair missing relevance.
Minutes 60–75: Are You Discoverable Across AI Search Surfaces?
Treat AI discovery as an extension of the same eligibility system, with crawler-specific access checks. Google says there are no extra technical requirements or special AI schema for AI Overviews and AI Mode. OpenAI says public sites can appear in ChatGPT search and specifically advises allowing OAI-SearchBot through robots.txt, the host, and the CDN or firewall.
- Confirm Google eligibility.
The page must be indexed, snippet-eligible, internally discoverable, and useful in visible text. That is the same foundation used by Google’s AI features.
- Confirm OAI-SearchBot access.
Check robots.txt and bot-protection rules. A crawler allowed by robots.txt can still receive a 403 from a WAF or JavaScript challenge.
- Check Bing’s web and chat reporting.
Bing Webmaster Tools includes web and chat sources in its performance reporting. Use the direct measurement available before buying a modeled visibility score.
Then inspect the content itself. Important facts should be present in readable text, definitions should be explicit, comparisons should include tradeoffs, and evidence should link to the original source. This is not special “AEO formatting”; it is a practical consequence of making each section understandable when retrieved on its own.
Minutes 75–90: Can a Reader Move From Answer to Decision?
Visibility that ends on an informational island is unfinished. Start at each page you identified in Search Console and follow the next useful click. There should be an editorial route to a deeper guide, a relevant comparison, a tool, or a commercial page, whichever genuinely serves the reader’s next question.
Google’s link guidance is refreshingly direct: links help it discover pages, relevance signals help it understand them, and descriptive anchor text helps people and Google know what the destination contains. Audit the paths as a user journey, not as a quota of “three internal links per post.”
- Flag priority pages with no contextual incoming links.
- Replace vague anchors when the surrounding sentence does not explain the destination.
- Remove links inserted only for keywords when they interrupt the argument.
- Verify that the page has a clear business action and that the action is measurable in analytics.
Turn Findings Into a Queue the Team Can Finish
Order findings by dependency first, then by likely consequence, confidence, and effort. Access and canonical conflicts precede content polish; a missing commercial page precedes another top-of-funnel article; measurement precedes a claim of success.
| Queue | Definition | Examples |
|---|---|---|
| P0 · Blocks eligibility | Prevents retrieval, index selection, or measurement | Template noindex, crawler 403, broken canonical, missing analytics |
| P1 · Suppresses a priority journey | Affects a page tied to qualified demand or conversion | Wrong intent, missing service page, orphaned comparison |
| P2 · Improves an established result | Strengthens a page already earning impressions or clicks | Refresh evidence, improve title promise, add a useful link |
| P3 · Maintenance | Reduces future drift without blocking current outcomes | Schema cleanup, template consistency, documentation |
Limit the active queue. A small team with twenty “priority” fixes has no priority. Ship the first dependency, annotate the release date, and define the earliest sensible check window before opening the next batch.
Questions About Search Visibility Audits
- Can a useful SEO audit really take only 90 minutes?
Yes, if the goal is triage rather than exhaustive certification. Ninety minutes is enough to identify the few constraints most likely to suppress discovery, clicks, or conversion and to produce a ranked fix queue. It is not enough for log-file analysis, a large international site, a migration investigation, or a penalty recovery.
- Which tools do I need for this audit?
Google Search Console, a browser, PageSpeed Insights, your analytics product, and Bing Webmaster Tools are enough for the core pass. A crawler helps on larger sites but is not required for a small site. ChatGPT search access can be tested from robots.txt and server or CDN logs rather than from a paid visibility score.
- Should the audit start with keywords or technical SEO?
Start with access and indexing because demand work cannot help a page that engines cannot retrieve or select. Stop technical investigation once the foundation is sound, then spend most of the audit on demand, page usefulness, internal paths, and conversion.
- How often should a small site repeat this audit?
Quarterly is a good full-audit cadence for a stable small site, with a shorter weekly Search Console review in between. Repeat it immediately after a migration, redesign, large publishing batch, or unexplained traffic change.
