# How to Do SEO Audit: A Step-by-Step Guide in the Right Order > How to do SEO audit in the right order: measurement, crawling and indexing, content, authority, then a rule for which findings to fix first and which to skip. - URL: https://missiongrowth.io/blog/how-to-do-seo-audit - Published: 2026-09-14 · Updated: 2026-09-19 - Author: Ömer Furkan Aktaş, Founder, Mission Growth - Publisher: Mission Growth. Company facts: https://missiongrowth.io/llms.txt An SEO audit checks whether Google and other search engines can find, crawl, index and trust every page on your site that should rank. If you're searching how to do SEO audit for the first time, you'll usually land on a flat checklist in whatever order occurred to whoever wrote it. That order is the problem. A content finding on a page Google hasn't indexed is wasted work. A backlink review run before you've fixed duplicate content just profiles noise. This guide runs the audit in dependency order instead: verify your measurement, then crawlability and indexing, then content quality, then authority. Every finding also passes through a decision rule before you touch it, because a flagged issue and a problem worth fixing aren't the same thing. In this guide: - Why the order you run checks in matters more than the checklist itself - The gate sequence, and why each gate depends on the one before it - How to tell real keyword cannibalization from a false alarm - A test that turns a pile of findings into prune, consolidate, improve or keep ## How to do SEO audit: what it actually checks An SEO audit is a structured check of whether Google and other engines can find, crawl, index and trust every page on your site that should rank. You run one before a redesign, after a traffic drop, after a domain migration, or on a recurring cadence. Each can quietly break something your normal editorial workflow never touches. Search interest here splits into two questions: what is SEO audit, and how do you run one well. Most SEO audit checklists answer the first and stop there, splitting the work into types instead: technical, on-page, content, off-page. That split is accurate, but it doesn't tell you the order to run checks in. This guide does. The real difficulty is knowing which flagged issues are worth fixing, and in what order. How often should you audit your website? Roughly every quarter, or whenever one of these happens: - You're about to redesign or replatform and want a clean baseline first - Organic traffic or rankings dropped and you need to rule out a cause you control before blaming an algorithm update - You just migrated domains, URL structures or a CMS ## Verify your measurement before you audit anything Verify your Search Console property, Analytics setup and tracking configuration before you check anything else. Every later finding in this guide reads from that data, so a wrong number here becomes a wrong decision three gates later. Few guides treat measurement as a step at all. They jump straight into crawling and content checks against whatever data already sits in the account. That's backwards. A misconfigured Search Console property, a duplicate Analytics tag firing twice, or a staging subdomain still sitting in the index produces a findings list that looks precise and is quietly wrong at every gate that follows. Before you touch a crawler or a rankings report, confirm: - **Property match.** Your Search Console property covers the exact domain and protocol you actually serve traffic on, with no orphaned old property catching a slice of the data. - **One tag firing once.** Analytics and conversion tags fire once per page load, not twice from a leftover container in your tag manager plus a hardcoded snippet. - **No staging leakage.** A staging or preview subdomain isn't indexed and isn't reporting into your production Analytics view. - **Tracking fires where you think it does.** Pixels and conversion events fire on the exact pages and actions you expect, verified individually rather than assumed. A free tracker audit tool checks pixel and tracking setup for exactly this gate: run it to [verify your tracking setup](https://missiongrowth.io/tools/tracker-audit) before you trust anything downstream of it. ## The audit order that actually works Run an SEO audit in gate order: measurement, then crawlability and indexing, then content quality, then authority. Each gate's findings are only trustworthy once the gate before it is clean. This SEO audit process only holds up when the gates run in that order, and reshuffling them makes the later findings meaningless. Here's why the dependency runs in that direction. A content-quality finding on a page Google hasn't indexed yet is wasted analysis: you're polishing something nobody's reading. A backlink review run before duplicate content and canonical issues are fixed can't tell you which URL in a cluster actually deserves the links pointing at it. None of it matters if the data feeding your dashboards was wrong to begin with. That's why measurement comes first. The SEO audit steps below follow that dependency instead of a brainstorm order. ::figure{src="/blog/figures/how-to-do-seo-audit-1.svg" alt="Flow diagram of the SEO audit process: measurement, crawlability and indexing, content quality, authority, and an optional AI visibility add-on, in sequence" caption="The SEO audit process gate order: each gate must be clean before the next one’s findings can be trusted" width="720" height="337"} Once the first four gates are clean, AI search visibility, llms.txt, brand mention tracking, citation monitoring, is worth adding as a fifth gate. It sits on top of a clean SEO foundation. It doesn't replace one. For the crawlability mechanics, see [the technical SEO checklist](https://missiongrowth.io/blog/technical-seo-checklist-2026) for the deeper, exhaustive list. ## Crawlability and indexing: what actually controls Googlebot's attention Technical SEO audit checks, crawlability and indexing, matter because how often Google revisits a page isn't a flat schedule applied across the whole site. The more useful fix is changing the inputs a scoring system reads. Tweaking a directive Google ignores accomplishes nothing. ### The crawl-priority mechanism, and the robots.txt correction Patent US10621241B2 describes a crawl priority score as a function of a page's PageRank squared, how often its content substantively changes, and the time since it was last crawled. The patent sorts URLs into three crawl tiers: - Base, or periodic: the default cadence - Daily - Real-time: for pages that change constantly Google hasn't confirmed it runs this exact formula in production. The patent still frames crawl frequency as a scored allocation problem rather than a schedule you set yourself, so you influence how often Googlebot returns by publishing content that substantively changes, not by asking nicely. That reframing also corrects a common assumption about crawl rate: - **Not supported:** the `crawl-delay` field in robots.txt. There's no dial in a Search Console dashboard either. - **Supported:** a temporary 500, 503 or 429 response for a short stretch, a couple of hours up to a day or two, when you need to cut Googlebot's crawl rate urgently. Using the error response longer than that can hurt how your site appears, and the crawl rate recovers on its own once the errors stop. ### What to check, in order Run these checks in this order: - **Index coverage.** In Search Console's Page Indexing report, check what's excluded and why: "Crawled, not indexed" and "Discovered, not indexed" both point at a quality or crawl budget problem rather than a technical block. - **Orphaned and unlinked pages.** A page with no internal links pointing to it is harder for Googlebot to find and re-find, no matter how good the content is. - **Canonical accuracy.** Canonicals that point at themselves on pages meant to rank, and correctly pointed canonicals on the duplicates that shouldn't rank. - **Rendering.** Confirm the content you want indexed shows up in your rendered HTML, the version Googlebot actually sees, and not only in your source HTML. We migrated our own React single-page app to prerendered static HTML for 20 marketing pages because AI crawlers do not execute JavaScript. The same rendering gap affects how Googlebot and other engines see a page rendered in the browser. - **Structured data.** Validate markup against Google's Rich Results Test. Broken structured data doesn't tank rankings on its own, but it forfeits the eligibility for a rich result the markup was meant to earn. Core Web Vitals belong in this gate too, since they're part of how a page's fetch and render performance is measured: - **Largest Contentful Paint** measures how fast the main content finishes loading. - **Interaction to Next Paint** measures how fast the page responds to a click or a tap. - **Cumulative Layout Shift** measures how much the layout jumps around while the page loads. A page that fails these doesn't drop out of the index. It competes from a weaker position in results where the content on both sides is otherwise close. ## Content quality: duplicates, thin pages and false-alarm cannibalization Content quality checks matter because Google's Helpful Content System runs inside every core update rather than as a separate periodic sweep. In March 2024, it "evolved and became part of our core ranking systems." Unresolved duplicate, thin or falsely flagged cannibalized pages keep suppressing the rest of the site's helpfulness signal, and they do it again with every core update that runs afterward. Duplicate or highly similar content has a straightforward fix, whatever form it takes: - The same product listed under different URL parameters - Syndicated content copied elsewhere - Thin location pages built from boilerplate text Canonicalize, consolidate or noindex the losing copy in each case. Cannibalization is where flagged lists get it wrong. Not every URL pair that shares a query in Search Console is cannibalization. It's real cannibalization only when both pages compete for the same primary intent. When one page treats the query as supporting content on its way to a different topic, Google's own page selection resolves the overlap without you doing anything. Test it before you consolidate anything: - Do both URLs actually target the same primary keyword, or does one just mention it in passing? - Does merging them lose a distinct angle either page currently serves, a comparison versus a definition, a beginner take versus an advanced one? - Has Search Console shown both URLs ranking for the query across multiple weeks, or was it a single crawl artifact? If only one page was ever meant to rank for that query, leave both alone and go fix something that's actually broken. ## Backlinks and authority signals Authority checks go beyond backlink count because a homepage's trust classification appears to extend to the rest of the site. An audit that only counts links misses half the signal, and links are only one of the systems covered in [how to rank higher on Google](https://missiongrowth.io/blog/how-to-rank-higher-on-google). Google's internal Content Warehouse API documentation leaked in March 2024, and Google confirmed it was authentic on May 29, 2024. Independent analysis of that leak found a `homePageInfo` field with four values, NOT_HOMEPAGE, NOT_TRUSTED, PARTIALLY_TRUSTED and FULLY_TRUSTED, used to classify a homepage's trust level. Exactly how that classification propagates to internal pages is analysts' inference from the leak, not something Google has confirmed line by line. The field and its four-value scale are real. That makes homepage transparency an authority audit task in its own right, alongside backlink work: - **A real About page,** not three sentences of boilerplate - **Verifiable contact information,** a physical address or a named team, not a contact form alone - **A privacy policy** that matches what the site actually collects - **Clear, visible ownership** of the site or business Layer that on top of the backlink work you'd expect: links pointing at 404s or retired URLs (redirect them), losses in referring domains since your last audit, and spam link patterns worth a disavow. Homepage trust and backlink health are both authority inputs. A site with strong links and a homepage reading NOT_TRUSTED is fixing the wrong half of the problem. ## Turn findings into a prioritized fix list Prioritize audit findings with one test that crosses a URL's traffic and backlinks against its content quality, sorting each into one of four outcomes: prune, consolidate, improve or keep. ::figure{src="/blog/figures/how-to-do-seo-audit-2.svg" alt="Quadrant chart for SEO audit findings with traffic on one axis and backlinks on the other, each quadrant labeled prune, consolidate, improve or keep" caption="Cross a page's traffic against its backlinks to sort every SEO audit finding into prune, consolidate, improve or keep." width="720" height="440"} A page with no traffic and no backlinks is a prune candidate, unless it serves a structural role: an internal link hub, a required legal page. A page with backlinks but no traffic is usually worth consolidating into whatever page on the same topic does get visits, so those links stop going to waste. A page with traffic but no backlinks and decent content quality is worth improving rather than replacing. It's already earning attention on its own. A page with both traffic and backlinks is the one you leave alone and protect. ### A worked example Say your SEO audit tools turn up three flagged pages: - An old post with no traffic and no backlinks: prune it. - An old comparison page that still holds a handful of referring domains but gets no visits: consolidate it into your current comparison page and redirect. - A product page that gets steady traffic but is thin on content and has no backlinks: improve it. Add real depth rather than replacing it. The decision, not the flagged list, is the actual deliverable of an audit. How you package these calls for stakeholders is the job of an [SEO report](https://missiongrowth.io/blog/seo-reporting), a separate question from making the calls correctly. ### Where to start An SEO audit only produces trustworthy findings when it runs in dependency order and every finding passes through a decision rule before anyone touches it. Skip the order or skip the filter, and you're left with a longer version of the same list everyone else already has. Building authority or ranking higher from a clean baseline is a different exercise than this one; this guide stops at what to fix. Start with the gate you're least sure about. If you can't say your Search Console property is clean, start there. If your data's fine but you've never mapped which pages are indexed against which pages should be, start with crawlability instead. Pick one gate, run it this week, and let what you find there decide whether the next gate is worth your time yet. ## FAQ ### Should an SEO audit include AI search visibility? Yes, but only as a final add-on gate once the SEO gates here are clean. It doesn't replace them. AI visibility work, llms.txt, brand mention tracking, citation checks, depends on the same crawlable, indexed, trustworthy foundation; see [the full AI SEO checklist](https://missiongrowth.io/blog/ai-seo-checklist) for the complete step-by-step list once your SEO gates pass. ### Is keyword cannibalization always something to fix? No. It's real cannibalization only when both competing URLs treat the query as their primary content. When one treats it as supporting content on the way to a different topic, Google's own page selection resolves the overlap without a redirect or consolidation. ### Can I run an SEO audit myself, or do I need outside help? The measurement, crawlability and content gates here are runnable in-house with free tools. Ongoing authority work and continuous monitoring is where teams more often bring in outside help. Mission Growth is an AI-led SEO and GEO growth service: AI catches the signal, our experts make the move, and you see the result, and it's one option worth considering. ### What's the most common mistake teams make when auditing their own site? Treating every tool-flagged issue as equally urgent instead of running the gates in order and testing each finding against a decision rule first. A long flagged list from your SEO audit tools isn't a priority list: most of it is noise once you sequence by dependency and filter by traffic, backlinks and quality. ### How often should I re-run an SEO audit? Roughly quarterly, plus whenever something changes the baseline: after a traffic drop, a domain or CMS migration, or before a redesign. Wait longer, and small drift, broken redirects, orphaned pages, expired canonicals, compounds quietly until the next algorithm update makes it visible all at once. ### Does this guide replace comparing AI search to traditional SEO? No. This guide audits your existing SEO foundation: crawlability, indexing, content and authority. If you're weighing how AI search changes strategy against traditional SEO more broadly, see [AI SEO vs traditional SEO](https://missiongrowth.io/blog/ai-seo-vs-traditional-seo) for that comparison.