How to do an SEO audit: what I actually check
What a real audit actually checks
Every URL in the sitemap gets crawled for its status code, then checked for a unique, appropriately-sized title tag and meta description, a correct canonical tag, valid robots directives, and no accidental duplication with another page. Structured data gets parsed and validated, not just visually spot-checked. Core Web Vitals get measured with a real tool (I use Lighthouse), not estimated.
Why most 'free SEO audit' tools are misleading
Many automated tools flag hundreds of low-value 'issues,' such as minor image compression suggestions or generic 'improve your score' prompts, without any sense of which ones actually move the needle. A wall of 200 warnings is not an audit. It's noise that makes it hard to see the five fixes that would genuinely help.
What I do with the findings
Issues get fixed in priority order: broken canonicals and missing metadata first, then duplicate or over-length titles, then structural issues like heading order and internal linking, then performance. After fixes, I re-crawl and re-run Lighthouse to confirm the fix actually worked, rather than assuming it did.
Step 1: crawlability and indexing
Start by confirming search engines can reach and store the pages that matter. Every URL in the sitemap should return a 200 status, carry a canonical tag pointing to itself, and have no stray noindex directive in the HTML or the X-Robots-Tag header. Check robots.txt is not blocking important sections. Then compare the sitemap with Search Console's page indexing report: pages marked crawled but not indexed or discovered but not indexed usually point to thin content, duplication or weak internal linking.
Step 2: duplicate URLs and redirects
The same page reachable at several addresses splits ranking signals. Check that www and non-www versions, http and https, and trailing-slash variants all redirect with a single 301 to one canonical form. Look for redirect chains, where one redirect leads to another, and for internal links pointing at redirected URLs instead of the final address. After a redesign or migration, map every old URL to its new equivalent so existing rankings and links carry over.
Step 3: on-page signals
Review each page's title tag and meta description for uniqueness and length, roughly up to 60 characters for titles and 110 to 160 for descriptions so they display fully. Each page should have exactly one H1 and a logical heading order. Check that images have descriptive alt text, that the main topic appears naturally in the title, H1 and opening paragraph, and that pages link to closely related pages with descriptive anchor text.
Step 4: structured data, speed and content quality
Parse every JSON-LD block and validate it, because broken or misleading schema can be ignored or trigger manual actions; marking up your own reviews on your own site is a common mistake. Measure Core Web Vitals with both lab data (Lighthouse) and field data (the Core Web Vitals report or PageSpeed Insights), since real users on slow phones often differ from a test on a fast laptop. Finally, flag thin pages with little unique content and topics covered by several overlapping pages.
How long does an SEO audit take?
For a small site of up to a few dozen pages, a thorough technical and on-page audit can be done in a day or two, including a written list of fixes in priority order. Large sites with thousands of URLs, multiple languages or a recent migration take longer, mainly because finding patterns across many templates needs more crawling and analysis. The audit itself is the quick part; fixing the issues and confirming in Search Console that Google has picked up the changes is what takes weeks.
What to do after an SEO audit
Fix the items that stop pages being indexed first, then the ones that confuse search engines about which URL to rank, then on-page improvements, and finally speed and content depth. After fixes go live, resubmit the sitemap, request indexing for the most important changed pages in Search Console, and watch the page indexing and performance reports over the following weeks. Re-crawl the site to confirm every fix landed. A short follow-up audit after a month catches anything that regressed.
| Priority | Check | Why it matters |
|---|---|---|
| 1 | Status codes, canonicals, noindex | Pages that cannot be indexed cannot rank |
| 2 | Redirects and duplicate URLs | Split signals weaken every version |
| 3 | Titles, descriptions, headings | Decide relevance and click-through |
| 4 | Structured data validity | Enables rich results and machine understanding |
| 5 | Core Web Vitals (lab and field) | Page experience and conversions |
| 6 | Thin or overlapping content | Common cause of pages not being indexed |
Frequently asked questions
How often should a site be audited?
After any major change (redesign, migration, new section), and otherwise roughly every 6-12 months to catch drift like broken links or outdated metadata.
Does a high 'SEO score' from a tool mean the site is fine?
Not necessarily. Scores from generic tools often weight low-impact issues heavily. A prioritized, human-reviewed list is more useful than a single number.
Can an audit find why a specific page isn't ranking?
It can rule out technical causes (noindex, canonical pointing elsewhere, thin content) but ranking also depends on competition and backlinks, which an on-site audit alone can't fully explain.
What tools do I need for a basic SEO audit?
Google Search Console, PageSpeed Insights or Lighthouse, Google's Rich Results Test, and a crawler that lists status codes, titles and canonicals for every URL.
Relevant service
Related reading
Want this done properly?
Tell me what you are building. I reply within a business day.
Start a project