A site I looked at last year had noindex sitting on its 3 highest-value pages. Not a penalty, not an algorithm update. A tag left over from staging that nobody removed at launch. It had been there 9 months, and the business had been paying for content the whole time.
So I run the same 30-minute pass before I’ll say anything about a site’s search performance. It isn’t a full audit, and it doesn’t replace one. It’s the first pass: 7 checks that find the expensive problems fastest, in an order that stops you optimising heading tags on pages Google has never indexed. You don’t need to be an SEO to do any of it, and every tool here is free: Search Console, PageSpeed Insights and a crawler. Screaming Frog’s free version does 500 URLs, which covers most small business sites.
The short version
- One leftover tag can bury your best pages. Your inbox knows what buyers really type.
- 4 crawl faults quietly knock pages out of Google. The other warnings can wait.
- The most telling check needs no tools, just 90 seconds in a buyer’s seat.
- Then rank what you found. The fix that pays most is also the toughest.
1. Is it indexed? (4 minutes)
Search Console, Indexing, Page indexing. You want the ratio of indexed to not-indexed pages, then the reasons underneath.
Why: an unindexed page can’t rank, so every check after this one is theoretical until you’ve done it.
A bad answer looks like more URLs in “Not indexed” than “Indexed” on a site under 500 pages. Read the reasons rather than the total. “Excluded by ‘noindex’ tag” on a commercial page is the staging-tag problem above. “Blocked by robots.txt” combined with noindex is worse than either alone: Google can’t read a tag on a page it isn’t allowed to fetch, so the noindex never gets processed. “Discovered - currently not indexed” in volume usually means thin pages, not a crawl budget crisis.
This report lags, sometimes by weeks. For anything you actually care about, run URL Inspection on the page instead. That’s live.
2. What’s ranking, and for what (3 minutes)
Performance report, last 3 months. Queries tab, then Pages tab.
Why: this is the only place that tells you what the site earns attention for, rather than what you’d like it to earn attention for.
A bad answer is your own business name taking most of the clicks. That’s not search performance, it’s people who already knew you. The other bad answer is a mismatch between the Pages tab and your revenue: a 2019 blog post carrying the site while the page that sells something gets 40 impressions a month.
If your property has Google’s AI Overviews and AI Mode report, read it as visibility, not value. It tells you whether you’re in there, not whether being there is worth anything. And if clicks are falling while enquiries hold, read why that’s usually good news before anyone tries to fix it.
3. What your buyers actually type (8 minutes)
Step 2 tells you what the site is found for. This one tells you what it should be found for, in the words people actually use.
Stay in the Queries tab and read the long ones, the 5- and 6-word queries. That’s people describing their problem in their own words, and it’s rarely how the site describes it. Then type your main service into Google and read what autocomplete finishes it with and what sits under “People also ask”. Then search Reddit and the forums your customers use for the same problem. Then your own inbox: the last 10 enquiries, reviews and sales emails. That’s the best source on this list, and the only one nobody else can see.
Why: most sites are written in the business’s words. The business says “hot water system replacement”. The customer types “no hot water”. The page that ranks is the one that uses the customer’s words.
A bad answer is a question that keeps coming up with no page on the site that answers it. Or a page that answers it in words nobody searches for.
4. The 4 technical failures that matter (5 minutes)
Crawl the site. Then look at exactly 4 things.
- Directives. Any
noindexornofollowon a page you want ranked. Bad answer: any at all on a commercial page. - Canonicals. Bad answer: a canonical pointing at a staging domain, at the homepage from every page or at a URL that 404s. Self-referencing canonicals are fine and boring, which is what you want.
- Redirects. Bad answer: any chain longer than 2 hops, and any chain that terminates in a 404 or a redirect loop. These usually appear after a site migration nobody cleaned up.
- Internal 404s. Bad answer: broken links in the main navigation or from a page that ranks.
Why these 4 first: each one is a hard failure that silently removes a page from consideration. The rest of a crawl report matters, but it’s work for the full audit, not the first 30 minutes.
5. Core Web Vitals, field data only (3 minutes)
PageSpeed Insights. Look at the top section: real Chrome user data from CrUX. Ignore the Lighthouse score below it, which is a lab simulation on a synthetic connection.
The 3 metrics and their thresholds: LCP under 2.5 seconds, INP under 200 milliseconds, CLS under 0.1. You pass when 75% of real page views hit the threshold, measured over a rolling 28 days.
Why field data: I’ve seen a lab score of 96 on a page that was failing LCP for actual visitors. The lab test doesn’t have your customers’ phones or their reception.
A bad answer is a failing metric at origin level, which means it’s site-wide rather than one page. “Not enough data” at URL level isn’t a failure: the page has too little Chrome traffic and PSI has fallen back to origin data. Search Console’s Core Web Vitals report groups failures across the site for you.
6. Titles and headings on the pages that earn money (4 minutes)
Not every page. The 5 or 6 that make money.
Why: the title tag is the single on-page element that changes whether someone clicks, and duplicate titles tell Google 2 pages are the same page.
Bad answers, in order of how often I find them: the business name in front of every title, so the first 20 characters are identical site-wide. 5 pages sharing one title. No H1. An H1 that’s the logo. H2s that describe the company and never name the thing being sold. Titles in the business’s words when step 3 found the customer’s.
7. Internal links, and the question a buyer is asking (3 minutes)
In the crawler, sort by inlinks.
Why: internal links are how importance gets distributed, and they’re the one ranking factor you fully control.
A bad answer is a money page with 2 internal links pointing at it while a blog post has 40. Fix the direction, not the count.
Then close the tools and read your main service page as a buyer. Does it say who it’s for, roughly what it costs and what happens next? A bad answer is 3 scrolls of company history before anything a buyer asked about. This check takes 90 seconds and finds more revenue than the other 6 combined, which is an uncomfortable thing for an SEO audit to conclude.
Taking it further
Step 3 is a sample. It tells you whether there’s a gap between how you describe what you sell and how people ask for it. It doesn’t tell you how big the gap is.
That’s what the paid tools are for. Social listening tools like Brandwatch and Sprout Social track what people say about a problem, a category or a competitor across social platforms, forums and reviews, over months rather than one afternoon. SparkToro tells you where your audience spends its time and who it listens to. Keyword research tools like Semrush and Ahrefs put a monthly search volume against each phrase, so you know which questions are worth a page and which are three people and a bot. AnswerThePublic turns a single topic into the questions people ask about it. Together they give you the full list of what your buyers are asking, ranked by how many of them are asking it.
It’s the first thing I add when the quick check turns up a gap. It’s also where the work stops being 30 minutes.
What to do with the findings
Sort them by cost to fix against cost of staying broken. Indexation and technical failures are cheap and urgent. Usually one person, one afternoon. Titles are cheap. A question with no page is one page. Core Web Vitals are slow, expensive and rarely the actual problem. Internal linking is free and always neglected. Copy is the hardest and the most valuable.
Then stop. Thirty minutes finds the room the problem is in. It doesn’t tell you why it’s there, and the “why” is what decides between an afternoon and a rebuild. That’s the full audit: competitors, backlinks and content at scale.
A fast audit isn’t a diagnosis. It’s the thing that stops you spending a quarter fixing the wrong part of the site.