Ask five agencies what is included in an SEO audit and you’ll get five checklists, most of them padded to look thorough. The dirty secret is that the checklist is the easy part — any crawler spits out 300 “issues” in ten minutes. What separates a real audit from an expensive PDF is triage: knowing which of those 300 findings will actually move traffic, which are cosmetic, and which are outright false positives you should delete before they waste a developer’s week. This guide covers everything a complete audit examines, but more importantly, how to tell signal from noise in the output.
An Audit Is a Diagnosis, Not an Inventory
The failure mode of most audits is treating “we checked it” as the deliverable. A page can have a missing meta description, a slow Largest Contentful Paint, and three H1 tags — and none of those may be why it isn’t ranking. The real cause might be that four other pages on the same site target the same query, so Google can’t decide which one to trust. A checklist audit flags the three cosmetic issues and misses the one that matters. So when you ask what is included in an SEO audit, the honest answer has two parts: the categories examined (below) and the reasoning that connects a finding to lost traffic. Without the second part, you have an inventory, not a diagnosis.
The mental model I use ranks every finding on three axes: traffic impact (does fixing it plausibly recover or unlock sessions?), effort (a title-tag edit vs. a re-platform), and certainty (do I actually know this is causal, or am I guessing?). High-impact, low-effort, high-certainty findings go to the top. Everything else waits. Ten findings ranked this way beat two hundred sorted alphabetically by category.
Layer 1: Crawlability and Indexation
This is the foundation, because a page Google can’t reach or won’t store cannot rank at any quality level. The audit checks robots.txt for accidental blocks, the XML sitemap for stale or missing URLs, HTTP status codes (301 chains, soft 404s, 5xx errors), canonical tags pointing where they should, and stray noindex directives left over from a staging site. The single most useful artifact here is Search Console’s Pages report — it tells you exactly which URLs Google excluded and why (“Crawled – currently not indexed” usually means thin or duplicate content, not a technical bug).
A good deliverable names the specific URLs, not just a count. “47 pages have canonical issues” is useless. “These 12 product pages canonicalize to a deleted category, so their equity is bleeding into a 404” is actionable.
Layer 2: On-Page Elements at Scale
This is what tools do best, and where thin audits pad their page count. The crawler reports on title tags (missing, duplicated, truncated past ~60 characters), meta descriptions, heading hierarchy, image alt text, internal linking depth, and structured data validity. The value isn’t the list — it’s the evidence. A report that shows the actual duplicated titles and the URLs sharing them lets you fix in an afternoon. A report that says “18% of pages have title issues” sends you hunting.
- Titles & meta: duplication and truncation matter; a missing meta description does not (Google rewrites it ~70% of the time anyway).
- Headings: hierarchy for scannability and entity clarity, not a rigid “exactly one H1” rule Google stopped caring about years ago.
- Internal links: orphan pages and shallow crawl depth are real problems; “not enough internal links” as a raw number usually is not.
- Structured data: fix errors that break rich results; ignore “recommended” property warnings that change nothing.
Layer 3: Site Speed and Core Web Vitals
A complete audit separates field data (real Chrome users, the numbers Google actually uses) from lab scores (a simulated test that’s directional at best). The three Core Web Vitals are Largest Contentful Paint, Cumulative Layout Shift, and Interaction to Next Paint. Here’s the honest caveat most reports won’t give you: Core Web Vitals are a genuine ranking input but a small one. A page that fails LCP but nails intent will still outrank a fast page that answers the wrong question. Fix speed because it lifts conversions and reduces bounce — not because you expect a five-position jump from shaving 400ms.
Layer 4: Content Quality and Coverage
This is the layer tools handle worst and that matters most. Automated content scores are proxies; they can’t judge whether your page actually satisfies the searcher. A real content audit builds an inventory of every URL with its traffic, impressions, and average position, then assigns each a verdict: keep, improve, consolidate, or remove. It flags cannibalization — the silent killer where two pages compete for the same query and split the signal, so neither ranks. And it benchmarks against the weakest page-one competitor, not the market leader, because that’s your realistic bar. This is where SEO Rocket’s competitor gap analysis earns its place: it runs content and keyword-gap comparisons across up to five real rivals on live Ahrefs data, surfacing the topics they rank for that you don’t — the highest-leverage findings in most audits.
Layer 5: Backlinks and Authority
The link audit maps referring domains, domain rating trend, anchor-text distribution, lost and newly earned links, and any genuinely toxic domains. Two honest caveats: first, “toxic” scores from tools are wildly overcounted — Google ignores most spam links automatically, and reflexively disavowing can do more harm than good. Second, domain rating is a third-party estimate, not a Google metric; treat it as directional. The useful output of a link audit isn’t a disavow file — it’s a gap: the domains linking to three of your competitors but not to you, which a real outreach campaign could plausibly win.
Layer 6: Measurement and Tracking Setup
The most-skipped layer. If your analytics are broken, every other finding is measured against a lie. The audit verifies Search Console and GA4 are connected and de-duplicated, conversion events actually fire, internal and bot traffic is filtered out, and rank tracking pulls top-100 snapshots rather than volatile single-day checks. It establishes the “noise floor” — daily two-to-three position jitter is normal, so a keyword that “dropped” from 8 to 10 overnight moved nothing. SEO Rocket handles this layer with a real-crawler site audit plus built-in rank tracking and AI-visibility tracking (how often you’re cited in AI Overviews and chatbots), so the diagnosis and the ongoing measurement live in one place instead of five tabs.
A Worked Micro-Example: One Page, End to End
Say a blog post targeting “commercial flooring maintenance” sits at position 14 and gets 30 clicks a month. A checklist audit flags: missing meta description, no schema, LCP of 3.1s. A diagnostic audit does something different. It pulls Search Console and sees the page ranks for 40 queries but is outranked on its own primary term by another post on the same site about “how to clean commercial floors” — classic cannibalization. It checks the page-one results and finds the weakest one is a 900-word piece with no maintenance-schedule table. The verdict: consolidate the two competing posts into one, add the schedule table and an FAQ, and 301 the loser. The meta description and schema are noted as five-minute cleanups, ranked near the bottom. That’s the difference — the diagnostic audit found the position-14 cause; the checklist found three symptoms.
Scope, Timeline, and Realistic Price
What is included in an SEO audit scales with the tier you buy, and the tiers are roughly:
- Quick scan — minutes, often free. A crawler pass and a Core Web Vitals check. Useful for triage, not for a strategy.
- Standard audit — typically one to two weeks, commonly in the low four figures for a mid-size site. Covers all six layers with prioritized recommendations. This is what most businesses actually need.
- Enterprise audit — four to eight weeks for large or multi-domain sites, priced into five figures. Log-file analysis, crawl-budget modeling, international/hreflang, and migration risk.
Prices vary widely by market and consultant, so treat these as ranges, not quotes. A cheap crawl dump masquerading as a “full audit” is common — the tell is whether you get prioritized reasoning or a raw issue list.
What an Audit Cannot Tell You
Honesty gate: an audit is a snapshot of correlations, not a guarantee of causation. It can tell you a page is thin and un-linked; it cannot promise that fixing both will rank it, because intent, competition, and domain authority all interact. It can’t see Google’s algorithm directly, so every “this hurts you” is an informed hypothesis. And it’s perishable — a core update, a competitor’s new content, or a site migration can invalidate findings within weeks. The right frame is a working diagnosis you re-check quarterly, not a permanent verdict.
How to Judge the Deliverable
When the report lands, apply three tests. First, is it prioritized? Ten findings ranked by impact beat two hundred grouped by category. Second, does it include evidence — actual URLs, actual title text, actual GSC data — or just percentages? Third, does it tell you what’s working, not only what’s broken, so you don’t accidentally “fix” a page that’s already ranking? A deliverable that passes all three is worth paying for. One that fails them is a crawler export with a cover page. This is the exact standard SEO Rocket’s audit and AI writer are built to — a playbook proven across 1,000,000+ ranking pages, where the validation is in the reasoning, not the page count.
Frequently Asked Questions
How long does an SEO audit take?
A quick automated scan runs in minutes. A proper standard audit covering all six layers with prioritized recommendations typically takes one to two weeks — most of that is the human analysis, not the crawl. Enterprise audits with log-file and crawl-budget work run four to eight weeks.
Can I do an SEO audit myself?
Yes, for the technical and on-page layers — Search Console, a crawler, and PageSpeed Insights are free and cover a lot. The layers that need experience are content triage (keep/improve/consolidate/remove decisions) and connecting findings to causes. That judgment is where DIY audits usually go wrong, flagging cosmetics and missing cannibalization.
How often should I run an SEO audit?
A full audit once or twice a year, plus a lightweight technical scan monthly and continuous rank and index monitoring in between. Audits are perishable — core updates and competitor moves change the picture — so treat the big audit as a reset and the ongoing tracking as your early-warning system.
What’s the difference between a technical audit and a full SEO audit?
A technical audit stops at crawlability, indexation, speed, and site architecture — layers one to three. A full audit adds content quality, backlink authority, and measurement setup, then ties everything into a prioritized action plan. The technical audit tells you the site works; the full audit tells you why it ranks or doesn’t.
