If your website is indexed but not ranking, the usual reason is that you have no page dedicated to the search you care about, or the page exists but someone else answers the question better or has more authority. Which of the two it is shows up in the average position in Google Search Console: a page stuck beyond 50, one stuck between 11 and 20 and one stuck between 4 and 10 have three different causes and need three different pieces of work.
The trouble is that hardly anyone looks at that number.
People whose site won't climb usually do one of three things:
- buy technical fixes (plugins, speed, resubmitting the sitemap) for a site Google already reads perfectly well;
- try to rank the homepage for service searches, where in our test it almost never appears;
- write more blog posts, while the page that's meant to sell stays where it is.
And why should you trust the way we diagnose it?
Because we've watched it happen, with the numbers to show for it. Sunscape, which develops solar farms and looks for land all over Italy, was getting 329 impressions and 46 clicks a month in March 2026, almost all of them on its own name. The site was live and indexed. It didn't rank because no page answered people searching "vendo terreno per fotovoltaico" (selling land for solar).
Six months later, with one page for each intent and each region, the tracked keywords on page one went from 1 to 38 out of 143.
Macropix, an Italian maker of LED walls, has the opposite problem: 79 keywords out of 135 on page one, but only 28 in the top three. There the cause isn't a missing page, and the fix is different.
In this article:
- the table we use to read average position and work back to the cause;
- a test on 8 Italian SERPs, run on 9 October 2026, to see what kind of pages really sit above you, with a check on Google UK;
- six causes in order of frequency, each with its fix;
- what is almost never the reason, so you don't pay for it.
Let's start.
First question: can Google see you? (thirty seconds to rule it out)
A site that doesn't rank and a site that isn't indexed look the same from the outside. They aren't: in the first case Google has the page on file and puts it at the bottom, in the second it doesn't have it at all.
To tell them apart, run the URL Inspection tool in Search Console on the page you care about. If it says the URL is on Google, you're in the right place and can keep reading. If not, the path is a different one and we've written it up in why my website isn't showing on Google, going through the Search Console labels one by one.
From here on we assume the page is in the index. The problem isn't getting Google to find you, it's persuading it to put you above everyone else.
Where your site has stalled (the table we use for the diagnosis)
Average position is the number Search Console shows in the Performance report: for each impression it records the topmost position held by a link to your site and averages them. Filtered by query and by page, it tells you where Google keeps you for each search.
Why does the band matter so much? Because clicks aren't spread evenly. Backlinko's study of around 4 million results, based on Search Console data from several accounts, found that the first result has a CTR of 27.6%, ten times that of the tenth, and that only 0.63% of searchers click anything on page two. The data is from 2019 and the study doesn't say which market it covers: treat it as an order of magnitude, not a UK figure.
Put simply, position 12 and position 60 bring in almost the same traffic, which is to say none. But they say two different things about the page.
| Average position for the query | What Google is telling you | Most likely cause | First fix |
|---|---|---|---|
| No impressions | It has no page of yours to show for that search | The page doesn't exist, or it's about something else | Create a page dedicated to that search |
| Beyond 50 | It sees you as only loosely relevant | Wrong intent or format compared with the SERP | Rewrite the page in the type of result Google rewards |
| 21-50 | The page is in the running but weaker than the others | Thinner content, few internal links, two pages on the same search | Merge duplicate pages, add internal links and substance |
| 11-20 | You're almost on page one | On-page details and internal links, sometimes cannibalisation | Title, missing sections, links from the strongest pages on the site |
| 4-10 | You're on page one but not at the top | Authority of the site and page compared with competitors | Links from sites in your sector, signs of real experience |
| 1-3 with few clicks | You're near the top, but something else sits above you | AI Overviews, maps, directories, video | Work on the SERP features, not on the position |
One caveat almost nobody mentions: the average for a query gets muddied when several of your pages appear for the same search. Before reading it, filter by single page. How to do that step by step is in our guide to Google Search Console.
Who sits above you: our test on 8 Italian SERPs (and what doesn't matter)
To see what separates the pages at the top from those on page two, we ran a test. On 9 October 2026 we used DataForSEO to pull the top 20 organic Google results on mobile, geolocated to Italy, for 8 commercial searches in sectors we work in. For every page in positions 1-3 and 11-20 we counted the domains linking to it, both to the page and to the whole site, in DataForSEO's backlink index. It's a test on the Italian market: repeating it on the UK would mean around a hundred backlink lookups, so here we report it as it is, and further down we compare the main examples with Google UK.
| Metric | Positions 1-3 | Positions 11-20 |
|---|---|---|
| Results analysed | 24 | 80 |
| Pages with no linking domain | 79.2% | 83.8% |
| Domains linking to the page (median) | 0 | 0 |
| Domains linking to the whole site (median) | 747 | 663.5 |
| Homepages among the results | 0 out of 24 | 1 out of 80 |
Three things stand out.
On niche Italian searches, links to the individual page don't separate the top three from page two. In both groups nearly eight pages out of ten have no links detected. Mind the limit: "zero" means zero in DataForSEO's index, which is incomplete for small sites, not zero in absolute terms.
Domain authority doesn't change much either: a median of 747 referring domains at the top, 663.5 on page two. Page two includes YouTube, Facebook, Instagram and Reddit, huge sites stuck between positions 12 and 20.
Homepages aren't there. Out of 160 positions between 1 and 20 we found just one, at position 11. Everything else is an internal page built around the search. The UK checks we describe below add one nuance: for "3d printing service", two of the top three were the homepages of small firms built around 3D printing. When the whole business is one service, the homepage is the service page. When you sell ten things, it can only rank for your name.
So backlinks don't matter?
They do, but further down the line. On these searches the first cut is made by relevance: having the right page, written for that question. Links carry weight when lots of sites have the right page, which is what we see in the 4-10 band.
Cause 1: you have no page for that search (the Sunscape case)
It's the most common cause and the hardest to spot, because it produces no error in Search Console. The query simply doesn't appear, or appears with a handful of impressions and a position beyond 70.
In March 2026 the Sunscape site was technically sound. It had a general page about the company and its projects. But landowners search with the name of their region and the verb they know: sell, lease, agrivoltaics. One page couldn't answer all three questions across twenty regions.
We built the architecture around two intents, selling and leasing, with dedicated regional pages. The result, from the case study:
- organic impressions from 329 in March to 5,809 in August 2026;
- monthly clicks from 46 to 195;
- keywords in the top hundred from 46 to 112 out of 143, over two months of tracking;
- "vendo terreno per fotovoltaico" (selling land for solar) at an average position of 6.6.
All without link building, which only starts in the 2026-2027 plan. Pages that didn't exist before began ranking as soon as they existed.
The test on the 8 SERPs says the same thing from another angle: for "lavorazioni cnc conto terzi" (subcontract CNC machining) two of the top three results on Google Italy were pages dedicated to that one service, with the service name in the address. Google UK looks the same: for "cnc machining services" on 10 October 2026 the top three organic results were a CNC milling page on Xometry, a premium machining service page and a page called "CNC Machining Tamworth". It isn't the keyword-in-the-URL trick. It's a sign that those companies have a page about that service, while their competitors have a "Services" page with ten services on it.
Choosing which searches deserve their own page is a job for keyword clustering: searches with the same SERP go on the same page, searches with different SERPs need different pages.
Cause 2: the page exists but answers a different question (intent and format)
Search intent is what the person wants to get done when they type the query: understand, compare, buy, find a supplier nearby. If your page offers something else, Google keeps it beyond position 50 even if it's well written.
The test has a clean example. For "stampa 3d conto terzi" (3D printing service) you'd expect the bureaus that do the printing at the top. On Google Italy the top three were a Protolabs white paper on printing technologies, an article explaining what 3D printing means and an event page. On that query, on that day, Google was rewarding content that explains. A service page with a quote form and three lines of text struggles there, and links have nothing to do with it. Run the equivalent search in the UK and the picture flips. For "3d printing service" on Google UK, mobile, on 10 October 2026, the top three organic results were all providers (JC3DDesign, 3DPRINTUK and Print My Part), under a local pack, with an All3DP roundup of the best UK services in fourth place. Same technology, different intent: it's the reason you can't borrow the diagnosis from another market, or even from another wording.
Hire searches are the clearest case in both countries. In Italy, six of the top ten for "noleggio carrelli elevatori" (forklift hire) were hire firms' pages with the town in the address, plus the Italian Yellow Pages in ninth. On Google UK, for "forklift hire", all ten organic results had a town or county in the title, from "Forklift Hire Manchester" to Yell's listing for London. A blog post explaining how a forklift works has no room there.
How can you tell without a tool?
Search your query on a smartphone and count: how many service pages, how many articles, how many directories, how many videos in the top ten. If your format is in the minority, change the format or change the search. We've explained it with more examples in our guide to search intent.
Paying for technical fixes on a page that's about something else?
You've just seen that a page written for the wrong intent stays beyond position 50 whatever you add to it. We open your Search Console together, put every page in its band and tell you which ones need rewriting, which should be merged and which are missing.
Book a strategy consultationCause 3: two pages compete for the same search (and Google usually shows two at most)
Cannibalisation is when two or more pages on the same site target the same query and Google alternates between them without giving strength to either. In Search Console it looks like this: same query, two URLs, positions swapping week by week between 15 and 40.
There's a written rule that makes it more expensive. In its guide to ranking systems Google explains that its site diversity system generally stops it showing more than two pages from the same site in the top results, unless they're especially relevant.
In our Italian test you could see it with the naked eye. For "noleggio carrelli elevatori" a hire portal had two city pages on page one, at positions 4 and 8. Another six city pages from the same site were lined up from 12 to 17, built on the same address pattern with only the town changing. Those six won't climb by working on the single page: the ceiling is the site. "Generally" is the key word, though. On Google UK, for "forklift hire" on 10 October 2026, one hire firm had three city pages in the top six organic results (Manchester, Chelmsford and Sheffield), and Yell had two more. The limit is a tendency, not a guarantee, and it tends to bend on searches where few sites have a dedicated page.
A warning so you don't overcorrect. Ahrefs analysed 9,700 cases where its own site ranked with several pages for the same keyword and, after checking 80 by hand, found just one that needed fixing. Two pages on the same query aren't a problem as long as they answer different intents. They become one when they're duplicates: near-identical product pages, copied city pages, duplicated language versions. On Macropix the technical audit found exactly that: duplicates created by WPML, the WordPress multilingual plugin.
If the duplicate pages are yours, the answer is to merge them or make them different, as we explain in the guide to duplicate content. If instead the same business runs two sites that steal the same search from each other, that's a different problem and we cover it in why multiple domains hurt SEO.
Cause 4: you're on page one but not in the top three (this is where links count)
When you have the right page and Google keeps it between 4 and 10, the most common reason is that the competitors above you have the same relevance and more authority: more sites citing them, more years of history, a name people search for.
Sunscape is exactly here. In September 2026 it had 38 keywords on page one and none in the top three. That's why the 2026-2027 plan includes link building on renewable energy publications: the pages are there, now they need to be cited.
Macropix, after years of work, has 28 in the top three out of 135 tracked. For "ledwall" (LED wall) its average position over the last twelve months is 5.5 with 1,495 organic clicks, more than three times the 452 brought in by the company name.
Age matters more than people think. In 2025 Patrick Stox of Ahrefs analysed the top ten results for 1.3 million US keywords: 72.9% of pages in the top 10 are more than three years old, and the page in first place is five years old on average. It's an American sample, but it describes a mechanism that doesn't depend on the country: older pages have had more time to collect links and confirmation.
What to do in practice is in our guide to backlinks. Just one recommendation here: the links that move a page from 6 to 2 come from sites in your sector, not from directories bought in bulk.
Cause 5: the site is young (and a year isn't long)
A new site that doesn't rank in its first months is statistically normal. The same Ahrefs study followed 1 million URLs published in September 2023: only 1.74% reached the top ten within a year. Of those that made it, 40.82% did so in the first month, so pages that don't climb straight away tend to climb slowly.
Sunscape went from 1 to 38 keywords on page one in six months, in a small niche with little competition. On contested searches it takes longer. How long to really wait, sector by sector, is in how long SEO takes to show results, and what to do in the first months is in our guide to SEO for new websites.
One special case: if the site used to rank and has stopped, it isn't a question of age. You start from the dates, as we explain in the guide to a drop in website traffic.
Cause 6: you've climbed but the clicks don't come (AI Overviews, maps, directories)
Sometimes the site ranks and the owner doesn't notice, because the position brings no clicks. It happens when something else sits above the first organic result.
Of the 8 Italian SERPs in the test, 4 had an AI Overview and 3 a local pack with the map. Among the organic results were Trovaprezzi (an Italian price comparison site) on "ledwall", the Italian Yellow Pages on forklift hire, YouTube videos and Facebook posts. The three UK searches we checked look no different: all three had a local pack above the organic results, and "forklift hire" also had Yell in the top ten twice. A page at position 3 under an AI summary and three Google Maps listings sits, on a smartphone screen, much further down than the number 3 suggests.
Here the job isn't climbing positions: it's getting into the elements that sit above. For local searches that means looking after your profile for the local pack, for informational ones getting cited in AI Overviews.
What is almost never the reason (and gets paid for anyway)
Some jobs sell well because they're concrete and can be ticked off. On an indexed site that doesn't rank, they're rarely the cause:
- resubmitting the sitemap: it helps Google discover pages, not rank them;
- switching SEO plugin: Yoast and Rank Math handle the same tags, and the content stays the same;
- pushing the speed score from 70 to 95: a very slow site loses positions, a decent one doesn't gain any by becoming excellent.
There is a real exception on speed: if your pages take several seconds to load on a smartphone, the problem exists and we've covered it in why is my website slow. But in the test, of the 24 results at the top, 8 were on sites with fewer than 200 referring domains and almost all were on pages with no links detected. What they had is what Google, in its guidance on helpful content, calls people-first content: a page that answers the person searching.
If after all these checks the drop came on a specific date, then yes, it's worth looking at algorithm updates and the Manual actions report, as we explain in our guide to Google penalties.
FAQs
It depends on where it has stalled. With no impressions on the query, a dedicated page is usually missing; beyond 50 the page answers a different intent; between 11 and 20 it needs internal links and fuller content; between 4 and 10 authority against competitors is what counts. The average position per page in Search Console tells you which case you're in.
It depends on the competition, but it's rarely a matter of weeks. According to a 2025 Ahrefs study of 1 million URLs, only 1.74% of new pages reach the top ten within a year. In a small niche like Sunscape's, it took six months to get 38 keywords on page one.
It depends on where you want to get to. In our test on 8 Italian commercial searches, 79.2% of pages in the top three had no links detected to the page: to reach page one, relevance comes first. To move from position 6 to 2, against competitors with equivalent pages, links from sites in your sector become the main lever.
Rarely, for specific commercial searches. In our test, across 160 positions from first to twentieth on 8 Italian SERPs, there was a single homepage, at position 11. The exception is a business that offers only that service: on Google UK, for "3d printing service", two of the top three were homepages of 3D printing firms.
It depends. If they answer different intents they can coexist: Ahrefs, looking at 9,700 cases on its own site, found only one that needed fixing out of 80 checked by hand. They become a problem when they're duplicates, because Google generally shows no more than two pages from the same site in the top results.
In brief
First check that the page is in the index.
Then look at average position, page by page.
No impressions: the page is missing. Beyond 50: wrong intent. Between 11 and 20: details and internal links. Between 4 and 10: authority.
The homepage rarely ranks for services: 1 out of 160 in our Italian test.
Sitemaps, plugins and speed scores are rarely the answer.
Something people in this industry rarely say: the sentence "my website isn't ranking" is already half the problem. Google doesn't rank websites, it ranks pages for searches. As long as the question is about the whole site, the answer will be a fix for the whole site: technical work, plugins, a redesign. Once the question becomes "which page, for which search, stuck at which position", the answer usually fits on a sheet with three columns and turns into a work plan. Sunscape went from 1 to 38 keywords on page one without any link building campaign: it changed the question.
If you'd like to build that sheet for your own site, see how we work in our SEO service or start straight from your data in a consultation. Until next time!
