Search Visibility
Most People Read Search Console Like a Scoreboard. Read It Like a Practitioner
Clicks up, good. Impressions up, good. Position down a bit, shrug. That's a report, not an audit. Search Console is the one tool that shows exactly what Google displayed and exactly what people did about it — if you know the one question to ask and the maths traps to avoid.
The Four Numbers
Clicks, impressions, CTR, position — and their edges
Search Console gives you four numbers, and three of them have edges most people miss. Clicks are the honest one: someone saw your result and clicked it. Impressions are looser than they sound — they count the times you were eligible to be seen, including results nobody scrolled to, not "people who saw you." CTR is clicks divided by impressions for that row — a ratio, not a count. Average position is a mean of where you showed across those impressions — again, not a count.
That distinction between counts and ratios is where most Search Console reading quietly falls apart. Hold on to it; it's the difference between a number you can trust and a number that just looks plausible.
The One Question
You rank here. Are you getting the clicks that ranking should earn?
That's the question, and everything else in Search Console is context for it. Rankings are an input; clicks are the output; the gap between them is a diagnosis you can make from data alone, without guessing at intent or crawling anything. Click-through by position follows a reliable shape — the top spot takes a big share, it falls away steeply through the top few, and the bottom of page one is low single digits. Page two is, in practical terms, no traffic at all.
Overlay your actual click-through on that shape and three readings appear. If it matches the curve for your position, the page performs as its ranking warrants — more clicks means a better ranking, which is a content-and-links job. If it sits well below, you're winning the ranking and losing the click — a title-and-listing problem, fixable this week without publishing anything. If it sits well above, something's over-performing, usually your brand — understand it before you touch it, and don't "optimise" a page that's already beating its position.
The Maths Traps
Three ways the arithmetic bites you
Never sum CTR, never average average position. CTR is a ratio — you can't recover a group's click-through by adding or averaging the rows, because a row with ten thousand impressions and a row with three each contribute one number and they're nowhere near equal. For a group, divide the summed clicks by the summed impressions. Average position is worse, because it's already a mean; averaging means gives you the mean of means, which isn't the truth unless every row has identical impressions. A spreadsheet will happily compute the wrong answer and it will look convincing.
The totals won't add up, and that's correct. Filter by query or page and the totals drop below the unfiltered numbers. That's not data loss — Google withholds rare, potentially identifying searches. The site total is the truth; the sum of the visible rows is a subset. Report the total, analyse the rows, and never present the two as if they reconcile.
Row-counting isn't analysis. "We rank for four thousand keywords" means almost nothing — most of those have three impressions and no buying intent. Weight everything by impressions, and when you sort, sort by impressions or clicks, never by CTR or position, which float the noise straight to the top. A query with one impression and one click has a perfect click-through rate. It is not your best keyword.
What ARIA Shows You
The conclusion, not the export
None of this is hard — it's arithmetic a careful reader with a spreadsheet can do. The problem is that it takes an hour or two, it has to be redone every month, and the parts that are easy to get wrong (weighting by impressions, dividing sums instead of averaging ratios, keeping brand out of the non-brand numbers) are exactly the parts that get skipped when you're busy running the business.
The finding, in plain English: "This page ranks near the top for a term with several thousand monthly impressions and earns roughly a third of the click-through that position should earn. Brand searches are separated out, so this is real. The title is generic — rewriting it is about an hour, no developer required."
What you get is the sentence at the end — the finding, scored, with the recommendation attached — not another CSV to read at eleven at night.
Related In This Guide
Where to go next
Two of the readings above have their own guides: the "well below the curve" case is the CTR gap, and the "just off page one" band is your striking-distance keywords. When the numbers move over time, read why rankings slide before you react. And the full end-to-end pass this fits inside is the SEO audit checklist.
Search Console shows what happened; it doesn't fix the technical reasons behind a fault. For the hands-on side — crawlability, speed, structured data — that's Audry's territory: Audry's Site Audit.
Read your own Search Console data properly
The data is free and you already have it. What's scarce is reading it correctly, every month, without the arithmetic traps. Connect your site and let ARIA do the reading for you.