How does Google truly count impressions in Search Console?

What is an impression according to Google?

An impression is recorded as soon as a URL from your site appears in a SERP, regardless of its actual position or its effective visibility to the user. Google does not check if the page has been scrolled, if the element is in the viewport, or if the user has even seen it.

This broad definition applies to all types of results: classic organic, People Also Ask, featured snippets, images, videos. As soon as your content is technically present in the response served by Google, the counter starts.

Why is Mueller's statement problematic?

Because it completely dissociates impression from actual exposure. A result in position 50 generates an impression just like a result in position 1. The average CTR of a page can therefore be artificially low, not because it performs poorly, but because it accumulates invisible impressions.

This mechanism creates a systematic bias in the Search Console analysis. High-volume queries with an average ranking generate hundreds of impressions without any chance of clicks, while a highly qualified query in position 3 may show a "low" CTR compared to standard benchmarks.

What are the implications for performance tracking?

The impressions/clicks ratio becomes a misleading indicator if not crossed with the average position. A page with 10,000 impressions and 50 clicks (CTR 0.5%) may actually perform much better than a page with 1,000 impressions and 100 clicks (CTR 10%) if the former oscillates between positions 8-12 and the latter is in positions 1-2.

The Search Console filters do not easily segment actually exposed impressions (top 3, top 5) from phantom impressions (position 15+). This technical limitation requires exporting the data and manually reprocessing it to get a reliable view.

• An impression does not mean effective viewing by the user
• Raw CTR is mechanically underestimated for pages ranked beyond the first screen
• Standard CTR benchmarks (position 1 = 30%, position 3 = 15%) only apply to actually visible impressions
• Search Console does not automatically filter impressions outside the viewport
• Performance analysis requires reprocessing raw data to be useful

Is this definition consistent with field observations?

Yes, and that’s precisely the problem. We regularly see pages with thousands of impressions and a ridiculous CTR, while their average ranking is on page 2 or 3. Google confirms here that it’s not a bug; it’s intentional.

This technical logic likely serves to simplify tracking on Google’s side—count what is served rather than what is seen—but it makes interpreting Search Console metrics much more complex for us. A count based on the viewport would have been more representative of user reality.

What nuances should we add to this statement?

Mueller does not specify if this rule applies differently depending on the device. On mobile, the viewport is reduced, so a position 5 may already be off-screen initially. On desktop with enriched SERPs (People Also Ask, images, Knowledge Graph), even a position 3 may require scrolling.

Another unclear point: [To be verified] how Google handles SERPs with dynamic lazy loading or results loaded in JavaScript after interaction. The official documentation does not detail these cases, and empirical tests show inconsistent behaviors depending on the type of query.

When does this rule really impact analysis?

The impact is maximal on content sites that rank for large volumes of long-tail keywords with medium positions (7-15). Their overall Search Console CTR will be systematically underestimated, giving the impression of poor performance while the visible pages normally convert.

Conversely, sites with few keywords but premium positions (top 3 consistently) show CTR close to reality. The bias still exists, but it is marginal compared to the volume of actually exposed impressions.

How to correctly interpret Search Console data?

The first rule: always cross-reference CTR and average position. A low CTR with a high average position (8+) is perfectly normal. A low CTR with an average position in the top 3 signals a real problem (unattractive title/meta description, competing SERP features, etc.).

The second rule: segment queries by position range. Export the data, isolate queries in positions 1-3, then 4-10, then 11-20. Compare the average CTR of each segment to suitable benchmarks. It’s tedious, but it’s the only way to get a reliable view.

What mistakes should be avoided in impression analysis?

Never use raw impression volume as an indicator of real visibility. A page with 50,000 impressions/month may only be seen 2,000 times if it oscillates in position 15. An impression is only a technical proxy, not a business metric.

A common mistake: optimizing overall CTR without considering position. You can infinitely improve your titles, but if your pages are in position 12, the CTR will remain mediocre because no one sees them. In this case, the priority is to gain positions, not to polish the wording.

What adjustments should be made to your SEO strategy?

Focus your optimization efforts on title/meta description for pages already well positioned (top 5 minimum). That’s where the return on investment is immediate. For pages in positions 8-20, prioritize gaining positions through backlinks or content optimization first.

Set up a custom dashboard that automatically filters queries by position range and calculates adjusted CTR. Many SEO tools (Looker Studio, Power BI) allow connecting to the Search Console API and automating this processing. Without this, you're driving blind.

• Export Search Console data monthly (queries + pages) with average position
• Segment queries by range: top 3, top 5, top 10, top 20, beyond
• Calculate the average CTR per segment and compare it to adjusted benchmarks
• Identify top 5 pages with abnormally low CTR → optimize title/meta
• Identify pages ranked 6-15 with good potential → prioritize gaining positions
• Never compare overall CTR between two periods without checking position distribution