SEO Audit Checklist: 10 Steps Ranked by Revenue

An SEO audit checklist is an ordered set of checks that finds what is costing a site money, starting with index coverage and ending with AI-citability. Work it top-down: indexation, canonicals, Core Web Vitals, internal linking, titles, schema, cannibalization, backlinks, crawl budget, then AI citability. Order matters more than length.

Most published checklists are alphabetical inventories. They list 200 items with no sequence, so teams spend a fortnight rewriting alt text while a rogue parameter quietly duplicates 40,000 URLs. The checklist below is sequenced by revenue impact, which means the first four items usually account for most of the recoverable traffic on a site that has never been audited properly.

The 10-point SEO audit checklist, ranked by revenue impact

Run these in order. Do not skip ahead because an item looks easy. Each check assumes the one above it has passed, and fixing item 5 on a site that fails item 1 is wasted effort.

The four checks that move money

1. Index coverage versus your real page count

Export every URL you intend to rank, then compare it against the Pages report in Search Console. The number that matters is not the total indexed count, it is the gap in both directions: pages you want indexed that are not, and pages indexed that you never wanted. Common causes of the second kind include tag archives, internal search results, paginated series, session-parameter URLs, staging subdomains that were never blocked, and soft 404s on discontinued products.

One rule catches most mistakes here. If a URL should leave the index, give it a noindex tag or a 410, not a robots.txt disallow. A blocked URL cannot be recrawled, so Google never sees the signal telling it to drop the page. Blocking is for saving crawl budget on URLs that were never indexed in the first place.

2. Canonical and parameter control

Every parameterised URL should either self-canonicalise on purpose or point at the clean version. Faceted navigation is where most ecommerce audits fail. Test three things by hand: whether a filtered URL such as ?color=blue&size=xl generates a crawlable link, whether its canonical points back to the bare category, and whether paginated pages self-canonicalise (they should) rather than all pointing at page one (they should not).

Then check the four-variant problem: HTTP, HTTPS, www and non-www, plus trailing-slash duplicates. All should collapse to a single version with 301 redirects, not 302s, and not through a chain of three hops. If you inherited a migration, verify cross-domain canonicals too. These are cheap technical SEO fixes with an outsized effect, because they consolidate ranking signals that are currently being split across duplicates.

3. Core Web Vitals on mobile

Judge performance on field data at the 75th percentile of mobile sessions, not on a lab score. A near-perfect lab result from a throttled desktop run tells you nothing if real users on mid-range Android phones are waiting 3.4 seconds for the main image.

These three thresholds are the pass marks as of 2026, and Google has changed the metric set before, so recheck the definitions each year. Treat the numbers as a floor rather than a goal: a page at 2.4s passes, but a page at 1.4s converts better.

4. Internal linking depth

Pull a crawl depth report and find every commercial page sitting four or more clicks from the homepage. Anything that earns revenue should be within three. The fix is rarely a navigation rebuild. It is usually a handful of hub pages, a related-items block on templates that lack one, and contextual links added from the blog posts that already attract traffic. While you are in the report, list orphan pages, meaning URLs in your sitemap that nothing links to internally, and either link them properly or remove them.

The supporting checks

• Titles and meta descriptions. Duplicates almost always come from a template collision, such as every paginated page reusing page one's title. Fix the template, not the individual pages. Expect Google to rewrite some titles regardless; write for the click anyway.
• Schema. Cover the templates that matter first, typically Organization, BreadcrumbList, Article and Product with offer data. Validate every type after deployment, because a single malformed field can invalidate the whole block.
• Cannibalization. In Search Console, look for queries where the ranking URL changes month to month. That flip-flopping is the signature. Resolve it by consolidating the weaker page into the stronger one with a 301, or by rewriting each page to serve a genuinely different intent.
• Backlink profile. The most useful audit output is a list of broken backlinks, meaning external links pointing at URLs that now return 404. Redirect them to the closest live equivalent. Disavowal is rarely necessary for a normal site.
• Crawl budget waste. Only server logs tell you the truth here. If a large share of crawler requests hit parameter URLs, redirect chains or paginated tails, you are paying for crawling that will never be indexed. This matters most above roughly 10,000 URLs; below that it is usually a symptom of items 1 and 2 rather than a problem in its own right.

10. AI-citability, the newest line on the checklist

Assistants and AI answer panels quote passages, not pages. Five things make a page quotable: an answer-first opening paragraph that resolves the query in its first sentence; headings phrased the way people actually ask questions; FAQ schema pairing each question with a self-contained answer; tables and specific figures that are hard to paraphrase away; and an llms.txt file at the root pointing to your most useful documents. Also confirm robots.txt does not block the AI crawlers you want citations from, since crawler names and access rules change and are worth rechecking each quarter.

A 90-minute triage version

When there is no budget for a full audit, this compressed run finds the expensive problems. It is roughly the sequence used to produce a free SEO audit before scoping a larger engagement.

Frequently asked questions

How often should I run an SEO audit?

A full pass quarterly is enough for most sites. Check index coverage monthly, because it degrades quietly, and run the whole checklist immediately after any migration, replatform, template change or CMS upgrade. Sites publishing daily should also watch cannibalization more often than quarterly.

How long does a proper SEO audit take?

Approximately half a day for a small brochure site, one to two days for a few hundred pages, and considerably longer for large ecommerce catalogues where log file analysis and faceted navigation testing dominate the work. The triage table above takes 90 minutes and catches the majority of high-cost problems.

Which tools do I actually need?

Search Console and PageSpeed Insights are free and cover items 1, 3 and 7. You need a crawler for items 2, 4 and 5; free tiers typically cap at around 500 URLs, though vendor limits and pricing change. A backlink index covers item 8, and item 9 needs server log access, which is usually the hardest thing to get hold of.

Do I have to fix everything on the list?

No, and trying to is the most common way audits fail. Fix items 1 to 4, ship them, measure for four to six weeks, then reassess. Items 5 to 10 are real but smaller, and their value depends on the first four being sound. A good technical SEO engagement works the same way: sequence first, volume later.

Does an SEO audit help visibility in AI search?

Largely yes, because the foundations are shared. A page an AI assistant cannot crawl, or one buried five clicks deep behind duplicated URLs, will not be retrieved or cited. Item 10 adds a formatting layer on top of that, but it only pays off once items 1 to 4 pass.

Get a free SEO audit

← All articles