To run an SEO audit, map crawl-to-content, fix index blockers, check Core Web Vitals, and log fixes in a repeatable checklist.
Search health dips when pages are hard to discover, slow to load, or thin on answers. This guide gives you a crisp, practical workflow to review a site from crawl to content, capture issues, and ship fixes that move traffic and revenue. You’ll finish with a punch list you can repeat each quarter.
Performing An SEO Audit: Step-By-Step
Work top-down. Start with access and discovery, then index state, then content quality, then experience. Close with structured data and tracking. The order matters because each layer leans on the one before it.
What You’ll Need
Access to Search Console, a crawler, PageSpeed tools, and a shared tracker. A staging login helps for template checks. If you don’t have a paid crawler, small sites can be scanned with free tools and careful manual passes.
Audit Checklist Table
| Area | What To Check | Tools |
|---|---|---|
| Discovery | Robots rules, sitemaps, server status, redirects | Search Console, server logs, curl |
| Crawl | Broken links, orphan pages, depth, duplicate paths | Site crawler, internal link map |
| Index | Canonical tags, noindex, blocked resources | URL Inspection, site: search |
| Content | Unique value, headings, intent match | Editor review, user testing |
| Experience | LCP, CLS, INP, images, script weight | PageSpeed Insights, Lighthouse |
| Structured Data | Eligibility, validity, coverage | Rich Results Test |
| Tracking | Analytics tags, events, goals | Tag assistant, analytics |
| Security | HTTPS, mixed content, HSTS | Browser devtools |
| International | hreflang, region targeting | Page source, Search Console |
| Maintenance | Error budgets, regressions, alerts | Monitoring, uptime checks |
Start With Discovery And Access
Begin at the root. Visit /robots.txt and read the rules. The file should live at the root, return 200, and stay within size limits. Rules should block only private areas and staging paths, not public content. Check a few key folders and assets to be sure nothing needed is blocked.
Scan sitemaps. A clean sitemap index lists the current sitemap files. Entries should point at canonical URLs that return 200 and match the preferred protocol and host. Fetch a handful to confirm freshness and coverage.
Open Search Console. Verify the property if needed. In the Pages/Indexing report, group errors by theme and by template. In the Sitemaps report, submit the current index and note last fetch time, size, and any parse errors.
Run A Quick Crawl
Set your crawler to respect robots rules and throttle if the server is fragile. Crawl a sample first, then the full site. Flag 4xx and 5xx, non-self 200s, loops, long chains, and non-200 canonicals. Export the list to your tracker with page type, template, and owner so fixes land with the right team.
Check Index State And Canonicals
Pick a batch of key URLs across templates. Use the URL Inspection tool to see whether a page is indexed, which URL Google picked as the canonical, and whether resources load. If the selected canonical isn’t your preferred URL, compare signals on both pages: the rel=canonical tag, internal links, hreflang, and sitemap entries. Align them so the same URL wins across signals.
Watch for soft 404s, parameter traps, session IDs, near-duplicates, and thin variants. Where multiple pages chase the same search task, fold them into one strong page and redirect the rest. Keep print views, test pages, and tracking parameters out of indexable paths.
Evaluate Content Quality And Intent
Open your top templates and a few high-traffic pages. Check whether the main task is answered near the top. Headings should preview the section that follows. Cut boilerplate that repeats across many pages. Add facts, steps, screenshots, or data that give readers real gain.
Scan for duplicated topics across your own site. Merge pages that chase the same query class into a single hub with clear subheads. When two pages must stay, set one as primary and link the other as a helper. Align each page with a distinct task to avoid cannibalisation.
Measure Experience With Core Web Vitals
Open PageSpeed Insights for a few key URLs. Look at field data first; that comes from real users. If there’s no field data, use the lab test as a guide and confirm on real devices. Track Largest Contentful Paint, Cumulative Layout Shift, and Interaction to Next Paint. These metrics map to loading speed, visual stability, and responsiveness.
For LCP, shrink and preload the hero asset, in-view images, and critical CSS. For CLS, reserve space for images and ads, and avoid layout shifts from late-loading modules. For INP, trim long tasks, defer non-critical scripts, and keep third-party code lean. Re-test after each change so you can see which tweak moved the needle.
You can read Google’s guidance on Core Web Vitals to see thresholds, tooling, and how field data feeds reports.
Speed Checks Beyond Vitals
Run a Lighthouse test and study the waterfall. Heavy bundles, render-blocking CSS, and unused JavaScript are frequent culprits. Ship modern image formats, lazy-load below-the-fold assets, inline tiny critical CSS, and cache static files with long expiry. Test on a mid-range Android device or a throttled desktop to mirror real-world load.
Lock Down Robots Rules And Sitemaps
Keep robots rules tight and readable. Block admin and test areas, not public pages. Avoid redirects on the rules file. Use UTF-8. Keep comments clear. Sitemaps should list canonical, live URLs only. Split large sites into logical sitemap files and group them under a single index. Submit the index in Search Console so you can see fetch logs and errors. This keeps discovery fast and clean.
For precise syntax and limits, see Google’s page on building and submitting sitemaps.
Add Structured Data Where It Helps
Mark up content that maps to supported features: articles, products, breadcrumbs, FAQs, events, and more. Validate with the Rich Results Test and fix any errors or warnings. Keep markup accurate, reflect what users can see on the page, and avoid fake reviews or hidden content. Start with site-wide wins: organization markup in the header, breadcrumb markup in templates, product markup on product pages.
Use Search Console To Prioritise Work
Open the performance report. Sort by click drop, then compare the last 28 days to the previous period. Drill into query and page pairs. Check whether the page still fits the query. If yes, refresh the content, tighten headings, and add internal links from related hubs. If not, adjust the page intent or split content so each page targets one clear task.
In the Pages/Indexing report, group errors by template to fix at scale. Use the URL Inspection live test to confirm a fix, then request re-indexing. Track time to first crawl and time to index to judge impact and to spot bottlenecks in deployment.
Internal Links And Navigation
Build a simple model of your site sections and key pages. Every important page needs text links from related hubs. Use short, descriptive anchors. Add links in the body where readers need them, not only in nav bars. Remove dead-end pages that hoard links with little value.
Check paginated lists, tag pages, and filters. Avoid infinite spaces that trap crawlers. Where lists are long, add view-all pages with server-side render if possible. Keep facets crawlable only if they serve a search use case and lead to unique, useful results.
Media, Images, And Alt Text
Compress pictures and serve responsive sizes with srcset. Set width and height to reserve space. Add alt text that names the subject. For video, provide transcripts and key moments markup where it fits. Host heavy media on a fast CDN and reuse assets across pages to cut duplicate downloads.
Accessibility And Mobile Checks
Open pages on a phone and a mid-range Android. Check tap targets, contrast, and font sizes. Test without JavaScript to spot content that never renders. Fix sticky elements that hide headings or block text. Keep pop-ups and nags from covering the main content.
Second Table: Fix Paths And Owners
| Issue Type | How To Verify | Who Owns Fix |
|---|---|---|
| Crawl Blocks | robots.txt test, log sample | DevOps |
| Index Gaps | URL Inspection, Pages report | SEO + Dev |
| LCP Slow | Field data, lab trace | Frontend |
| CLS Jumps | Layout shift debug | Frontend |
| INP Lag | Long task audit | Frontend |
| Duplicate Pages | Canonical map | SEO + CMS |
| Thin Content | Template review | Editorial |
| Schema Errors | Rich Results Test | SEO + Dev |
| Broken Links | Crawl export | SEO + Eng |
| Tracking Gaps | Tag assistant | Analytics |
Set Up A Repeatable Cadence
Make a simple rhythm: monthly spot checks for new errors, a quarterly deep pass, and an annual reset of crawl rules and templates. Keep a shared tracker with owners and due dates. Add a column for expected impact so teams pick the best next task. Roll wins into playbooks so fixes repeat across templates and sections.
Mini Walkthrough: One Page, End To End
Step 1: Fetch
Load the page in a private window. Check that the main content appears fast and above the fold. Disable JS and reload. If the page hides content without scripts, render it server-side or use a hybrid approach so the first view carries the core text.
Step 2: Inspect
Run the URL through the inspection tool. Confirm the indexed URL, mobile-friendliness, and whether Google can fetch all resources. If assets are blocked, allow those files or inline the critical bits so layout stays stable.
Step 3: Speed
Open PageSpeed Insights. Note real-world data at the top. Compare with the lab test below. Fix the biggest pain first: giant hero images, render-blocking CSS, or long tasks from third-party scripts. Ship a small change, re-test, and record the delta.
Step 4: Content
Read the page like a new visitor. Does the intro answer the task fast? Are headings clear and in order? Trim fluff and add specifics. Link to a related hub where a reader would go next. If a query shifts, adjust the page so it serves that new intent.
Step 5: Re-check
Re-run the live test, request indexing, and log the change. Watch clicks and Core Web Vitals over the next few weeks. If rankings dip, check which queries moved and whether the page still matches the task those queries imply.
What Good Looks Like
A tidy site has robots rules that only shield private areas, clean sitemaps, stable canonicals, and pages that answer the task near the top. It loads fast on common phones, shifts little, and responds to input in a snap. It uses markup where it helps readers and keeps tracking clean and lawful.
Keep Learning From Source Docs
Two must-reads round out this playbook: the SEO starter guide for fundamentals across crawling, indexing, and content, and the detailed guidance on Core Web Vitals for user experience metrics and tooling.