When a business owner tells us their SEO is not working, the report in front of them usually says the opposite.
Rankings are up. Impressions are up. The agency is pleased. The phone has not rung any more often than it did in January.
Both things can be true at once, because a ranking is not an enquiry. Between the two sit four or five places where the value leaks out, and the leak is nearly always in one of them.
This article is the order we check those places in. Each check takes under ten minutes in Search Console or GA4, and each has a specific fix. Start at the top, because the first cause is the commonest and the last is the easiest to rule out.
The order to check things in
Check the queries first, then the positions, then the search results page, then the landing page, and only then the tracking. That order follows how often each cause turns out to be the culprit when we open up a client's data.
flowchart TD
A["Rankings up, enquiries flat"] --> B{"Are the ranking queries commercial?"}
B -- "No, mostly questions" --> B1["Cause 1: wrong queries. Build and link pages for buying terms."]
B -- "Yes" --> C{"Average position 1 to 3?"}
C -- "No, 4 to 10" --> C1["Cause 2: low CTR. Rewrite titles, push the money pages up."]
C -- "Yes" --> D{"Impressions up but clicks flat?"}
D -- "Yes" --> D1["Cause 3: AI Overview takes the click. Shift effort to local and commercial terms."]
D -- "No, clicks are up too" --> E{"Do visitors contact you?"}
E -- "No" --> E1["Cause 4: landing page. Phone above the fold, speed, trust."]
E -- "Can't tell" --> F["Cause 5: tracking. Send a test enquiry and watch GA4."]
The chart is a guide to likelihood, not a strict sequence. If you already know your form is broken, fix that first. Most people do not know, which is why it sits at the end.
Cause 1: you rank for questions, not for buying
The commonest reason rankings and enquiries come apart is that the site ranks for informational searches, and people asking questions are not ready to buy. "How much does a boiler service cost" and "boiler service Northampton" look similar in a rankings report. They bring very different visitors.
Blog-led SEO produces this pattern almost by design. Articles are easier to rank than service pages, so the article count grows, the rankings report fills with page-one positions, and the person searching is three months away from needing anyone.
How to check in Search Console
Open the Performance report, set the date range to the last three months, and look at the Queries tab. Google's own filtering guide explains the regex option: choose Custom (regex) and enter something like (cost|price|near me|quote|hire|buy|company|services) to see only the queries that carry buying intent.
Compare the clicks on that filtered list with the unfiltered total. If buying queries are under a fifth of clicks, the SEO has been growing the wrong audience.
The fix
Give the service pages their own targets and links. Each page that should take an enquiry needs a buying phrase in its title, a town or area where that matters, and links from the articles that currently take all the traffic. The articles then do a job: they pass readers along to the page that sells.
Cause 2: positions 4 to 10 get almost no clicks
Ranking on page one and ranking where people click are different achievements. Backlinko's analysis of 4 million search results, drawn from Search Console data, found the first organic result takes 27.6% of clicks and the top three together take 54.4%. Page two gets 0.63%.
First Page Sage's 2025 figures, as reported by Semrush, show the same shape with a steeper top: 39.8% for position one, 18.7% for two, 10.2% for three, 7.2% for four, 5.1% for five, then a slide from 4.4% to 1.6% across positions six to ten. The full ten-point breakdown is tabulated here.
xychart-beta
title "Organic click-through rate by Google position (First Page Sage, 2025)"
x-axis ["1", "2", "3", "4", "5", "6", "7", "8", "9", "10"]
y-axis "CTR (%)" 0 --> 45
bar [39.8, 18.7, 10.2, 7.2, 5.1, 4.4, 3.0, 2.1, 1.9, 1.6]
So a page that moves from position 14 to position 7 has, on paper, had a great quarter. In clicks it has gone from nothing to roughly 3 in every 100 searches. For a term searched 200 times a month in your area, that is six visits.
How to check in Search Console
In the same Performance report, switch on the Position metric and sort the Queries table by impressions. Google defines position as the average spot of your topmost result for that query. Look at your buying queries from the cause 1 filter. If most sit between 4 and 10, the rankings are real and the clicks are the problem.
The fix
Two jobs, in this order. First, rewrite the titles and descriptions of the pages that sit at 4 to 10 so they say what a buyer wants to see: the service, the place, something concrete such as "same-week appointments" or "fixed quotes". That can lift CTR without moving position.
Second, put the link-building and content effort behind the three or four pages that would pay for the whole programme if they reached the top three.
Cause 3: the AI Overview took the click
For informational searches, the answer now often sits above your result in an AI Overview, and the click never happens. Ahrefs compared 300,000 keywords across March 2024 and March 2025 using Search Console data. Where an AI Overview appeared, the average position-one click-through rate fell from 7.3% to 2.6%, which the authors put at 34.5% below what it would have been without the overview.
The same study notes that 99.2% of keywords triggering an AI Overview were informational. That matters for the diagnosis. If your rankings are mostly on question-type queries (cause 1), you are also the business most exposed to this.
Pew Research found the same from the user's side. In a study of 900 US adults in March 2025, people clicked a traditional result on 8% of searches that showed an AI summary, against 15% when there was none. Only 1% clicked a link inside the summary itself.
How to check in Search Console
Use the date comparison in the Performance report: this quarter against the same quarter last year. Look for queries where impressions rose and clicks stayed flat or fell, with position unchanged. That pattern, on informational queries, is the AI Overview absorbing the click. It is UK data, since Search Console reports your own searchers.
The fix
Stop paying for rankings on queries where the click no longer exists. Move the budget to the buying and local terms, where AI Overviews appear far less often and the searcher wants a supplier, not an explanation. Keep the informational pages, because being the cited source still has some value even in zero click search, but stop measuring the programme on them.
Cause 4: the page gets the visit and loses it
A page that ranks and gets clicked can still produce nothing, and this is where most owners never look. The visitor arrives, cannot see how to get in touch, waits for the page to load on a phone, finds no reason to trust the business, and leaves.
A thread on UK Business Forums from 2010 is still the clearest example we know. A security alarm company ranked and took 10 to 15 visitors a day for six weeks, and the owner wrote that "we still have not had a single phone call/lead". The first reply came from a copywriter who had looked at the site.
"Having seen the website, the answer is simple: there's no sales copy." directmarketingadvice, UK Business Forums, October 2010
The second reply asked whether the phone number was at the top of every page, and whether the page ever actually suggested the reader should call. Fifteen years on, we still open client sites where the answer to both is no.
Speed belongs here too. Google's web.dev team sets the Largest Contentful Paint threshold at 2.5 seconds for a good experience, measured at the 75th percentile of page loads, and calls anything over 4 seconds poor. A service page that takes five seconds on a 4G connection loses a share of its visitors before they see the number.
How to check in GA4
Open Reports, then Engagement, then Landing page. Find the pages that take your organic traffic and look at two things: average engagement time and the key events column. A service page with a 20-second engagement time and no key events is losing people. Then load the page on your own phone, on mobile data, and time it.
The fix
Put the phone number and a short form in the first screen on mobile, with no scrolling. For clinics and dental practices, that means the booking link sits above the fold, not a form buried three screens down. State the three or four facts a buyer checks before calling: where you cover, what it costs or how quoting works, how soon you can come, and who you are. Add the proof you already have, such as review counts, accreditations and photos of real jobs.
Then fix the speed, which on most small business sites means image sizes and one or two plugins.
Where the page is beyond patching, this becomes a website design job, and no amount of SEO will patch it. The two are often sold separately and should not be.
Cause 5: the enquiries happened and nobody recorded them
Sometimes the SEO has been working for months and the reporting has not. Form submissions go to an inbox nobody checks, the thank-you page was renamed in a site update, or the key event in GA4 was never set up. The rankings report says "up", the enquiries report says "zero", and the truth is somewhere in a spam folder.
This is last in the tree because it is the quickest to rule out, not because it is rare. We check it on every site we audit, before we read a single ranking, because it changes the meaning of every other number.
How to check in GA4
Submit a test enquiry through every form on the site, and phone the tracked number if there is one. Then open GA4, go to Reports, then Realtime, and watch the "Key events by event name" card.
Google's verification guide also describes DebugView, which shows each event as it arrives with its parameters, for cases where Realtime is not detailed enough.
If the test enquiry does not appear within a minute, the SEO report has been counting a number that was never being measured. Check the inbox the form sends to as well, and the spam folder on that inbox.
The fix
Make form submission, phone clicks and tracked calls into key events in GA4, and check them monthly by sending a test. Ask whoever reports on SEO to show enquiries, not sessions, as the first line of the report. If they cannot, that is its own finding.
The ten-minute version
If you only have ten minutes before a meeting with your agency, do these in order.
- Search Console, Performance, Queries, regex filter for buying words. What share of clicks are buying queries?
- Same report, sort buying queries by impressions, read the Position column. How many sit at 4 to 10?
- Same report, compare this quarter to last year. Where did impressions rise while clicks did not?
- GA4, Landing page report. Which organic landing pages have engagement time under 30 seconds and no key events?
- Submit a test enquiry and watch GA4 Realtime. Did it arrive?
| Cause | Where you see it | What it looks like | First fix |
|---|---|---|---|
| Wrong queries | Search Console, Queries | Buying terms under a fifth of clicks | Target and link the service pages |
| Positions 4 to 10 | Search Console, Position column | Buying terms ranking but 1% to 7% CTR | Rewrite titles; push the money pages |
| AI Overviews | Search Console, year-on-year compare | Impressions up, clicks flat, position steady | Move budget to commercial and local terms |
| Landing page | GA4, Landing page report | Short engagement, no key events | Phone above the fold, proof, speed |
| Tracking | GA4 Realtime after a test | Test enquiry never appears | Set up key events, test monthly |
Most of the time, the answer comes out of steps one and two, and the conversation with the agency changes from "why is SEO not working" to "which pages are we going to move into the top three, and how long that should take". That is a question an agency can answer and be held to.
The rule we would offer is this: judge SEO on the number of buying queries in positions 1 to 3, and on the enquiries that follow them. Everything else in the report is a lagging or a vanity indicator, depending on how it is used.
What to expect from an agency more broadly is set out in our guide to what a UK SEO agency should cost and deliver. If you would like us to run this check on your own Search Console and GA4 and tell you which of the five it is, our SEO audit does exactly that, or you can get in touch with the question.