Skip to main content

Google clicks fell 39% despite better index and position

Our write-up of a story first reported by Cossa

Google clicks fell 39% despite better index and position

An auto service site lost 293 Google clicks over identical 28-day periods — down 38.7% from 758 to 465. Pages in the index grew 9%, excluded URLs fell 34%, and the site's average position improved. A breakdown of individual URLs showed 95% of the drop came from the services section, not the blog or the site evenly.

Why average position masked the problem

Average position does not reveal which specific queries and pages changed. A commercial page can drop from fifth to tenth and lose most of its clicks while dozens of weak queries improve from 50th to 30th at the same time. The average looks better, but business traffic shrinks.

In the initial analysis, the site's average position improved slightly even as clicks fell nearly 39%. Google recommends not giving too much weight to absolute average position and looking for patterns in specific pages when traffic drops.

A few URLs explained most of the drop

Comparing URLs from the Pages tab across both periods showed that most of the loss was concentrated in a limited group of commercial pages:

  • Air conditioning repair: minus 64 clicks, roughly 22% of the entire site drop. Impressions, CTR and position all declined.
  • Radiator cleaning: from top five to the lower edge of page one. CTR fell nearly threefold, but the page also dropped from roughly fifth to tenth.
  • ABS module: clicks 18 → 2, impressions 1,075 → 418, average position 16.87 → 42.07.
  • Timing belt: clicks held steady (2 before and 2 after), but impressions fell from 2,558 to 1,453 — down 43%. Over 16 months the page had more than 56,000 impressions.

Seasonality explained only part of the decline

Auto service is a seasonal niche. Applying last year's dynamics to the current 758 clicks would suggest an expected level around 650. The actual result was 465. Seasonality may account for part of the decline, but not the full scale.

The acceleration in decline coincided with the May 2026 Core Update (21 May to 2 June). The June 2026 Spam Update began 24 June, the August 2026 Spam Update on 18 August. But temporal overlap does not prove causation. A core update is not a penalty against a specific site but a broad re-evaluation of results.

Why the entire site was not rewritten

The site has hundreds of service pages. After a drop it is easy to launch mass optimisation: rewrite all titles, expand texts, add keywords, rebuild H1s and order another hundred articles. But the data did not show the problem was evenly distributed.

Site-level CTR fell from 1.56% to 1.07%. Of 293 lost clicks, roughly 75 can be explained by fewer impressions at the previous CTR; the larger part is tied to lower click rates on existing impressions. But CTR depends on more than title and description: position, query type, ad blocks, maps, rich results, device and user intent all matter.

On the radiator cleaning page, CTR dropped from 12.3% to 4.4%, but position also fell from 4.95 to 10.07. Rewriting the snippet without accounting for this would treat the symptom.

The blog was not the source of the drop. In the first period articles received 3 clicks, in the second 16. A new article on brake fluid loss had 6 clicks, 942 impressions and an average position around 7.45; an article on suspension noise had 4 clicks, 288 impressions and a position around 8.6. The team continues developing informational content but does not use it as a universal fix for dropping commercial pages.

Instead of optimising several hundred URLs, the team formed a priority list: 8 to 10 pages accounting for most of the loss. For each page the analysis covers queries before and after the drop, search intent, the current top ten, content and commercial factors, internal linking, possible cannibalisation, and external signals.

TG X Copied

Project sponsors

Companies that keep this analytics open