Every site ranks for things nobody planned. A paragraph buried in a two-year-old post happens to answer a question, and Google starts showing it. Those accidental appearances are the most honest content research available to you, because the demand is proven and the audience is already yours.
The report is free, it is in Search Console, and it is usually opened to check whether the line went up.
What are you actually looking for?
A mismatch between a query and the page answering it. Three patterns matter, and each implies a different action.
| Pattern | What it means | Action |
|---|---|---|
| Impressions, position 8-20, page barely covers it | Demand exists, your answer is thin | Write the dedicated page |
| Impressions on a page about something else | Google is guessing on your behalf | Write the right page, then link to it |
| Good position, poor click-through rate | The listing is losing the click | Rewrite title and description |
The second row is the valuable one. When a page ranks for a question it does not really answer, Google has told you there is demand and nothing good serving it — including from your competitors, or a better page would be there instead.
An accidental ranking is a keyword tool that has already been validated against your own audience. No volume estimate can tell you what a real impression on your own domain tells you.
How do you pull the list?
Open the performance report, set the date range to the last twelve months, and export queries. Then work through it in this order, because the early filters remove most of the noise.
- Drop branded queries. They are demand you already own and they will dominate the top of any sort.
- Filter to positions between 6 and 25. Above that you are already winning; below it the query is usually incidental.
- Sort by impressions. Volume here is measured on your site rather than modelled.
- For each of the top rows, check which page is ranking. That is the column people forget to look at.
- Ask whether that page genuinely answers the query. Read it as a stranger would.
Twenty rows is enough for a quarter of content work. The list is also self-updating, which is what makes it better than a one-off keyword project.
Why does this beat a keyword tool?
Because third-party volume estimates are modelled and this is counted. A tool can tell you a phrase gets an estimated 1,300 searches a month somewhere in the country. Search Console tells you that your page appeared for it 400 times, at position 11, and got four clicks.
It also surfaces phrasings no tool will show you. Google has said a meaningful share of daily queries have never been seen before, and those cannot appear in any keyword database. They can appear in your report, because someone searched them and you showed up.
Writing a new page for every interesting query found. Half of these belong as a section in an existing page. Splitting them out creates the internal competition that consolidating pages exists to fix.
When is it a section rather than a page?
When a reader asking the query would be satisfied by the existing page plus two paragraphs. If the query is a sub-question of what the page already covers, it is a heading; if it is a different job with a different outcome, it is a page.
The practical version: could you write a heading in the reader's own words, answer it in the first forty words beneath, and have that sit naturally where it is? Then it is a section, and adding it is an hour of work rather than a week.
What about queries you cannot serve?
Leave them. Some accidental rankings are for questions your business genuinely has nothing to do with, and chasing them produces content with no commercial connection to anything you sell. Impressions are not the goal; they are only the evidence.
A reasonable filter: would a person asking this ever become a customer, even indirectly? If not, the impressions are noise, however many there are.
Make it a habit rather than a project
Once a month, twenty minutes, same filters. The value compounds because each new page starts ranking for its own set of unplanned queries, which feeds the next month's list.
Open the report and set the filters now. The first ten rows will name your next two articles, and you will not have guessed at either of them.
Sitting on a year of query data nobody has mined?
We will turn your accidental rankings into a prioritised content plan built on demand you have already proven.
Get in Touch →Does this apply to your site?
Reading about it is one thing. Point the scan at your own site and see whether this applies to you, and what it is worth fixing.
Free and unlimited. No account, no card, and you get every finding rather than a teaser.