On a large site, scoping is most of the work. A crawl can return a million URLs, but a few templates usually explain most of what matters.
Quick answer
Agree what the audit is for, segment the site by template and section, choose the data sources, sample where a full crawl is not practical, and decide what the deliverable looks like before starting. Then review each segment against its own risks, in an order that follows commercial priority.
Step 1: agree the question
An audit with no question becomes a catalogue of everything wrong. Ask what the business needs to know.
- Why is organic performance not matching the strength of the product or content?
- Is a migration, replatform or redesign coming?
- Which parts of the site earn revenue or leads?
- What can the team realistically change, and by when?
The answers decide what to leave out as much as what to include.
Step 2: segment by template
Treat a large site as a set of page types, not a set of URLs. A typical ecommerce site breaks down like this:
- Category and listing pages
- Product pages
- Faceted and filtered URLs
- Search and parameter pages
- Editorial or blog content
- Utility pages such as basket and account
Each type has its own risks and its own owner. On a storage retailer with 288 indexed URLs, one template, the faceted navigation, produced around 130 near-duplicate pages. On a larger site the same fault could produce thousands. Finding the template is the point of segmenting.
Step 3: choose the data
No single source is complete. Combine them.
- A crawl, ideally of the whole site, or a sample per template if the site is too large
- Search Console for impressions, clicks, indexing status and query data
- Analytics for landing page traffic and conversions
- Server logs, where available, to see what Google actually crawls
- A list of the pages the business cares about most
Before any of this, compare your expectation of the site's size with what the crawl finds. On my own migration, I expected about 40 URLs and the crawl found 101. On a large site, the gap between expected and actual page counts is often wider and always worth explaining.
Step 4: decide how to prioritise
On large ecommerce sites, prioritise by commercial value and by risk.
- Categories and listings that drive revenue come first
- Templates that generate many low-value URLs, such as filters, come next, because they affect crawling and duplication at scale
- Product pages follow, with attention to stock status, variants and structured data
- Editorial content is reviewed for overlap and decay
Fixes to one template reach every page that uses it, so a single change on a category template is often worth more than dozens of page-level edits.
Step 5: set the deliverable
Agree this before the work starts.
- A scorecard by area, and a prioritised plan with owners
- How findings are presented, for example by template, not by URL
- Which findings need evidence from the crawl and which from Search Console
- What is out of scope, written down
See how to turn an SEO audit into an action plan for how the output should look.
Step 6: plan for implementation
Large sites have release cycles, approvals and several teams. A scope that ignores this produces a plan nobody can deliver. Ask how changes get released, and shape the plan around it. See why recommendations do not get implemented.
Common mistakes
- Starting with the crawler instead of the question
- Auditing every URL with the same depth
- Missing the templates that generate the most pages
- Skipping Search Console and relying on a crawl alone
- Agreeing no deliverable up front
When to get support
For a large or complex site, the Visibility Review (part of my SEO audits work) can be scoped by template and priority, and Technical SEO covers deeper crawl and indexing work. The technical SEO checklist for large websites covers the checks themselves.
FAQ
Do I need to crawl the whole site?
If it is feasible, yes. If the site is too large, crawl a representative sample of each template and use Search Console to check the rest.
How long does a large-site audit take?
It depends on size, access to data and how many templates matter. Scoping first gives a realistic estimate, so it is worth doing before agreeing a timeline.
Is a larger audit better?
No. Focus beats volume. A scoped audit that covers the templates that matter is more useful than a full export nobody reads.