Most audits I inherit are 90–page PDF exports with 300 issues sorted by severity, and nobody has touched them in six months. They fail for the same reason every time: they document problems instead of ranking them by what the problem is actually costing the business.

A useful SEO site audit is shorter, more opinionated, and ends with a sequenced list of work. This is the process I run — what to crawl, what to check by hand, how to decide what matters, and how to turn the findings into rankings rather than a report nobody reads.

What Is an SEO Site Audit?

An SEO site audit is a structured review of a website's technical health, on-page quality, content performance, and authority signals, carried out to find the specific issues holding back organic visibility and to sequence the fixes by impact. It ends in a prioritised action plan, not a list of errors.

The distinction matters. A crawler will happily tell you that 412 pages have meta descriptions over 160 characters. That finding is real and almost worthless. The audit's job is to separate the noise from the four or five things genuinely suppressing traffic.

Audit, Health Check, and Diagnosis

Three different jobs get called the same thing:

• A full audit covers everything and takes several days. Right at the start of an engagement, after a migration, or annually.
• A monthly health check covers indexation, Core Web Vitals, broken links and new errors. Half a day.
• A diagnosis answers one urgent question — why did traffic drop on 14 March — and ignores everything unrelated.

Running a full audit when you needed a diagnosis wastes a week. Running a health check when the site has never been audited misses the structural problems entirely.

What You Need Before You Start

Gather access first. Half the delays in an audit come from waiting on logins.

• Google Search Console, with all property variants verified
• Analytics with at least 13 months of history so you can compare year on year
• A crawler — Screaming Frog, Sitebulb, or the site audit module in Ahrefs or Semrush
• Server log files if you can get them, ideally 30 days
• CMS admin access and, where possible, staging
• The business context: which pages make money, what changed recently, what the traffic goal is

That last item is the one people skip, and it is the one that makes an audit useful. Without knowing which 20 URLs generate revenue, every finding gets equal weight.

Step 1: Establish the Baseline

Before touching a crawler, plot organic sessions and clicks over 16 months and mark the inflection points. Then annotate them against known Google ranking update dates, site releases, and migrations.

A gradual decline usually points to content decay or competitors improving. A cliff edge points at a technical event or an algorithmic hit. Knowing which shape you are looking at changes the entire audit before you have crawled a single URL.

Segment the Site

Split URLs into templates — blog, category, product, location, landing pages — and pull clicks per segment. Problems are almost always template-level rather than page-level. If every product URL under one category is losing impressions, you are looking for one broken pattern, not 400 individual issues.

Step 2: Indexation and Crawlability

This is the first place real money is lost. If a page is not indexed, nothing else about it matters.

Check	Where	What Bad Looks Like
Indexed page count vs real page count	Search Console Pages report	Large mismatch in either direction
Discovered – currently not indexed	Search Console	Growing bucket, usually a quality or crawl budget signal
Robots directives	Crawler, robots.txt	Noindex or disallow on pages that should rank
Canonical tags	Crawler	Canonicals pointing to unrelated or redirected URLs
XML sitemap accuracy	Sitemap vs crawl	404s, redirects or noindexed URLs listed
Parameter and faceted URLs	Crawler, logs	Thousands of near-duplicate crawlable combinations

Work through the specifics of each in my notes on robots.txt, creating a sitemap, and the recurring discovered currently not indexed problem. If large sections are simply missing from the index, my guide on why a website is not indexing covers the diagnostic order.

Read the Log Files if You Can Get Them

Logs answer questions no crawler can: which URLs Googlebot actually requests, how often, and how much of the crawl is being spent on parameters, redirects and error pages. On sites over roughly 10,000 URLs this regularly turns up 40 percent of crawl activity going somewhere worthless.

Step 3: Site Architecture and Internal Linking

Crawl depth is one of the most reliable predictors of whether a page performs. Pull a depth distribution from your crawler and look for commercially important URLs sitting four or more clicks from the homepage.

• Orphan pages: URLs in the sitemap or analytics with zero internal links. Fix by linking them from a relevant hub, or remove them.
• Thin hubs: category pages that link to children but carry no content of their own.
• Link concentration: which pages receive the most internal links, and whether those are the pages you want ranking.
• Redirect chains: every hop dilutes signals and wastes crawl.
• Anchor text: hundreds of internal links reading "read more" tell Google nothing.

My guides on fixing orphan pages, redirect chains, and website architecture go deeper on each.

Step 4: On-Page and Content Quality

Now move from whether pages can be found to whether they deserve to rank. Export every indexable URL with clicks, impressions, average position, word count, and last updated date, then sort into four buckets.

Bucket	Signal	Action
Performing	Steady clicks, position 1–5	Leave alone, refresh annually
Nearly there	Impressions, position 6–20	Highest priority — expand, match intent, add internal links
Decaying	Clicks down year on year	Update facts, re-check the SERP, rewrite intro
Dead weight	No clicks, no impressions, 12 months old	Consolidate, redirect or remove

The "nearly there" bucket is where audits pay for themselves. Pages already ranking 6 to 20 have proven relevance; they usually need better coverage of the subtopics competitors include, a tighter title, and three or four internal links from related content.

Duplicate and Overlapping Content

Look for several URLs competing for one query. Search Console's query report will show two or three of your own pages rotating for the same term — a clear consolidation candidate. My guide to fixing duplicate content issues covers when to canonicalise, when to merge, and when to leave well alone.

Titles, Headings and Intent

For your top 50 commercial URLs, check the title tag against the actual SERP: what format is ranking, what modifiers appear, and whether Google is rewriting your title. A rewritten title is a signal that yours does not describe the page well. Check the H1 exists and is unique, and that the first 100 words answer the query rather than warming up.

Step 5: Performance, Mobile and Rendering

Pull field data from the Core Web Vitals report rather than relying only on lab scores, which frequently disagree with what real users experience.

• LCP — usually a hero image, a font, or a render-blocking script.
• INP — heavy JavaScript on interaction, often third-party tags.
• CLS — images and ad slots without reserved dimensions.
• Mobile parity — confirm the mobile version contains the same content and links as desktop; indexing is mobile-first.
• JavaScript rendering — compare the raw HTML against the rendered DOM for your key templates.

Guidance for each metric is documented on web.dev. For diagnosis, see my walkthroughs on Core Web Vitals failures and JavaScript rendering issues.

Step 6: Off-Page Signals

Authority is the slowest thing to change, so audit it for direction rather than absolute numbers.

• Referring domain growth over 12 months, compared against two competitors
• Which pages hold the links, and whether they are pages that matter commercially
• Anchor text distribution, watching for unnatural exact-match clusters
• Lost links worth reclaiming, particularly from redirects and removed pages
• Unlinked brand mentions

Use a backlink checker to pull the profile and my notes on backlinks for how to read it. If growth is flat while competitors add 50 domains a quarter, that is your medium-term constraint regardless of how clean the technical picture looks.

Step 7: Write the Report People Will Actually Use

Five sections, no more:

• Headline: three sentences on the single biggest constraint.
• Evidence: the traffic chart, the segment table, and the two or three findings that explain it.
• Action plan: every fix with an owner, effort estimate, and expected impact.
• Sequenced roadmap: quick wins in weeks one to four, structural work in months two and three, authority work ongoing.
• Appendix: everything else, so it exists but does not distract.

A useful discipline is to cap the main report at ten recommendations. If a finding cannot displace something already on that list, it belongs in the appendix. My template for a readable SEO report follows the same shape.

Audit Cadence and Tooling

Task	Frequency	Typical Tool
Full audit	Annually, or after a migration	Screaming Frog plus manual review
Index and error check	Monthly	Search Console
Core Web Vitals	Monthly	Search Console field data
Content decay review	Quarterly	Search Console plus analytics
Backlink review	Quarterly	Ahrefs or Semrush
Log file analysis	Twice a year, large sites	Server logs plus Screaming Frog Log Analyser

You can run a credible first pass entirely on free tooling — Search Console, the free Screaming Frog tier, and a manual SERP review will surface most structural problems. My roundup of free SEO tools lists workable options, and the fuller website audit checklist gives you a line-by-line version of this process.

How to Prioritise What You Find

Every audit produces more findings than anyone will act on. Scoring them properly is the difference between a plan that gets executed and a backlog that gets ignored.

Score on Three Axes

Rate each finding one to five on business impact, confidence that the fix will work, and effort required. Impact means proximity to revenue — an issue on a category page that drives enquiries outranks the same issue on an old blog post. Confidence is honest self-assessment: a noindexed money page is a five, a speculative content rewrite is a two. Divide impact by effort and the order writes itself.

Separate Blockers from Improvements

Some findings prevent everything else from working. A page that is not indexed cannot benefit from better copy; a redirect chain undermines every link you build to it. Fix blockers first even when they look small, then move to improvements. In practice that usually means indexation and canonical problems in week one, internal linking and on-page work in weeks two to four, and content or authority builds after that.

Agree the List with Whoever Will Build It

An audit handed to a developer without a conversation loses about half its recommendations to competing priorities. Walk through the top ten items, let them estimate effort in their own terms, and adjust the sequence to fit a release cycle you already have. Findings that fit an existing sprint get done; findings that require a special project rarely do. My SEO audit overview and technical SEO checklist both work well as the shared reference document in that conversation.

Mistakes That Make Audits Useless

• Exporting the tool's report as the deliverable. The analysis is the value, not the crawl.
• Weighting every issue equally. Two missing alt attributes are not a priority next to a noindexed category tree.
• Auditing without business context. Fixes on pages nobody buys from do not move revenue.
• No re-crawl. Verify fixes were actually deployed — a surprising number never reach production.
• Ignoring what competitors changed. A drop with no site changes usually means someone else improved.

Why Work With Sadekur Rahman

I run audits to answer one question: what is the cheapest change that produces the largest gain in organic revenue over the next 90 days. That means a short report, an explicit sequence, and a re-crawl afterwards to confirm the work landed — not a PDF handed over and forgotten.

If your site needs that, look at my SEO services and case studies, compare plans and pricing, or book a meeting to walk through your current data.

Frequently Asked Questions

How long does an SEO site audit take?

Two to four days of work for a site under 5,000 URLs, and one to two weeks for a large ecommerce or multi-market site where log analysis and template testing are needed. Access delays add more time than the analysis itself.

How often should I audit my site?

A full audit once a year, plus a monthly index and error check and a quarterly content decay review. Migrations, replatforming and sharp traffic drops all justify an unscheduled audit immediately.

Can I run an audit myself without paid tools?

Yes, for the fundamentals. Search Console covers indexation, queries and Core Web Vitals, and the free Screaming Frog tier crawls 500 URLs. Backlink analysis and competitor comparison are where free tooling runs out.

How soon will fixing audit issues improve rankings?

Indexation fixes can show within days of recrawling. On-page improvements to pages already ranking 6 to 20 typically move in four to eight weeks. Architecture and authority work takes three to six months to settle.