Guides
What Is an SEO Audit? Scope, Evidence & Deliverables
An SEO audit is not a score. It is an evidence-based review that connects search performance problems to reproducible causes, owners, and validated fixes.
Run a fresh DomainLens audit and use the report as your priority list.
An audit explains why—not only what failed
A useful SEO audit answers four questions: can search engines discover and retrieve the URL, can they render and index the intended version, does the page satisfy a real query better than alternatives, and can users complete the task without avoidable friction? The report then connects each failure to affected URLs, business impact, a fix owner, and a verification step.
The four layers of an SEO audit
These layers interact but should not be conflated. A poor LCP is not a robots problem; an indexable page is not necessarily worthy of indexing; a canonical is not a substitute for removing a useless duplicate from navigation and sitemap.
| Layer | Evidence | Typical decisions |
|---|---|---|
| Discovery and crawling | Links, sitemap, robots.txt, logs, status codes | Expose orphan pages, remove crawl traps, repair broken paths |
| Indexing and canonicalization | Robots meta, canonical, rendered HTML, GSC states | Choose indexable pages, consolidate duplicates, fix conflicting signals |
| Relevance and content | Queries, intent, page copy, competing URLs, internal anchors | Rewrite, merge, redirect, or create clearer information architecture |
| Experience and performance | CWV field data, lab traces, mobile UX, conversion paths | Fix shared template bottlenecks and interaction failures |
Collect a baseline that can survive the audit
- Export Search Console pages, queries, clicks, impressions, index states, and sitemap status for the agreed date range.
- Record analytics landing pages and conversions by device and country.
- Crawl the canonical production host and retain raw URL-level results.
- Group URLs by template and indexability rather than reviewing a flat list.
- Save representative rendered HTML, screenshots, headers, and performance traces for high-impact findings.
- Document releases, migrations, domain changes, CMS constraints, and seasonal events before attributing traffic changes.
Turn evidence into a prioritized backlog
A noindex on every product page outranks metadata polishing. A template-wide wrong canonical outranks one broken image. A proposed rewrite with no query evidence should not outrank a reproducible checkout interaction failure simply because content is easier to assign.
| Factor | Question |
|---|---|
| Impact | How much traffic, revenue, or index coverage can this change affect? |
| Reach | Is this one URL or a shared template affecting thousands? |
| Confidence | Can the cause be reproduced, or is it only a correlation? |
| Effort and risk | What engineering/editorial work and regression risk are involved? |
| Dependency | Must another migration, tracking fix, or content decision happen first? |
What the final deliverable should contain
- 1Executive summary: the few causes that explain most of the missed opportunity.
- 2Scope and limitations: hosts, locales, data windows, authentication, crawl limits, and unavailable sources.
- 3Finding cards: evidence, affected examples/count, expected outcome, owner, effort, priority, and validation.
- 4URL decisions: keep, improve, consolidate, redirect, noindex, or remove—with destination where applicable.
- 5Implementation sequence: dependencies and release-sized batches rather than one undifferentiated spreadsheet.
- 6Measurement plan: baseline, deployment annotation, leading checks, lagging organic metrics, and review dates.
Where DomainLens fits
DomainLens supplies repeatable URL evidence: fetch status, robots, canonical, headings, metadata, links, structured data, security, and performance diagnostics. Run it on representative URLs from every template and attach the relevant output to findings.
It does not know your margins, content ownership, strategic query set, or whether two pages should be merged. Combine the automated evidence with Search Console, analytics, server logs where available, and manual review. Then rerun the same URL sample after each release instead of declaring success when a ticket closes.
For the response-layer portion of the work, use the technical SEO checklist as a verification sequence rather than expanding the audit scope ad hoc.