Do I need paid tools to audit a website?

No. Search Console, a free crawler and a spreadsheet cover the six steps. Paid tools speed up crawling and historical tracking, they do not replace judgement.

What is the first thing to fix after an audit?

Whatever blocks indexation of pages that already earn impressions. Fixing a page Google cannot index produces no result, no matter how good the content is.

How often should a website audit be repeated?

A full pass each quarter and a light crawl monthly. Re-run the full audit after any migration, template change or major content release.

How to Do a Website Audit: Six Steps, Three Numbers

WebSEOGrowTechnical SEO and content Talk to an auditor

Audit steps · field guide

How to Do a Website Audit: Six Steps, Three Numbers Each

A website audit is not a screenshot collection. It is a fixed sequence of checks that ends with one ranked list: what to fix first, what to ignore, and what will still matter after the next template update. Below is the exact order we use, and the only three figures you need to write down at each stage.

One working day of data collection. Two pages of findings. No dashboards you will never open again.

The six stages are fixed. Skipping stage one is the most common reason a finished audit cannot be defended six months later.

What does a website audit actually check?

A website audit checks four things a search engine needs before it can rank a page: whether it can reach the URL, whether it keeps the URL in its index, whether it can read the page structure, and whether the page deserves to exist at all. Every finding you write down belongs to one of those four, and anything that fits none of them is a design opinion, not an audit item.

Before step one

Why most audits change nothing

We have inherited dozens of audit documents that were 40 pages long and produced zero deployments. The failure is rarely the tool. It is one of four habits.

01Recording symptoms

"Meta descriptions are missing on 214 pages" is a symptom. It says nothing about whether those 214 pages receive impressions, earn clicks, or convert. An audit item without a traffic figure attached cannot be prioritised, so it sits in a spreadsheet until someone is embarrassed enough to close it.

02Skipping the baseline

If you never record clicks, impressions and revenue per template before the work starts, you can never separate your improvements from a seasonal dip or a competitor outage. The baseline is not paperwork. It is the only thing that makes the next quarter defensible.

03Auditing past the first week

An audit that takes three weeks has already drifted: pages were published, redirects were added, and the crawl no longer matches the live site. Fast, shallow and current beats slow, exhaustive and stale every single time.

04Treating every issue as equal

A broken link in a footer and a noindex tag on a category page both go in the "issues" column. One costs a small amount of crawl budget. The other removes a page that earns enquiries. If your report does not separate them, your developer will fix them in the order they appear.

The method

Six steps, three numbers per step

Each step below asks the same two questions: what do you look at, and what do you write down. Three figures per step is deliberate. It forces a decision about what matters and keeps the final report short enough that someone will read it.

01Baseline

Record where you are before you touch anything

Open Search Console and your analytics, set the range to the last 90 days, and split by landing page and by template. You are not looking for problems yet. You are freezing a photograph of the site as it stands, because every later judgement depends on this photograph existing.

What to look at

Clicks and impressions by page group, conversions by template, and the date of the last structural change. If nobody remembers the last change, that itself is a finding.

What to record

90-day clicks, indexed pages, converting templates. Three numbers, one row, dated. Everything downstream references this row.

02Crawl

Crawl the site the way a bot would, not the way a browser would

Run a crawler with JavaScript rendering off first. Most commercial sites are readable that way, and the plain-HTML view exposes exactly what a crawler sees when scripts are ignored. Then compare the count with your sitemap and with the pages your navigation actually links to.

What to look at

Status codes, redirect chains longer than one hop, orphan URLs that only exist in the sitemap, and parameter pages multiplying every filter combination.

What to record

Total URLs found, non-200 URLs, orphan URLs. The gap between URLs found and URLs linked is your crawl budget leak.

03Index

Compare what you submitted with what actually got indexed

Search Console pages report splits every URL into indexed, crawled but not indexed, discovered but not indexed, and excluded. Read the exclusions by reason, not by count. A thousand "alternate page with proper canonical" is healthy. Sixty "soft 404" is not.

What to look at

Whether your money pages sit in the indexed bucket, whether canonical tags point at themselves, and whether robots rules block anything that earns impressions.

What to record

Indexed share, pages excluded by noindex or canonical, impressions lost from excluded URLs. The third number is the one that gets budget approved.

04Pages

Check the page-level basics on the templates that earn money

Do not audit all 4,000 URLs. Group them into templates, then sample five pages per template. Titles, headings, internal links and image alt text repeat by template, so five pages tell you the truth about five hundred. Spend the saved time on the templates that convert.

What to look at

Duplicate or templated titles, two H1s on one page, product copy that stops at 60 words, and internal links that all point at the homepage instead of the next step in the journey.

What to record

Templates audited, pages per template, pages with a unique title. If that third figure is low, one template fix moves hundreds of URLs at once.

05Content

Decide which pages deserve to exist

This is the step almost every DIY audit skips because it requires a judgement rather than a screenshot. Take the pages that receive impressions but almost no clicks, and ask what query each one is trying to answer. Where two pages answer the same query, one of them is cannibalising the other and should be merged or rewritten.

What to look at

Query overlap between pages, thin category text, and blog posts that rank for nothing but sit in the sitemap feeding crawl budget to nobody.

What to record

Pages with overlap, thin pages under 300 words, pages to merge or retire. Deleting is a legitimate audit output, not a failure.

06Priority

Rank every item by money at risk, not by tidiness

Multiply the traffic a fix could protect or unlock by the commercial value of that traffic, then divide by the effort. That ratio sorts the list for you. The result usually surprises people: the noindex tag on a category page outranks forty broken footer links that look worse in a screenshot.

What to look at

Which items block indexation, which unblock revenue from existing pages, and which are cosmetic. Three buckets is enough; more buckets means nothing gets done.

What to record

Items in bucket one, combined monthly impressions protected, hours of development needed. That last figure is what your developer will ask about first.

The record sheet. If a row has fewer than three numbers, the step is not finished.
Step	What you watch	Number one	Number two	Number three
01 Baseline	Clicks, impressions, revenue by template	90-day clicks	indexed pages	converting templates
02 Crawl	Status codes, orphans, parameter sprawl	URLs found	non-200 URLs	orphan URLs
03 Index	Exclusion reasons in Search Console	indexed share	excluded money pages	impressions lost
04 Pages	Titles, headings, internal links, alt text	templates sampled	pages per template	unique titles
05 Content	Query overlap, thin pages, dead posts	overlapping pages	thin pages	merge candidates
06 Priority	Traffic at risk versus effort	bucket-one items	impressions protected	dev hours

Want the six steps run on your site this week?

Send the domain and we will come back with the baseline, the crawl gap and the three fixes worth doing first. No deck, no retainer pitch.

Step six, extended

The first 72 hours after the audit

An audit has a shelf life. Pages get published, a developer ships a template change, and your crawl is history. Three days is enough to lock in the findings and start the highest-value fix while the evidence is still fresh.

First 24 hours

Freeze and share

Export the crawl and the Search Console pages report to one dated folder
Send bucket one to whoever controls the deploy, not to a committee
Write the one-sentence version: what is broken, what it costs per month
Book the deploy slot before the details are debated

24 to 48 hours

Ship the smallest blocking fix

Remove the noindex, the stray canonical or the robots rule that hides a money page
Resubmit the affected URLs and confirm the indexed count moves
Repoint orphan URLs with one internal link each, in the template, not one by one
Add the missing numbers to the record sheet the same day

48 to 72 hours

Set the recheck

Schedule a re-crawl for day 14, not day 90
Note the two items you deliberately deferred and why
Add the next template fix to the sprint that owns the header
Re-read the baseline before anyone claims a win

Honest limits

Where a do-it-yourself audit stops

You can run all six steps yourself with free tools. What you cannot easily do alone is the judgement in steps five and six, because it requires knowing which pages convert, what the margin looks like, and what your developer can actually ship this sprint. That context rarely lives with the person holding the crawler.

The other gap is scale. On a catalog of 30,000 URLs, the crawl takes hours and the template grouping takes a full day, and by then the first crawl is out of date. At that size the work needs a pipeline rather than a browser tab.

If you want the six steps run with that commercial context attached, our SEO Audit Services follow exactly the order above and end with a ranked list you can hand to a developer without translation. If you would rather keep it in-house, start with a written SEO Audit Checklist so that two people running the same audit produce the same numbers.

What a paid audit should contain

A dated baseline you can compare against in six months
The crawl gap, stated as a number, not a list of examples
Every money template checked, not just the homepage
A ranked list with effort estimates your developer accepts
Explicit "do not fix this" calls, so the queue stays short
A recheck date and the metric that decides success

Inside the engagement

How we run the same six steps

Same order, different ownership. You keep the account access, we do the crawl, the grouping and the ranking, and you get a document a developer can start from on Monday morning.

Day oneAccess and baseline. Search Console, analytics and the last deploy date are recorded before any change.

Day twoCrawl plus index comparison. Every exclusion reason is read, not just counted.

Day threeTemplate sampling and content overlap. Five pages per template, then the merge list.

Day fourRanking, effort estimates and the 72-hour action list, delivered as one page plus an appendix.

One example, no names

A 214-page industrial parts catalog came to us after a redesign. The baseline showed 61 clicks a month from 214 URLs. The crawl found 340 URLs, of which 126 were filter parameter pages reachable only from a JavaScript menu. The index comparison showed 44 product pages excluded by a canonical tag pointing at the category. The template sample found a single title tag repeated across all 92 product pages.

Only two items reached bucket one: remove the canonical from the product template and give the product template a unique title pattern. Both were template-level changes, shipped inside a week. The parameter pages moved to bucket three and were handled by a robots rule two months later. Nothing in the report required a designer, a copywriter, or a new CMS, which is why it was finished rather than filed.

Questions we get

Website audit FAQ

How long does a website audit take?

One working day for data collection and one for ranking, on a site under 500 pages. Larger catalogs need extra crawl time. The six-step order does not change with size; only the sampling depth does.

Do I need paid tools to do this?

No. Search Console, a free crawler and a spreadsheet cover all six steps. Paid tools speed up crawling and keep historical snapshots, which is convenience, not method.

Should I audit before or after a redesign?

Always before. A baseline recorded before the redesign is the only way to prove afterwards whether a traffic change came from the work or from the market. Audits taken after a redesign can only describe the new state.

What should be fixed first?

Whatever blocks indexation of pages that already earn impressions. Fixing the copy on a page Google cannot index produces no measurable result, and it makes the whole audit look ineffective to whoever approved the budget.

How many pages do I need to sample?

Five per template is usually enough to establish the pattern, because titles and internal links are generated by the template. If the five disagree with each other, the template is not the unit and you need a different grouping.

Does a website audit affect rankings directly?

Not on its own. The audit is a document. Rankings move when the fixes ship and are recrawled, which is why the ranked list and the deploy slot matter more than the report format.

How often should the audit be repeated?

A full pass each quarter, a light crawl monthly, and a fresh full audit after any migration, template change or large content release. Anything faster turns into monitoring, which is a different job.