A Technical SEO Audit You Can Run in an Afternoon

You can do the useful 80% of a technical SEO audit yourself, in one afternoon, with Search Console, PageSpeed Insights and your browser. Three hours, no licence fees, no retainer. The hard part is not finding problems — any crawler will hand you four hundred warnings — it is knowing which numbers are genuinely bad. So this runs in three one-hour blocks, and each one comes with the benchmark figures from the 2025 Web Almanac so you can tell a real fault from a cosmetic one.

The afternoon plan

Block out three hours and work in this order, because each block depends on the one before it. There is no point tuning images on a page Google cannot index.

1. Hour one — reach and indexing. Can Googlebot fetch your pages, and has it chosen to index them? Search Console’s Pages report and the URL Inspection tool answer both.
2. Hour two — speed and stability. What do real visitors experience? Search Console’s Core Web Vitals report and PageSpeed Insights, field data only.
3. Hour three — on-page signals. Titles, headings, alt text, structured data and internal links on your fifteen most important URLs.

Write findings in a single document with three columns: what is wrong, what it costs you, and who can fix it. If the third column is empty, the item is not a finding, it is a complaint.

Hour one: can Google reach and index your pages

Start in Search Console, Indexing, Pages. The chart splits your URLs into indexed and not indexed, and the reasons list underneath is where the real information lives. “Crawled, currently not indexed” usually means Google read the page and judged it not worth storing. “Discovered, currently not indexed” often means a crawl budget or server speed problem. “Excluded by noindex tag” means somebody made a decision, deliberately or not.

Then check your sitemap. Google’s documentation says it supports the formats defined by the sitemaps protocol and has no preference between them, so the format matters far less than the contents. A sitemap listing redirected, noindexed or 404 URLs is actively unhelpful. On WordPress the sitemap is usually generated by your SEO plugin, so the fix is a plugin setting rather than a file edit.

Finish the hour with URL Inspection on five pages: your home page, your most important service page, your best blog post, a page you suspect is broken, and one page you have never thought about. Use “View crawled page” to see the HTML Googlebot actually received. If your main content is missing from that HTML, you have found the most important fault on the site.

robots.txt does not mean hidden

This one causes real damage, so it is worth being precise. Google’s own introduction to robots.txt says the file tells crawlers which URLs they can access, that it is used mainly to avoid overloading a site with requests, and that it is not a mechanism for keeping a web page out of Google. To keep a page out, you block indexing with noindex or you password-protect it.

The failure mode is specific. A URL disallowed in robots.txt can still be indexed if something links to it, because Google never fetches the page and therefore never sees your noindex tag. You end up with a URL in the index and no snippet. If you want a page gone, let Google crawl it and tell it to go away on the page itself.

Canonicals, duplicates and parameter sprawl

Canonical tags are the most over-used tool in the box. Google’s canonicalization documentation describes the process as selecting the representative URL for a piece of content, and the help page on specifying a canonical is clear that none of the methods are required — if you do not specify one, Google identifies the version it considers best for users.

So do not spend your hour on canonicals. Spend it on the three duplicates that actually exist on most small sites: the same page reachable with and without a trailing slash, the same page on http and https, and filtered or sorted listing URLs with query parameters generating endless near-identical pages. The first two are redirects. The third is usually a theme or plugin setting.

Use URL Inspection to see which URL Google picked as canonical. If Google’s choice and yours differ, Google is telling you something about how the pages are linked internally. That is more useful than overriding it.

Hour two: Core Web Vitals, measured on real visits

Core Web Vitals are three field metrics, and the thresholds are fixed. web.dev’s explanation of how they were defined sets the pass bar at the 75th percentile: if at least 75% of page views to a site meet the good threshold, the site is classified as good for that metric. Good means Largest Contentful Paint within 2.5 seconds, Interaction to Next Paint within 200 milliseconds, and Cumulative Layout Shift of 0.1 or less.

The 75th percentile matters more than the numbers. It means your average visitor is irrelevant; the slowest quarter decides. A site that feels instant on your office fibre and your two-year-old laptop can still fail, because the measurement includes the visitor on a mid-range Android phone on a train.

In PageSpeed Insights, read the field data at the top and treat the lab score below it as a diagnostic hint only. A green lab score of 98 with failing field data is common and means your real visitors are slower than your test. Only the field data affects Search.

Two things jump out of that table. Interactivity is almost solved on desktop and clearly not on mobile, a 20-point gap caused by slower phone processors chewing through JavaScript. And layout stability is the one metric where mobile beats desktop, at 81% against 72%, largely because desktop pages have more room for late-loading sidebars and ad slots to shove content around.

Where the web actually sits on each metric

Context stops you panicking. The Web Almanac’s Performance chapter, drawing on CrUX data from July 2025, found that 48% of mobile sites and 56% of desktop sites achieved good scores across all three Core Web Vitals. In 2024 those figures were 44% and 55%, and back in 2021, the first year of the metrics, they were 32% and 41%.

Popularity does not help. The same chapter reports that among the top 1,000 websites, 59% of desktop sites have good Core Web Vitals against 51% on mobile, and mobile gets worse further down the ranks — 42% for the top 10,000 and 37% for the top 100,000. Big brands ship heavy pages. You are not competing against a perfect web.

One very specific finding is worth acting on today: the Almanac found about 16% of pages still lazy-load their LCP image, which delays the one piece of content the metric is measuring. If your hero image has loading="lazy" on it, remove that attribute and you may fix LCP in a single edit.

Hour three: titles, headings, alt text and markup

The last hour is the cheapest and most often skipped. Take your fifteen most important URLs — home page, service pages, top blog posts — and check five things on each: a unique title under about 60 characters that leads with the thing people search for, a meta description that reads like a reason to click, exactly one non-empty H1, a sensible H2 outline, and alt text on every image that carries information.

The benchmark here is embarrassing in your favour. The 2025 Web Almanac’s SEO chapter found an H1 on 70% of pages but only 66% non-empty, and at the median just 60% of images on mobile pages had alt attributes, up from 58% in 2024. Getting these right on fifteen pages takes an hour and puts you ahead of a large slice of the web.

Structured data goes in the same hour. JSON-LD appeared on 43% of home pages in the 2025 crawl and only 39% of inner pages on desktop, so adding Organization and Article markup is still differentiating. Keep it truthful: markup that does not match the visible page is worse than no markup.

• Titles — unique, front-loaded, written for a human scanning a results page.
• One H1 — with words in it. On most content systems this is the post or page title, so check your theme is not adding a second.
• Alt text — describe the information, not the file. “Rack of network switches with labelled patch cables” beats “IMG_4821”.
• Internal links — every important page should be reachable in three clicks from the home page, with anchor text that says where it goes.
• Structured data — Organization once, Article on posts, Breadcrumbs sitewide. Nothing you cannot see on the page.

What to fix first, and what to leave alone

You will finish the afternoon with a list. Fix in this order: anything stopping indexing, then anything a visitor experiences, then anything a crawler merely notices. An accidental noindex outranks a slow image. A slow image outranks a missing canonical.

And here is the part no audit tool will tell you. Leave alone: keyword density, the exact length of your meta descriptions, the handful of 404s from links nobody follows, and every warning about a third-party script you cannot change. Technical SEO has a point of diminishing returns and it arrives sooner than the industry admits.

Once the technical floor is sound, the gains come from content and from choosing where to spend. We cover the content side in content marketing that actually works, the groundwork in SEO fundamentals every business should know, and the budget question in SEO vs paid ads.

If your afternoon ends with findings that need developer time, we pick up from there remotely — indexing faults, Core Web Vitals work and structured data. Our digital marketing services include a written audit you keep, not a dashboard you rent. Get in touch with Eudora Technology to talk about your project.