A business can have a page dedicated to its main service and still find that Google keeps sending searchers to an advice article instead.
That was the problem in a project I’ve been working on. The client had a main service page, more specific service pages and a substantial advice section. For some important searches, an article about tackling the problem yourself kept appearing ahead of the page intended to explain the professional service.
The obvious question was: how do we get the right page ranking?
The investigation became more interesting than that. The intended page had more recorded backlinks. It performed differently for different types of searches. At times, it appeared as a supporting link in an AI Overview while the advice article appeared in the regular results. After content changes, some performance signals improved while others needed closer examination.
This is an account of that investigation at that stage, rather than a claim that one change solved everything. I’ve left out the client, sector, search terms and identifying figures, but kept the observations and reasoning that made the work useful.
First, what did we mean by the wrong page?
From the business’s perspective, the main service page was the natural destination for someone looking for professional help. It explained the service and offered a route to an enquiry.
The advice article served a different purpose. It addressed people considering doing the work themselves, including the practical difficulties and reasons they might need a professional.
But a broad search doesn’t always tell us which of those visitors is searching. Someone typing the name of a service might want a supplier, an explanation of the options or instructions for doing it themselves.
So I needed to distinguish the page the business wanted to rank from the page that best answered the search. Calling it keyword cannibalisation was a useful starting point, but it didn’t explain why Google was making that selection or what we should change.
The backlink comparison didn’t settle it
One early comparison made the result seem particularly puzzling. The main service page had a much stronger recorded backlink profile. Our link data wasn’t showing backlinks to the competing advice article.
That was a finding about the data available to us, not proof that Google knew of no links to the article. Even so, it meant that simply assuming the article was winning because it had more external links didn’t fit the evidence.
It also gave us a practical reason to investigate the content and internal structure before deciding that building more backlinks was the answer. The existing links weren’t enough to make the intended page the consistent choice.
The advice article was answering more of the question
Looking beyond the page titles helped explain the overlap.
The article covered the difficulties of doing the work yourself, why attempts could fail, the available approaches and when professional intervention made sense. Although its stated subject was DIY, it answered quite a lot of what someone researching the broader service might want to know.
The main page had an important role as a hub, directing visitors towards more specific services and methods. However, it needed to make a stronger case as a useful destination in its own right.
Our working hypothesis was that the advice article could be satisfying parts of the broader search more directly. That wasn’t something we could prove by reading the pages, but it gave us a specific weakness to address.
The question became: if someone lands on the main service page with no prior knowledge, can they understand their options and make a useful next decision without having to piece the answer together across several other pages?
The navigation wasn’t reflecting the page’s intended importance
The internal-link review added another important piece.
The main service page was much less well linked internally than some of the more specific service pages. The navigation offered routes into particular customer categories, while giving the broader parent page less prominence than its intended role suggested.
In other words, we regarded it as a central page, but the routes through the website didn’t consistently reflect that.
One concrete change was to add the main service page to the relevant guide navigation. We also reviewed links between the main page and its related content as part of the wider work.
This was a sensible correction to the structure, not proof that navigation alone had caused the ranking problem. Google’s internal-link guidance also recommends linking to important pages from other relevant pages and using descriptive link text.
We were also reviewing a large collection of location pages, many with overlapping content and very little search traffic. That raised a related question: which pages provided useful local information, and which could be consolidated into stronger regional resources? This formed part of the wider structural work, rather than something we could identify as the cause of Google choosing the advice article.
We rewrote the main page around the decisions a customer needed to make
The rewrite gave the main page a fuller explanation of the service, while retaining routes to specialist information.
It addressed the questions someone comparing their options would reasonably have:
- What do the different approaches achieve?
- Which circumstances suit each approach?
- What affects cost and timescales?
- How do requirements differ between customer types?
- What should someone understand about guarantees and limitations?
- When might doing it yourself be inappropriate?
Relevant case studies were retained, and the main page continued to link to more detailed material. The aim was to make the overview useful on its own, with the supporting pages available when someone needed greater depth.
That distinction mattered. Making every page cover everything would have left us with more overlap. Each page needed a clear job, even where some explanation naturally crossed over.
Why not just remove the competing article?
It can be tempting to remove the page you don’t want ranking and hope the preferred page takes its place. But the advice article addressed a real question. Removing it could mean losing a useful resource without securing the result we wanted.
For this situation, improving the main page’s usefulness and position within the website was a more defensible starting point.
That’s different from finding two near-identical pages doing the same job, where consolidation might be appropriate. A useful advice article and a service page can both have a purpose.
A canonical tag isn’t a general way to tell Google which distinct page you’d rather rank, either. Google’s canonicalisation guidance concerns duplicate or very similar pages. I wouldn’t use it simply to express a commercial preference between two different resources.
The results depended on which searches we looked at
As the work developed, the main page showed strong visibility for a more specifically commercial version of the search. The broader phrase remained less straightforward, with the advice page continuing to appear in checks.
That made a single statement such as “the page is ranking better” inadequate. We needed to ask: for which searches, and for which audience?
A strong result for a narrower phrase was encouraging. It didn’t establish that the broader page-selection issue had been resolved. Equally, finding the advice article for a broad phrase didn’t mean the service page was performing poorly across every related search.
The AI Overview added a different observation
At points during monitoring, the main service page appeared as a supporting link in an AI Overview, while the advice article appeared in the regular organic listings.
That was worth recording, but it wasn’t a reliable basis for declaring that Google had finally selected our preferred page, or for explaining why it selected different pages in those placements.
Google’s documentation on AI features explains that these experiences can use multiple related searches to identify supporting information. I treated the observed placements separately, rather than expecting the AI Overview to mirror the regular listings.
The practical questions remained: where were the pages appearing, for which searches, and were relevant visitors reaching the business?
A better average position wasn’t the whole result
Another complication was that the updated main page showed an improvement in average position alongside a fall in impressions.
That combination needs investigation. It doesn’t, on its own, tell us whether the content changes were successful.
For example, imagine a page stops appearing for a large number of searches where it previously ranked poorly. Its average position across the remaining impressions could improve even without gaining ground for the searches that matter most. That’s an illustrative possibility, not a confirmed explanation of this client’s figures.
Changes in demand or in the mix of searches could also affect the picture. Google’s Search Console performance report allows us to examine queries and pages alongside clicks, impressions, click-through rate and average position. Those comparisons are more useful here than treating one headline metric as the verdict.
What had we established?
We had identified a mismatch between the main page’s intended importance and its internal prominence. We had also found that the advice article answered substantial parts of the broader question, while the main page needed to provide a more complete overview.
The work addressed those weaknesses through navigation and content changes. Monitoring then showed a mixed picture: encouraging visibility for some searches, continuing appearances of the advice page and metrics that needed to be read together.
What we hadn’t established was a single cause, a permanent resolution or a clean uplift attributable to one edit. Several related changes had been made, and search performance was still developing during the period described.
That matters when assessing SEO work. A reasonable change, an encouraging observation and a proven outcome are different things.
What this means if Google is choosing a different page on your website
Start by comparing the pages for the particular searches where the mismatch occurs. Look at what each page helps a visitor understand, how the rest of the website links to it and whether the intended destination is actually the strongest answer.
Technical checks matter too, including whether the intended URL is indexed and whether redirects or canonical signals point somewhere unexpected. But the response should follow the evidence: clearer content, better internal links, consolidation and technical corrections solve different problems.
In this investigation, the useful progress came from examining the content, structure and search data together. The business’s preferred landing page gave us a goal; it didn’t explain why Google should choose it.
If Google keeps sending visitors to an unexpected page on your website, get in touch. I can help you investigate what’s happening, identify the changes worth making and assess the results without reducing everything to a single ranking.