How to Run an SEO Audit on Your Own Site in Under an Hour

Most SEO audits fail for the same reason: they start in the wrong place. Somebody opens a tool, gets a list of 340 issues sorted by nothing in particular, fixes the ones that are easiest to understand, and changes nothing that mattered.

An audit is useful when it runs in dependency order. There is no point optimising a title tag on a page Google cannot reach, and no point worrying about page speed on a page that is not indexed. Work in the right sequence and an hour is genuinely enough for a small site.

Step 1: Can Google reach the page at all? (10 minutes)

Everything else depends on this, and it is the step people skip because they assume the answer is yes.

• robots.txt. One careless Disallow can remove a whole section from search. Check the specific paths that matter rather than glancing at the file — the Robots.txt Checker tests a URL against the rules the way a crawler does.
• noindex tags. Staging sites ship with noindex and it survives the move to production more often than anyone admits. If an important page vanished after a redesign, check this first.
• HTTPS. An expired certificate stops crawling outright, not just browsing. The SSL Checker shows the expiry date in a few seconds, and we wrote about why HTTPS matters for SEO separately.
• Sitemap. Submitted, fetching successfully, and listing the pages you care about.

If anything here fails, stop and fix it. Nothing further on this list will help.

Step 2: Is it actually indexed? (5 minutes)

Reachable and indexed are different things, and only Search Console answers the second honestly.

Open the Pages report and look at the reasons under “Not indexed”. The two that matter most are Discovered — currently not indexed, which means Google knows about the URL but has not prioritised crawling it, and Crawled — currently not indexed, which means it looked and was not convinced.

The first is usually an authority and internal-linking problem. The second is a content problem. They look similar in the interface and need completely different responses, which is why reading the specific reason matters more than the total count.

Step 3: Does the page say what it is about? (20 minutes)

This is where most of the actual wins are on a small site, and where the analyzer earns its keep.

• One <h1> per page, describing the page rather than the brand.
• A title tag under about 60 characters that front-loads the thing people search for. We went into this properly in the title tag guide.
• A meta description that earns the click. It is not a ranking factor, but it decides whether anyone chooses your result over the one above it.
• Headings in a sensible order. H2s for sections, H3s underneath them. Not chosen for how big the text looks.
• Image alt text that describes the image. Empty alt on a decorative image is correct; missing alt on a meaningful one is not.
• Internal links pointing at the page. A page nobody links to looks unimportant, because by your own site’s structure it is.

Step 4: Technical signals worth checking (15 minutes)

Shorter than people expect, because most technical SEO advice is aimed at sites far larger than yours.

• Mobile. Google indexes the mobile version. If your mobile layout hides content that exists on desktop, the hidden content is what gets ignored.
• Core Web Vitals, but only the field data in Search Console. Lab scores from a fast laptop on fibre tell you very little about real visitors.
• Structured data. Worth it for the page types that earn rich results — articles, products, FAQs, recipes. Not worth it everywhere. Check what you have with the Schema Checker.
• Canonical tags. Self-referencing on the canonical version, pointing at it from the duplicates. Getting this backwards quietly de-indexes the page you wanted.
• Duplicate titles and descriptions across pages, which is a strong hint that two pages are competing for the same query. We covered fixing duplicate meta descriptions in more detail.

Step 5: Decide what to fix first

You now have a list. The temptation is to work top to bottom. Do not — sort by impact instead.

A short rule that holds up well: five pages fixed properly beats fifty pages touched lightly. Find the handful of URLs that already get impressions in Search Console and work on those, because they are the ones Google has already decided are nearly good enough.

Reading the Search Console data that actually guides the work

Most of the decisions above come from one report, and most people look at the wrong column in it.

Open Performance, set the range to the last three months, and switch on all four metrics. Then read them in this order:

• Impressions without clicks. Pages ranking somewhere nobody looks, or ranking fine with an unappealing title. This is the cheapest win on the whole list — Google already thinks the page is relevant, so rewriting the title and description can change traffic within days, with no new content at all.
• Average position between 8 and 20. These are the pages close enough to page one that a modest improvement moves them. A page at position 60 needs a different kind of work entirely; a page at 11 might need one good internal link.
• Queries you did not expect. Sort by impressions and read down. You will usually find two or three terms you never targeted that the page happens to rank for. Those are free briefs for your next piece of content, and they tell you how your audience actually phrases things.
• Pages with falling impressions. A slow decline over months usually means the content has aged rather than anything technical breaking.

What not to read into it: day-to-day position wobble. Average position is averaged across every query, device and country, so it moves constantly for reasons that have nothing to do with you. Compare months, not days.

A worked example

Say your Performance report shows a page with 1,400 impressions, 11 clicks and an average position of 9. That one row tells you a lot.

Position 9 means you are at the bottom of page one — Google already considers the page a reasonable answer, so this is not a content quality emergency. But eleven clicks from fourteen hundred impressions is well under one percent, which is low for that position. The ranking is not the problem; the result itself is not attractive enough to click.

So the first move is not more content, more links, or faster loading. It is rewriting the title and meta description, then waiting a fortnight and comparing. If the click rate improves, you have gained traffic for twenty minutes of work. If it does not, you have cheaply ruled out the easy explanation and can look at the page itself.

Compare that with a page showing 40 impressions and position 48. Nothing about the title will fix that one — it needs either substantially better content or internal links from pages that already have authority. Same report, same two numbers, completely different response.

What to ignore

Plenty of audit findings are noise, and acting on all of them is how an afternoon becomes a fortnight.

• Keyword density. Not a thing search engines measure in any way worth optimising for.
• Meta keywords. Ignored by Google for roughly two decades.
• A perfect 100 score in any tool, including ours. These are heuristics that approximate good practice, not the ranking algorithm.
• Tiny speed gains on a site that already loads in two seconds. The gap between 2.0 and 1.8 seconds is not what is holding you back.
• Every broken external link ever. Fix them in navigation and important content; stop short of auditing comments from 2019.

What you need, and what it costs

Nothing on this list requires a subscription, which is worth saying because most audit advice assumes a tool that bills monthly.

Google Search Console is free, and it is the only source that tells you what Google actually did with your pages rather than what a third party guessed. If you set up nothing else, set this up. Verification takes a few minutes and the data starts accumulating from that day, which is a good reason not to put it off.

Free checkers cover the mechanical parts — robots rules, certificate expiry, structured data, on-page tags. They answer the yes-or-no questions quickly so you can spend your hour on judgement rather than lookups.

Paid crawlers earn their money on sites of a few thousand pages and up, where finding every orphaned page by hand is genuinely impractical. Below that they mostly produce longer lists of smaller problems, which is the opposite of what a first audit needs.

The constraint on a small site is almost never missing data. It is deciding which of the things you already know about to fix first, and no tool does that part for you.

How often to repeat this

Quarterly for a stable small site, and immediately after any redesign, platform migration or host change — those are when noindex tags and broken robots rules escape into production. Keep a short note of what you changed and when, so that when something moves three months later you have some idea which change to look at.

If your audit turns up weak or missing titles and descriptions across a lot of pages at once, writing them by hand is miserable. The Meta Tag Generator will draft them in bulk from a list, and there is a walkthrough in writing meta tags in bulk.