Search Console Performance Report: Clicks, Impressions, CTR, Position

The Performance report is the most-opened section of Search Console, and it rests on four numbers: clicks, impressions, CTR and average position. The problem is that the everyday meaning of those four words is not the same as Google's precise definition of them — and a misreading leads straight to a bad decision.

The classic example: "our average position went from 12 to 18, our rankings got worse." In reality this usually happens because the site started appearing for many more new queries — new queries enter at low positions, dragging the average down, while total clicks go up. The number got worse; the situation got better.

This article works through every element of the report using Google's official definitions, and shows how to turn the numbers into concrete tasks. For the tool as a whole, see what Google Search Console is and how it works.

The four metrics and their exact definitions

Google defines them as follows:

• Clicks: "The number of times a user clicked your site from Google Search results."
• Impressions: "How many times your site appeared in Search results."
• CTR: "The click count divided by the impression count."
• Average position: "The average position of the topmost result from your site."

The word to notice in that last definition is topmost. If two of your pages rank 3rd and 9th for a query, the report records position 3, not 6.

Impressions: when they are counted

An impression means your site appeared on a results page. Several subtleties follow:

First, impressions vastly outnumber clicks, and that is normal. Ten thousand impressions and 200 clicks is a 2% CTR — perfectly ordinary for lower positions.

Second, an impression is not viewability. Even if the user never scrolls down to the eighth result, that result is usually still counted as an impression. So "impressions but no clicks" is entirely expected at low positions.

Third, different result types are counted separately. Image results, news carousels, "People also ask" blocks — each is tracked under the Search appearance dimension. Rich results depend on structured data, which is a topic of its own: schema markup and entity linking.

Fourth, a sudden jump in impressions is not always good news. If the growth comes entirely from irrelevant queries — your site sells SEO services but impressions arrive from "free image download" style queries — that is a signal of content or intent mismatch.

Average position: the most misleading number

Average position is an average — and averages always hide structure. There are three traps:

Trap 1 — the scale effect. When a site starts appearing for new queries, those queries typically enter around position 30–50. The average worsens while coverage expands. Always read average position together with impressions.

Trap 2 — the "topmost" rule. If three of your pages appear for one query, only the highest is counted. That makes cannibalisation invisible in the average — to find it you must filter by the query and switch to the Pages tab.

Trap 3 — personalisation and context. Google's own documentation warns: "Search results are specific to the time, place, device, and recent history of the person searching." The position you see in your browser may not match the report's average. The report is an average over thousands of real queries; your browser is one person at one moment.

CTR: meaningless without context

CTR is a simple formula — clicks ÷ impressions — but "good CTR" is not a universal number. It depends on three things:

Position. A 25% CTR can be normal at position 1; at position 15, even 1% is good. CTR only makes sense compared within the same position band.

Query intent. Navigational queries (your brand name) produce very high CTR. Informational queries where the answer is given directly in the SERP — rich results, "People also ask" — produce low CTR through no fault of yours.

How crowded the SERP is. Ads, a map pack, a video carousel all push organic results down and reduce CTR even when your position has not moved.

So the practical question is not "is our CTR good?" but "why is this page's CTR lower than our other pages at the same position?" The answer is usually in the title and meta description.

Dimensions and filters

The report table has six dimensions, each answering a different question:

The real power lies in combining filters. A few practical examples:

• Page filter + Queries tab → which queries this page appears for (finding content gaps)
• Query filter + Pages tab → which of our pages compete for this query (cannibalisation)
• Country filter + date comparison → did something change in this market
• Device filter = Mobile + sort by CTR → are our mobile snippets being truncated

There is also a comparison mode that places two date ranges side by side. It is the main tool for analysis after an algorithm update or a site migration. Getting the redirects right during a migration is a topic of its own: 301 vs 302 redirects.

Limits of the data

This section matters most for reading the report correctly, because "numbers that don't add up" are almost always explained here.

Totals do not equal the sum of table rows. Google explains that the chart total can be higher than the sum of the table rows — because tables display a maximum of 1,000 rows, and "Tables in the performance reports omit rare queries to protect user privacy."

Anonymized queries. Officially: "To protect user privacy, the Performance report doesn't show all data. For example, we might not track some queries that are made a very small number of times or those that contain personal or sensitive information." On small and new sites the hidden share can be large — do not be surprised if the clicks in the query list add up to half the site total.

Filtering by query changes the numbers. The moment you apply a query filter, anonymized queries leave the calculation. So "the sum where query = X" and "the overall total" will never fully reconcile.

Data lags by two to three days. "Normally, however, collected data should be available in 2-3 days." The most recent days are preliminary and can still change.

The date range is 16 months at most. Longer comparisons require regular exports. On large sites, the way past the 1,000-row limit is the Search Console API.

Five practical workflows

Opening the report and staring at numbers is not work. The five workflows below turn numbers into concrete tasks.

1. Quick CTR wins. Filter: last 3 months, average position 1–8. Sort the table by CTR ascending. The pages at the top rank well but are not being chosen — they are the first candidates for a title and meta description rewrite.

2. "Second page" opportunities. Filter: average position 8–20, high impressions. These queries are already close. Deepening the content and adding internal links often gains several positions — for the structural logic see topical authority and content clusters.

3. A cannibalisation check. Filter one query and switch to the Pages tab. If two or three of your pages appear for it, Google is struggling to pick the most relevant one. The fix: consolidate, differentiate, or 301 redirect one into the other.

4. Diagnosing a drop. When clicks fall, the top-level chart never shows the cause. Turn on comparison mode (last 28 days vs previous 28), then sort the Pages and Queries tabs by the difference column. The drop is usually concentrated in a handful of pages.

5. Tracking new content. After publishing, apply a page filter. The sequence should be: impressions appear first, then clicks. If impressions never appear at all, the problem is not performance — it is indexing: the page indexing report.

Common mistakes

1. Treating average position as a KPI. As coverage expands, the number naturally worsens. Use clicks and conversions as your KPI instead.

2. Comparing CTR without regard to position. Putting a page at position 3 next to one at position 20 tells you nothing.

3. Forgetting an active filter. A country or device filter left in place distorts every subsequent reading.

4. Treating mismatched totals as a bug. The 1,000-row limit and anonymized queries are documented behaviour.

5. Making decisions on the last two or three days. Those days are preliminary and will change.

6. Not separating brand queries. Brand traffic inflates the total. Without a filter that excludes your brand name you cannot answer "is our SEO working?".