The Super Cleaners: first on the map, invisible in search, and the ten days we spent finding out why

The situation

The Super Cleaners is a domestic and commercial cleaning company working across Portsmouth, Southampton and Chichester. It runs three Google Business Profiles, one per city, and they do well. Between April and early September 2026 those three listings produced 206 phone calls, 328 requests for directions and 614 clicks through to the website.

The website tells a different story. Over the 90 days from 12 June to 9 September 2026, Google Search Console recorded 513 clicks and 64,200 impressions. That is roughly 175 clicks a month. An impression means the site was shown somewhere in a set of search results, so the site appeared sixty four thousand times and was clicked five hundred.

Two numbers explain the rest of this article. First, 78% of those 513 clicks came from the website button on the Business Profiles rather than from an organic search result. Second, the 36 service pages on the site collected 5 clicks between them in 90 days. Not five each. Five in total.

We started work on 12 September 2026. Ten days is nowhere near long enough to change a ranking, so this is not a results article. It is an account of what an audit turns up when a business is visible on the map and absent from the results below it, and what we chose to fix first.

Why Maps can be strong while Search is empty

Google is running two systems on the same page. One ranks business listings, and it cares about where you are, how far away the business is, how complete the profile is and what the reviews say. The other ranks web pages, and it cares about what the page says, whether Google can read it, and who links to it. The two barely share signals, so a business can sit at position 1 or 2 on the map and be nowhere in the ten blue links underneath.

That is this site. The three profiles rank first or second for the terms that matter locally. The pages built to catch the same searches are not competing.

The clearest symptom was internal. On 12 of the 20 terms we track, the homepage outranked the company's own city page for that term. For "end of tenancy cleaning southampton", the homepage sat at position 26 and the dedicated Southampton end of tenancy page sat at position 70.

Position 26 is the third page of results. Position 70 is the seventh. In practical terms neither exists, but the pattern matters more than the positions. Google looked at the site, could not work out which page answered the question, and defaulted to the generic one. The purpose built page, the one with the right service and the right city in its title, was the one being ignored.

The same shape appeared in carpet cleaning. The hub page, the one meant to sit above the city pages and hold the topic together, recorded zero impressions. Its three city pages pulled 1,328 impressions between them at positions 46 to 67. The children were faintly visible. The parent was invisible.

What the audit turned up

Pages Google could not see

Of 55 URLs that should have been listed, 11 were not in Google's index at all. Four of those carried a noindex instruction, a line of code that tells Google to leave a page out of its results, recorded during a single crawl in December 2023. Google had not been back to those pages since. They are heavily linked from the site's own navigation, which usually gets a page recrawled quickly, and they had still been sitting untouched for close to three years.

A page can be fixed on your website and still be broken in Google's copy of your website, and the gap between the two is measured in months, not minutes.

Pages Google could see, and was displaying badly

A search result is two lines of text: the title and the description underneath. They are the only advert most people will ever see for the business, and they were in poor shape.

• 13 titles ran over 60 characters and were being cut off mid phrase in the results.
• 23 descriptions ran between 230 and 244 characters. Google shows roughly 155. About half of each one was written, published and never seen by anyone.
• 5 hub pages had no description at all, so Google assembled its own from whatever text it found first.
• One page had lost the first letter of its description. Google had been showing "rusted commercial cleaning in Portsmouth..." where the page meant to say "Trusted".

That one costs nothing to fix, and there is no way to know how many people read it and moved on.

No measurement at all

Google Analytics had never been installed. The property existed in the account, correctly set up, and had never received a single hit. So there was no record of which pages people landed on, which ones they left immediately, or which ones produced an enquiry. Every decision about the website up to this point had been made without data.

Housekeeping that had drifted

The XML sitemap, the file that lists your pages for search engines, had last been fetched by Google in March 2025 and still contained entries dating back to 2021. The robots.txt file, which is the first file a search engine reads on any site, had no line pointing to the sitemap at all. Neither is dramatic. Both make the site harder to crawl than it needs to be.

On speed, the site scored 54 out of 100 on a mobile Lighthouse test, with 1,072 KiB of page weight and a largest contentful paint of 7.3 seconds. That last figure is the time until the main content of the page actually appears. Somebody standing in an empty flat searching for an end of tenancy clean on their phone is waiting more than seven seconds to see anything.

The site also runs PHP 8.0.30, a version that stopped receiving security updates. That is not a ranking issue. It is the kind of thing that is fine until the day it is not.

Two spam campaigns, running at once

Both started within days of each other in the spring.

The first was a backlink campaign from roughly mid April. Links to a site are a ranking signal, and a cheap way to attack a competitor is to point thousands of junk links at them. We disavowed 583 domains on 12 September, which is the formal way of telling Google to ignore them.

The second was comment spam, starting 31 March. The site had collected 1,009 comments from 557 distinct IP addresses. Every one was junk and none came from a customer. The sending domains included services called seoautomationpro.com and verifiedlinklist.com, which at least have the honesty to say what they are. All 1,009 were removed and the comment form was closed, so the route is shut rather than merely swept.

We cannot prove either campaign cost the business anything, and we are not going to claim it did. What we can say is that both were live, both were pointed at this site, and neither needed to stay.

What we fixed first, and why in that order

The ordering rule is simple. Anything that depends on Google coming back to look gets started on day one, because the waiting is the slow part, not the work. Everything else can be done while that queue moves.

So, in the first ten days:

• The disavow was filed on the first day of the engagement, covering 583 domains.
• The sitemaps were cleaned and resubmitted. Google re-downloaded them within three days, after eighteen months of not bothering.
• robots.txt was corrected to point at the sitemap.
• Author archive pages were set to noindex. These are automatically generated pages with no content of their own, and on a small site they dilute the crawling effort that should be going to service pages.
• A missing H1 was added. The H1 is the main heading of a page, and its absence leaves Google guessing what the page is for.
• 37 titles and descriptions were rewritten and deployed, including the one missing its first letter.
• Analytics was installed, so that from 12 September onwards there is a record.
• The comment spam was removed and the form closed.

None of that is glamorous and none of it required a redesign. It is the unopened post of a website, and it had been accumulating for years while the Business Profiles quietly carried the business.

What happens next, and when

Here is the honest version of the timeline, because this is where SEO is most often oversold.

The disavow will take weeks to have any effect, and Google will never confirm it worked. The file is accepted and then the effect, if any, appears gradually in rankings that also move for a dozen other reasons.

The four stuck pages will be recrawled when Google decides to recrawl them. We have resubmitted them and cleared the obstacles. A page that has not been visited in nearly three years is not at the front of the queue, so weeks is realistic and longer is possible.

The rewritten titles and descriptions are the fastest moving change, and they show up as pages are recrawled over days and weeks. They will not move a page from position 26 to position 3. What they can do is change how many people click a result the site already has, which on 64,200 impressions is worth having on its own.

Analytics starts from zero. There is no way to backfill the years that were never recorded. The first genuinely useful read on behaviour comes after about a month of collection.

Then the real work starts, and it is structural rather than technical. The homepage is competing with the city pages and winning, which means the internal linking and the on page focus have to be rebuilt so that each city page is unambiguously the answer for its own city. The carpet cleaning hub has to earn its position above its three children rather than sitting below them. The 36 service pages need to justify existing, individually. And the site needs to get faster, on a phone, where the searches actually happen.

We will publish a follow up at month three and again at month six, with the numbers as they land, including the ones that do not move. The starting line is on the record above, which is the point of writing this now rather than later. For a sense of what the same work looks like once there are numbers to show, our eleven week write up of a local SEO retainer for a Montreal print shop picks up where this one stops.