What Is SEO Audit and How Does It Find Problems

An SEO audit is a structured check of whether search engines can find, crawl, index and understand a site's pages, followed by a prioritized list of fixes. The most useful result isn't a perfect score or a spreadsheet full of warnings; it's knowing which problems deserve your attention first.

The popular advice is to run an audit tool and fix everything it flags. That sounds tidy, but it treats a missing description on an old announcement like a blocked service page that should generate inquiries. Those aren't equivalent problems, no matter how enthusiastically the software colors them red.

A useful audit connects technical evidence to business decisions. It tells you what's wrong, why it matters, who should handle it and how you'll know whether the fix worked.

What an SEO Audit Checks Before It Checks Rankings

Rankings are an outcome. Before evaluating them, an audit needs to establish whether search engines can access your important pages and make sense of what they contain.

Google's explanation of how Search works describes three foundational stages: crawling, indexing and serving results. Each stage asks a different question:

• Crawling: Can automated programs discover the page and download its content?
• Indexing: Can Google analyze its text, images, videos, title elements and alternative text, then store eligible information?
• Serving: Does Google select the indexed page as a relevant, useful result for someone's search?

A page can succeed at one stage and fail at another. Being reachable doesn't mean it's indexed, and being indexed doesn't mean it will appear for the searches you care about.

Separate access from visibility

Consider a hypothetical service page linked only from a downloadable PDF. A visitor with that document might reach it easily, but the website provides no ordinary internal link to help crawlers discover it. Adding a relevant link from the services navigation addresses discovery, not necessarily the page's usefulness or ranking potential.

That's why an audit looks beyond keywords and backlinks. It examines navigation, redirects, sitemaps, access restrictions and whether important content remains understandable when search engines process the page.

The work combines website analysis with business context.

For French-speaking small-business owners, Wispra's guide to réaliser un audit SEO pour PME offers a companion resource for approaching the task.

An audit should identify where a page's path to visibility breaks, not simply report that traffic is disappointing.

You should leave with a diagnosis and a sensible order of work. Removing blockers improves the opportunity to compete; it doesn't come with a reservation at the top of Google.

Five SEO Audit Areas and What Each One Checks

Technical and on-page checks solve different problems

Technical work examines the delivery and accessibility of your pages. On-page work examines what those pages communicate. A technically accessible page can still leave visitors uncertain about what you offer, whom you serve or what to do next.

Review each important page for a descriptive, distinct title element and a meta description that accurately summarizes its content. Use headings to organize the explanation, not just to make selected sentences large. Informative images need useful alternative text; decorative images don't need invented keyword descriptions.

Search intent means the task behind someone's query. A person searching for emergency repair needs availability and contact details, while someone researching repair options may need explanations and comparisons. An audit should judge whether the page helps that person finish their task, not merely whether it contains the search phrase.

Structured data is machine-readable markup describing page content. Google's General Structured Data Guidelines require the underlying page to remain accessible to Googlebot, without robots.txt blocking, noindex or another access restriction, and require the markup to represent visible content. Passing a syntax test isn't enough.

The areas fit together like this:

Reputation, location and usability need context

Off-page checks review links from other websites. Look at the relevance of referring domains, the wording used in links and signs of purchased or manufactured link patterns. A raw backlink total says little about whether those links represent genuine recommendations.

Don't turn an automated toxic-link label into an emergency cleanup project. Investigate clear manipulation rather than treating every unfamiliar referring domain as a threat. Some reports are much more confident than their evidence deserves.

Local checks matter when customers choose you partly because of where you operate. Review your Google Business Profile's name, address, phone number and categories against the website and major directories. Handle reviews openly and consistently, without manufacturing praise or pressuring people to revise criticism.

UX means user experience. Test whether visitors can read the page, find their way around, tap controls and complete forms on a phone. A beautiful contact form isn't useful if a consent banner covers its submit button.

Keep these boundaries in mind:

• Shared work: Performance improvements often address both technical and UX findings.
• Unequal evidence: A verified broken form deserves more confidence than a third-party estimate of link quality.
• Conditional relevance: Local visibility deserves attention when geography affects customer choice, not just because an audit template includes it.

Technical Checks That Decide Whether Pages Count

Crawling is the download step; indexing is the decision about storing and interpreting what was downloaded. Treating them as interchangeable leads to fixes that solve the wrong problem.

Google Search Central's explanation of Search separates those processes and makes an important limitation clear: discovery doesn't guarantee indexing. Your technical inventory should record what really happens at each important URL.

Build a URL-level inventory

A crawler follows links and collects page information. Configure it to inspect the live site, then review these findings:

• HTTP status codes: Investigate important URLs returning 4xx errors, which indicate client-side problems such as missing pages, or 5xx errors, which indicate server problems.
• Redirect paths: Replace unnecessary chains with a direct redirect to the intended destination.
• Canonical declarations: Check that each declaration identifies the intended preferred URL rather than an unrelated page.
• XML sitemaps: Include the pages you want indexed, using their final, preferred addresses.
• Robots controls: Review robots.txt rules and page-level indexing directives.
• Internal links: Confirm that navigation and contextual links reach important pages.
• Rendered content: Check whether JavaScript-dependent text and links appear when the page is rendered.
• Template leftovers: Look for accidental noindex directives carried over from staging or an old template.

A canonical declaration is a preference signal about duplicate or similar URLs. It isn't a substitute for correcting broken links or deciding which pages should exist.

Match the control to the objective

Robots.txt restricts crawling. It shouldn't be used as the mechanism for keeping a page out of the index, because a blocked address can still be known through links elsewhere.

Use noindex when a publicly accessible page should stay out of search, and let crawlers access the directive. Use authentication when the content should be private. These controls solve different problems.

Practical rule: Before changing an exclusion, write down whether the page should be public, crawlable and searchable. Those answers determine the correct fix.

Google's Search Essentials also cautions that meeting its requirements doesn't guarantee crawling, indexing or serving. Google may select another canonical URL or decide that a page doesn't add enough useful content.

The remediation list should specify affected URLs, the intended behavior and the person responsible. Restore valuable broken pages, simplify redirects, expose essential content and reconcile sitemap entries with canonical destinations.

Verify the changes in Search Console and with another crawl. Track indexed pages, exclusion reasons, crawl anomalies and coverage of important organic landing pages, rather than celebrating a smaller error total. For more background on the underlying work, see what technical SEO involves.

Core Web Vitals Thresholds and Page Experience

These boundaries come from Google's Core Web Vitals report documentation. They turn a vague request to make the website faster into measurable targets.

LCP describes when the largest visible content element appears, often a prominent image or text block. INP measures how promptly the page responds to interaction, while CLS measures unwanted movement. A page can load its main image quickly and still respond poorly when someone taps a menu.

Read real visits alongside controlled tests

Search Console's report uses field data, collected from real visitors. Lab tools simulate visits under controlled conditions, which makes them useful for investigating a specific problem.

Don't assume a strong desktop lab result describes a customer using an older phone over a weak connection. Segment results by mobile versus desktop, page template, geography and connection type where your tools provide that detail. Separate a site-wide issue from one confined to a particular template.

The diagnosis should connect the metric to a plausible cause:

• Slow main content: Oversized hero images and render-blocking resources can delay LCP. Resize and compress images, and remove unnecessary blocking work.
• Delayed interaction: Heavy JavaScript and third-party tags can worsen INP. Defer noncritical scripts and reduce main-thread processing.
• Unexpected movement: Missing image dimensions, injected banners and late-loading fonts can increase CLS. Reserve space and make font loading predictable.

Google's Core Web Vitals guidance supports using performance measurements to connect loading, responsiveness and stability problems to engineering work.

Verify the fix beyond the test run

Set performance budgets in deployment checks so a new image, script or tracking tag doesn't quietly undo the improvement. Test the affected template, not just the homepage.

Field reporting uses a rolling window, so better lab results won't immediately replace older visitor experiences. Review the trend over the following weeks rather than declaring victory after a fresh test.

For a plain-language explanation of the metrics and how to interpret them, use this guide to Core Web Vitals. Performance work should produce a more usable page, not merely a nicer screenshot of a score.

What to Gather Before the First Crawl

Start with access and a baseline, not a shopping trip for software. Have these items ready:

• Verified Search Console access: Confirm you can inspect index status, crawl statistics, sitemaps and the Core Web Vitals report.
• A crawler: Choose Screaming Frog SEO Spider or another crawler appropriate for the site's size and rendering needs.
• Business-critical URLs: Export pages tied to leads, sales, donations or calls, and identify the main organic landing pages.
• A dated baseline: Record indexed pages, exclusion reasons, template-level performance status and analytics events such as completed forms or qualified calls.
• An implementation owner: Decide who can edit content, change templates and approve technical work.

If your team uses MCP-connected tooling, NotFair's documentation explains how to access Search Console through MCP. Check permissions and data access before connecting any integration.

Establish the page inventory

Compare the crawler's output with the pages your organization expects to exist. This guide to finding all pages on a website can help assemble that inventory.

Resolve missing Search Console access before attempting the final checklist. A crawler alone can't confirm Google's indexing decisions.

Tools That Cover Each Audit Area

No single tool covers every audit area. Choose the tool that answers your question, rather than collecting dashboards and hoping clarity appears.

Start with a small working set

Search Console, a crawler and PageSpeed Insights make a sensible starter combination. Together, they provide search-engine evidence, site-wide inspection and focused performance testing.

Add a paid suite when you need backlink analysis or competitor comparison. Don't buy one just because an audit feels more official with a subscription attached.

For automation options, The SEO Agent is another resource to evaluate against your reporting needs. Before adopting automated recommendations, check whether you can inspect their evidence, control site changes and verify the output.

A tool can identify a missing element. It can't decide whether fixing that element is your most valuable use of a developer's afternoon.

Use this overview of SEO optimization tools to compare additional options when your questions become more specialized. The most important limitation remains the same: software reports findings, while a person must decide what matters.

Ranking Findings by Impact, Effort and Risk

Sort findings into fix now, test next or monitor. Assess impact on business-critical pages, effort in people and implementation work, and risk.

Record risk in both directions: the cost of leaving the problem unresolved and the chance that changing it causes damage. A broad navigation rewrite might have substantial potential value but still deserve testing before release.

This framework makes the sorting process visible:

Turn findings into assigned work

1. Export the findings with affected URLs and supporting evidence.
2. Group related findings by audit area and shared cause.
3. Assess impact, effort and risk.
4. Assign an owner and a deadline.
5. Record the baseline for the affected metric.
6. Implement the change.
7. Verify the result in Search Console or the relevant testing tool.

In a hypothetical 40-page service website, 14 pages have missing meta descriptions, while the contact page carries an accidental noindex directive from an old template. If that contact page should appear in search, correcting the directive is a high-impact, low-effort candidate; the missing descriptions can wait.

Keep that judgment conditional. First verify that the exclusion is accidental and that the contact page has a meaningful search role. A scary label isn't evidence of business importance.

Keep evidence separate from recommendations

Evidence might say that a page contains noindex. The recommendation explains whether to remove it, why and who should approve the change.

Google's guidance on search fundamentals for AI features reinforces familiar priorities such as crawlability, internal linking, textual content and page experience. It doesn't prescribe a universal audit score or business priority order.

A longer export isn't a better audit. A short list of important bottlenecks can be more useful than a long queue of cosmetic warnings.

Track form submissions, calls, donations or qualified visits alongside visibility. Rankings indicate search exposure; they don't establish whether the website helps the organization achieve its purpose.

Your First Audit Pass and When to Repeat It

Use Search Console and your crawler to work through this checklist:

1. Important pages return successful status codes.
2. Key pages are indexed, and exclusions have understood reasons.
3. Robots.txt and noindex rules match each page type's intended purpose.
4. The sitemap contains only pages intended for indexing.
5. Key pages have distinct, accurate titles and meta descriptions.
6. Main templates meet Core Web Vitals targets in available field data.
7. Structured data matches visible content and remains accessible.
8. Internal links reach every service page.
9. Local profile details match the website when location matters.
10. Baseline metrics are recorded before changes begin.

Use the visual checklist as a companion, not a substitute for investigating each finding.

Repeat after meaningful change

Run a full audit after redesigns, migrations and major content changes, when URLs, templates and metadata are especially likely to shift. The SEO Handbook timeline of Google algorithm updates also illustrates why reviewing only keywords misses changes involving content quality, meaning and usability.

Between major projects, schedule lighter checks according to site size and publishing frequency, then compare findings with your baseline. Treat the cadence shown in the graphic as a planning option, not a requirement for every website.

If a video explanation suits your learning style, this provides another format:

Run the first pass yourself, or explore the scope of SEO services if you'd rather hand off the detailed work.

Bruce and Eddy offers SEO audits and website development support to connect search findings with practical fixes. Visit Bruce and Eddy to discuss which pages matter most to your organization and what an appropriately scoped audit should investigate.