Website analytics provider Diib found an unusually high rate of sharp Google traffic losses during September’s spam update. In its sample, 5.8% lost half their traffic from unpaid Google search while traffic from elsewhere stayed comparatively stable. The baseline rate in periods without an update was 1.3%.

Google confirms that the global update ran from 24 September to 8 October. Its announcement names no affected industries and provides no list of websites classified as spam. Diib’s finding is an observed traffic pattern, not a count of penalties issued by Google.

For businesses commissioning SEO services for Melbourne businesses, a decline on that scale warrants an explanation tied to the affected pages. Losing traffic to a general advice article is a different problem from losing visibility for an emergency service. The first may reduce readership; the second can remove a source of bookings. A site-wide percentage conceals which happened.

September’s sample shows a bigger fall, with an uneven comparison

Diib’s September comparison covered 34,597 connected websites receiving at least ten search visits a day. It counted sites where search traffic halved while other traffic remained within 25% of normal.

There is a timing wrinkle. The September measurement covered the rollout, whereas the other update comparisons generally used a period after completion. The sample also excludes sites below its traffic threshold, so a very small local business may not resemble the websites measured.

A separate analysis, published on 24 October, followed sharp losses around five earlier update periods. Of the declines it tracked, 68% remained at least 25% below their earlier traffic six weeks later. Only 18% returned to within 10% of their previous level.

That follow-up covered 7,092 drops across 6,566 websites and used a different loss threshold. September wasn’t in the follow-up. Too little time had passed, so its recovery figures can’t predict what September’s affected sites will experience.

The findings still give owners grounds to question a promise that a large fall will resolve itself in a few weeks. Waiting is a decision with a cost if the missing visits were bringing customers. The consequence is smaller if most of the loss came from an advice article that rarely produced enquiries.

A technical fault can also produce a search-specific decline. Stable traffic from other sources doesn’t rule out a blocked page or a failed redirect. Neither Diib analysis establishes which explanation applies to an individual site.

Repeated suburb pages need more than different place names

Google’s spam policies describe doorway abuse as pages created for similar searches that lead readers through less useful intermediate pages. Its scaled-content policy concerns large amounts of material produced primarily to manipulate rankings rather than help users.

The policy covers content made by people, automation or both. Hiring a human writer doesn’t, by itself, settle whether a page is useful. Nor does using AI establish that the scaled-content policy has been broken.

For a service business, suburb pages are a practical place to examine that distinction. A company may need to explain where it works and show projects completed in different areas. But changing the place name above identical paragraphs gives the next reader little extra information.

A project page can do more. It can explain an access problem, the work completed and what the customer needed to arrange beforehand. Someone with a similar property can use those details when deciding whether to request a quote.

The editorial decision then depends on the pages already published. Several pages might be competing to answer the same question, in which case combining them could make the information easier to use. Elsewhere, a page may omit a real service limitation that customers repeatedly ask about.

Deleting all local pages would discard those differences along with the repetition. Expanding every page would leave the duplication in place. A page-by-page review needs to identify the reader’s unanswered question before adding another paragraph.

Google also treats links created primarily to manipulate rankings as link spam. A report giving only the number acquired leaves the owner unable to examine where they appeared or why. The publication and arrangement behind each link belong in the review.

A manual notice gives an owner different evidence

Google’s Search Console documentation says a manual action follows a human review. Google reports it through the Manual Actions report and message centre, giving the site owner an identified issue to investigate.

An empty report doesn’t rule out an automated ranking change. Google describes spam updates separately as improvements to automated detection, so a business may see a decline without receiving an individual notice explaining it.

The first investigation should preserve dates and URLs. A service page moved during a redesign, an altered redirect or a tracking change can coincide with an announced update. Correcting the wrong problem wastes time and can obscure what happened.

Search Console can help separate fewer impressions from fewer clicks, while its indexing information shows whether pages are appearing as expected. Analytics then records visits, and the business’s enquiry records show whether the lost audience was asking for work.

For a manual action, the stated issue can be corrected and a review requested. An unexplained decline around an automated update needs a broader investigation into the pages and searches affected. The two cases don’t support the same recovery timetable.

Google’s guidance says a site may improve over months as its automated systems learn that it complies with the policies. An agency promising recovery on a fixed date is offering more certainty than that guidance provides.