Reading Search Console honestly
Search Console reports four metrics per query: clicks, impressions, click-through rate and average position. These are real data, not estimates. They say almost nothing about causes, though, and their own counting limits are documented by the tool itself.
What are the four metrics, and what do they say?
The performance report provides four metrics: clicks, impressions, click-through rate defined as clicks divided by impressions, and average position.
officialSource : Google Search Console, “Performance report (Search results)”
Each answers a different question, and the value of this tool comes from combining them rather than from reading them separately.
Impressions say your page was offered. It is the measure of the demand you are attached to, and the only one that does not depend on your writing choices.
Clicks say it was chosen. Between the two sits all the work on the title and the displayed promise.
Click-through rate is their ratio. It is never read alone: a high rate on three impressions says nothing, and a low rate on a query where your page has no business is good news.
Average position places the context, and it is the most misunderstood of the four.
How is average position calculated?
In the charts, average position is that of the topmost result from your site; in the tables, that of the address or grouping shown on the row.
officialSource : Google Search Console, “Performance report (Search results)”
Two different definitions depending on where you read it, and that is enough to produce disagreements between two people looking at the same account. In a chart it is the best result from the whole site; in a table it is the row shown. Comparing one with the other makes no sense.
There is a more serious trap, and it is arithmetical: that average improves when you lose badly placed impressions. A page that stops being offered on the queries where it appeared near the bottom sees its average position rise, with no progress having taken place. The reading rule that follows is simple and not negotiable: never position without impressions.
Why checking “by hand” proves nothing
Even where a query appears in your list, you may not see your site when running it yourself, because of the time, the place, the device and the recent history of the person searching.
officialSource : Google Search Console, “Performance report (Search results)”
That sentence, published by the tool itself, disqualifies the trade’s most widespread reflex: open a browser, type the query, conclude. What you see is one draw among many, influenced by parameters you do not control and are not aware of.
Deciding from real impressions rather than from a manual check
hypothesis
- Effort
- low
- Time to effect
- weeks
- Risk
- none
- Durability
- structural
What the tool says about its own limits
It is rare and it is worth reading, because it stops you attributing to your site gaps that come from the measurement.
The crawl stats report lists most crawl requests, but some may not be counted for various reasons.
officialSource : Google Search Console, “Crawl Stats report”
The addresses given as examples in that report are not exhaustive: they are representative examples.
officialSource : Google Search Console, “Crawl Stats report”
Two practical consequences. First, a gap between your server logs and what the tool reports is not necessarily an anomaly to explain. Second, a list of example addresses is not an inventory: treating the few addresses shown as the whole set of affected cases leads to declaring a problem solved when only a sample has been fixed.
Which conclusions must you refuse to draw?
Small position changes happen at any time, upwards included, with no action needed.
officialSource : Google Search Central, “Debugging drops in Google Search traffic”
A three-day drop is not a trend. The background noise is documented, and it includes spontaneous recoveries. Comparing one week with the previous one on a single page mostly produces randomness put into a chart.
A variation is not a cause. The tool reports what happened, never why. The temptation to attribute every movement to the last thing you changed is nearly irresistible, and it is the best way of learning false things confidently.
Absent data is not absent display. Very low-volume queries do not all come through, which makes hunting for a particular query you are sure exists a waste of time.
The reading that stays possible is slower and far more reliable: over several weeks, identify the queries with high impressions and few clicks — those are badly phrased promises — then the pages losing impressions lastingly, which signal a subject changing hands.
What does not work, or no longer works
Your average position went from 12.4 to 8.7 this week: your optimisation is working, keep publishing at the same pace.
One use is worth dropping outright: the shared dashboard displaying a position curve without its impression volume. It makes the only honest reading of that metric impossible, and it makes the flattering conclusion inevitable. A measurement you cannot read wrongly beats a measurement you have to remember to correct in your head.
Sources
- Performance report (Search results): Overview and basic setup — Google Search Consoleaccessed on
- Crawl Stats report — Google Search Consoleaccessed on
- Debugging drops in Google Search traffic — Google Search Centralaccessed on
- SEO Starter Guide: The Basics — Google Search Centralaccessed on
Block revalidated on
Frequently asked questions
- Why does average position not match what I see?
- Because it aggregates all your real impressions, in all their contexts, whereas your personal search produces only one. The documentation explicitly warns that a query present in your list may not show your site when you run it yourself.
- Is a rising average position good news?
- Not necessarily. It can rise because you lost impressions on queries where you were badly placed, which improves the average by reducing exposure. So it always has to be read with the impression volume beside it.
- Should you check Search Console every day?
- No, and the habit degrades the quality of the decisions taken. Daily variation comes partly from fluctuations the official documentation describes as spontaneous, upwards included. A weekly or monthly reading, driven by a precise question, teaches far more.
- Are the Search Console figures accurate?
- They describe events that actually happened, which sets them clearly apart from third-party estimates. They do carry documented counting limits, though: the crawl report itself states that some requests may not be counted, for various reasons.