Search Console shows four headline numbers and all four are defined differently from how most people read them. Average position is the riskiest, because it looks simple and bad decisions come out of it.
This is about working with those four numbers rather than about setting Search Console up. If you have not verified your site yet, do that first and come back.
What average position averages
It averages over impressions, not over days and not over queries.
Suppose one page sits at position three for a high volume query and was shown a thousand times, and at position forty for ten marginal queries shown ten times each. The headline figure stays near three and a half. Now move those ten marginal queries from forty to twenty, which is real progress, and the overall number barely moves.
The practical consequence is that site level average position is close to meaningless. It only becomes useful once you have filtered by page and by query at the same time.
And position ten is not the tenth blue link
This is where the misreading happens, and the way out is to look at the results themselves.
On 15 August 2026 we pulled live Google results for seven Persian commercial queries, Iran location, Persian language, to a depth of twenty. Here is what occupied those twenty slots:
| Query | Ordinary results in the top 20 | Of those, Aparat or YouTube | Other blocks |
|---|---|---|---|
| WordPress theme | 18 | 4 | Images |
| Website design with WordPress | 17 | 2 | Images, video |
| Search Console | 17 | 3 | Images |
| Logo design | 17 | 3 | Images |
| Cloud hosting | 18 | 1 | Images |
| Site speed test | 19 | 1 | None |
| WordPress security plugin | 19 | 1 | None |
That is 125 ordinary results across the top twenty of those seven queries, and 15 of them, roughly twelve percent, were Aparat or YouTube videos. Five of the seven also carried an image block sitting in the middle of the results.
So when Search Console tells you your position is eight, there may be an image block and two videos between you and the top of the screen. The number is accurate. Your sense of how far you are from the fold is not.
Impressions and click through rate
An impression is recorded when your result appears in the results page, even if the user never scrolls that far. Rising impressions on their own are therefore not good news; you may simply be appearing for less relevant queries.
Click through rate is a plain division of clicks by impressions, so any query bringing lots of impressions and few clicks drags the headline down. We covered its relationship to ranking separately in what click through rate has to do with position.
One detail that eats a lot of time: the query rows never add up to the site total. Google drops very rare queries to protect privacy. This is not a bug and you should not go looking for a cause.
The three reports that change a decision
The others are not bad. These three are the ones you can act on.
The pages report and its not indexed reasons. Open this before anything else. A page that is not indexed makes every other number meaningless, and most projects that start with "we are not ranking" end right here.
The performance report, filtered to one page. Last twenty eight days against the previous period, for one specific URL. In that state average position becomes meaningful again.
The Core Web Vitals report. The only place you see genuine field data from your own users, grouped by page type. How it differs from what a testing tool returns is covered in the three metrics and how they are measured.
What Search Console will not tell you
Where your competitor sits. Your actual position on a given day. And whether the page that ranked is the one you wanted to rank.
We run our own rank tracking and we still open Search Console first. The reason is simple: a tracker only reports on the queries you already thought of. Search Console shows you queries you did not know you were appearing for, and most of the real opportunities are in there. The exact boundary between the two is in where Search Console stops and rank tracking starts.
When not to trust these numbers
When the volume is small. A site pulling a few hundred organic visits a month does not have enough impressions at query level for a two position move to mean anything. That move is usually noise rather than a result of your work.
In that situation, stop looking at the chart weekly. Set the range to three months and read the direction only. Separating normal fluctuation from a real drop is covered in the line between normal movement and an actual decline.
Common questions
Why does the rank I see in Google differ from Search Console
Because what you see is personalised, from one location, at one moment, and the Search Console figure is an average across every real impression in the selected range. They never converge exactly. The measurement methods are compared in the ways of checking your position in Google.
How long is the data kept
The performance report holds the last sixteen months. If year on year comparison matters to you, export monthly and keep it.
Why is a page in my sitemap not indexed
A sitemap is a suggestion, not an instruction. The real reason is written in the pages report. Start there rather than resubmitting the sitemap.
Where to go next
If you open the pages report and the list of not indexed reasons is long, the order you tackle them in matters. The route to follow when pages are not being seen gives you that order.
And if these numbers lead you to conclude the problem is structural rather than editorial, a technical and content structure review starts from these same reports. For a site that does not exist yet, building with an indexable structure removes half of this list from the start.
Sources
- Search Console performance report documentation for the exact definitions of clicks, impressions, CTR and position
- Page indexing report documentation for what each not indexed reason means
The search results data was pulled on 15 August 2026 and the sources were checked the same day.
Comments & Questions
Have a question about this article? Ask, we'll answer.
No comments yet; be the first.