Article
What a Real SEO Audit Checks (and What the Free Ones Skip)
By INTSEO Media SEO Team10 min read
A real SEO audit diagnoses why important URLs underperform and produces a sequenced backlog tied to business outcomes. A free automated scan lists issues. In July 2026, the gap is still judgement: which findings touch money pages, which are noise, and what should ship first. If the output cannot become a roadmap with owners, it was a report, not an audit.
What an audit is for
The job is decision support under uncertainty. You are trying to explain the current organic system: can key pages be indexed, do they match the intents that matter, is content coverage adequate, and does the site have enough corroboration to compete? Then you turn that explanation into work.
Audits are time-bounded. They are not a substitute for ongoing reporting. They should establish baselines, prioritised hypotheses, and quick verification steps. A good audit makes the next ninety days obvious. A weak one dumps severity labels and walks away.
Scope matters. A six-page brochure site and a faceted ecommerce catalogue need different depth. Agree on markets, templates, and success metrics before crawling begins. Otherwise you get a generic checklist in a custom PDF cover.
What free automated audits miss
Free or cheap scanners are useful for inventory. They are bad at strategy. They often overweight easy-to-detect technical nits and underweight intent mismatch, weak offers, cannibalisation, and competitive gaps.
| Scanner often flags | What it might mean | What an audit still asks | Typical priority mistake |
|---|---|---|---|
| Missing alt text | Accessibility and image clarity debt | Does this hit money templates or rare blog images? | Sitewide alt projects before commercial rewrites |
| Thin pages | Low word count or low unique content | Is the URL meant to be thin (login, filters) or a failed money page? | Padding word count instead of consolidating |
| Slow URLs | Performance risk | Are these URLs in the conversion path? | Optimising forgotten landing pages first |
| No schema | Missing structured data | Would rich results or entity clarity actually change outcomes here? | Schema theatre with no SERP upside |
| Broken links | Crawl and UX debt | Which broken links sit on high-traffic templates? | Fixing ancient footer orphans before nav issues |
Automation also struggles with JavaScript rendering nuances, soft 404 behaviour, international hreflang intent, and whether a ranking page is simply the wrong answer for the query. Humans still have to open SERPs and analytics.
Use scanners to generate candidates. Use practitioners to accept or reject them against money URLs. That is the same triage mindset as prioritising SEO fixes on a limited budget.
Business and market context
A real audit starts before Screaming Frog. Who is the customer, what is a qualified lead or sale, which geographies matter, and what can the company actually ship? An audit that ignores CMS constraints or legal review queues produces fiction.
Review analytics and Search Console with brand/non-brand splits, landing page contribution, and conversion definitions finance will accept. Look at seasonality and recent releases. Many “SEO problems” are release regressions or measurement breaks.
Competitive sampling should be SERP-specific. Who ranks for the money queries, with what page types, and what gaps are obvious? Broad domain metrics without query context flatter or frighten without instructing.
Technical and indexation layer
Technical review asks whether search systems can fetch, render, and index the right URLs efficiently. Coverage issues, robots rules, canonical conflicts, status codes, sitemaps, pagination, facets, duplicate clusters, and Core Web Vitals on key templates all belong here.
The audit should distinguish sitewide hygiene from template-level blockers. A noindex on the main category template is an emergency. A pile of 404s on retired UTM experiments is housekeeping. Technical SEO findings only earn priority when they change access to demand.
Rendering and internal linking deserve special attention on JavaScript heavy sites and large catalogues. If important content is invisible to crawlers or buried five clicks deep, relevance work on that content will under-read in results.
Relevance, content, and SERP fit
This layer checks whether money URLs say the right thing to the right intent. Titles, headings, information architecture, thin or outdated copy, cannibalisation between URLs, and missing page types all appear here. SERP inspection is mandatory: if competitors rank with comparison tables and you offer a vague overview, the audit should say so plainly.
Content coverage analysis looks for gaps and decay. Which topics earn demand you do not address? Which old winners are slipping? Decay belongs in audits, not only in ongoing retainers. So does clarifying whether blog posts are stealing commercial intents from service or category pages.
On-page recommendations should be template-aware. Fixing one hero page by hand is fine for proof. Sustainable gains usually need pattern fixes across a template: shared title patterns, module order, and internal link blocks that every URL in the set can inherit.
Corroboration and competitive reality
Off-site review asks whether other sites and entities support your right to compete for the topic. Link quality and relevance, toxic or accidental patterns, unlinked brand mentions, local citations for location-led businesses, and digital PR gaps all sit here.
Raw referring domain counts without destination analysis mislead. Links into useless URLs, or from irrelevant networks, do not fix a weak category page. An audit should map corroboration to the URLs you want to move, and say when on-site work must come first. Off-site advice without that map is shopping, not diagnosis.
Local and ecommerce programmes need specialised checks on top of the same spine: Business Profile integrity, NAP consistency, facet indexation, and category strength. Use a shared vocabulary across technical, relevance, content, and corroboration so sections do not turn into silo theatre.
What a useful deliverable looks like
Expect an executive summary in plain language, a prioritised backlog with effort and impact notes, evidence for major findings (screenshots, URL lists, queries), and explicit non-findings (“we checked X; it is not the bottleneck”). Appendices can hold raw exports. The core doc should stay readable by a non-specialist operator.
Useful audits also define verification: how you will know a fix worked. “Improve rankings” is vague. “Restore indexation for these twenty category URLs and recover non-brand clicks within two review cycles” can be managed.
Be wary of deliverables that are only tool screenshots, only severity heatmaps, or only a sales bridge into a retainer with no diagnostic clarity. For vendor evaluation questions that pair with audit literacy, see choosing an SEO agency.
Takeaway
A real SEO audit connects crawl facts to business URLs, SERP realities, and a sequenced backlog. Free scanners can inventory symptoms. They cannot assign meaning, ignore noise, or respect shipping constraints. Demand context, evidence, prioritisation, and verification steps. Keep the PDF only if it can survive contact with a sprint board.
When you commission or perform the next audit, insist on a money-URL sample early, a clear split between blockers and polish, and a ninety-day plan that names owners. That standard protects you from both inflated decks and false comfort from green automated scores that never touched revenue.