What 2.9 Million Search Console Queries Reveal About Customer Intent
Google Search Console is usually treated as a keyword report. You open it to check clicks, impressions, rankings and click-through rate (CTR), then try to work out what is performing and what needs work. That is useful, but it misses one of the most valuable parts of the data.
Search Console does not just show what people searched for. It gives us clues about what people were trying to do when they searched. Some queries show that a person already knows the brand. Some show they are comparing options. Others suggest they are pricing something up, solving a problem, researching a topic, or looking for a local supplier. The language of the query itself reveals intent.
To test how far that holds, I analysed query-level exports from 123 websites: 2,915,076 query rows, covering 2,751,996 unique queries, 3,416,004 clicks and 310,205,480 impressions, across 16 months of exports (the maximum GSC allows).
| Data point | Total |
| Websites analysed | 123 |
| Query rows analysed | 2,915,076 |
| Unique queries | 2,751,996 |
| Total clicks | 3,416,004 |
| Total impressions | 310,205,480 |
This is an intent study, not an attribution study. Search Console cannot tell you whether a query became a lead or a sale. What it can tell you is how people use language before they click, enquire, compare or buy, and that turns out to be more useful than a ranking list.
How I Ran the Study, and Where It Stops
I grouped queries into broad intent categories based on the language used and the modifiers attached to them: branded, non-branded, commercial, cost and price-led, comparison, problem-led, local, question-based, and broad or unclear. The grouping keyed off common modifiers such as how, what, best, cost, price, near me, vs, review and calculator, because those words usually signal what stage of the decision someone is at.
Two things are worth being upfront about, because they change how you should read the numbers.
The categories overlap and are not mutually exclusive. A query like how much does X cost near me is a question, a cost query and a local query at once. So the intent groups below do not sum to the total, and they are not meant to. Branded and non-branded is the only clean, exhaustive split in the dataset; everything else is a modifier-based lens applied on top of the non-branded pool.
Classification is modifier-based, so it under-counts fuzzy intent. Matching on keyword patterns is fast and consistent, but it only catches the language it is told to look for. The problem-led group is the clearest example: at 7,560 queries it is almost certainly an undercount, because real problem searches use words a tidy match list misses (stopped, won’t, error, help, and countless product-specific phrasings). Read the smaller categories as directional, not precise.
One more caveat on scale. Blending 123 sites across different sectors hides a lot of variance. A local trades business and a national ecommerce brand have very different branded CTRs and very different intent mixes. The headline figures below are directionally sound, but no single site should expect to match them.
What the Intent Groups Showed
| Intent type | Unique queries | Clicks | Impressions | CTR | What it usually shows |
| Branded | 14,287 | 671,972 | 3,141,677 | 21.39% | People already know the business |
| Non-branded | 2,738,484 | 2,744,032 | 307,063,803 | 0.89% | Searching without naming a brand |
| Commercial | 96,657 | 42,148 | 9,673,594 | 0.44% | Looking for a product, service or supplier |
| Cost and price-led | 65,460 | 24,496 | 3,351,704 | 0.73% | Thinking about budget or affordability |
| Comparison | 90,827 | 44,599 | 6,383,575 | 0.70% | Weighing up options or alternatives |
| Problem-led | 7,560 | 1,215 | 293,358 | 0.41% | A specific issue to solve |
| Local | 111,157 | 85,909 | 9,586,110 | 0.90% | Looking for something in an area |
| Question-based | 248,717 | 71,354 | 16,416,231 | 0.43% | Researching or trying to understand |
| Broad or unclear | 2,013,899 | 2,403,319 | 251,798,516 | 0.95% | Intent too vague or broad to judge |
The largest group by far was broad or unclear. That is not a failure of the data; it is the point. Most searches do not fit neatly into one box. They are too broad, lack context, or could mean different things depending on the page, sector or result shown. Search Console needs interpretation. Exporting a query list and treating every impression as an opportunity is how reporting goes wrong.
The other clear message is that intent cannot be read from impressions alone. Broad or unclear searches produced 251 million impressions and the lowest-value intent. Local searches produced a tenth of that and a stronger CTR, because someone searching in a specific area is usually further along. High visibility is not the same as high value.
Branded Search Is Valuable, but It Is Not a Guaranteed Click
The sharpest finding was the gap between branded and non-branded behaviour.
Branded queries generated 671,972 clicks from 3,141,677 impressions, a 21.39% CTR. Non-branded queries generated 2,744,032 clicks from 307,063,803 impressions, a 0.89% CTR. No surprise in the direction: someone searching your name already has a reason to click.
The number worth sitting with is the other side of that CTR. Even with branded intent, roughly 78.6% of branded impressions did not lead to an organic click. That does not mean branded search is low value. It means the click went somewhere Search Console cannot see: a paid ad, a Google Business Profile, a social profile, a review site, a directory, a marketplace, or the answer sat in the results and no click was needed.
This is the part that matters for reporting. Branded organic traffic should not be booked as an SEO-only win. Branded search is usually where existing demand becomes visible, not where that demand was created. The demand behind a branded search may have come from PPC, social, email, PR, word of mouth or a previous customer experience. SEO, paid and brand activity do not work in isolation, and branded search is the clearest place that shows up.
Non-Branded Search Is the Wider Opportunity, if You Filter It by Intent
Non-branded queries made up over 98% of all impressions. These are people who may not know the business yet: comparing providers, researching a problem, checking prices, reading reviews, or working out their options. It is also far harder to win. A branded searcher arrives with a reason to click; a non-branded searcher has to be earned against every other result on the page, which is why title, meta description, content angle, page type and trust signals all pull weight.
The low blended 0.89% CTR is partly an artefact worth naming: enormous broad, high-impression queries sitting at low positions drag the average down. It reads like underperformance, but a lot of it is just visibility for searches the site was never well placed to win.
So the useful question is not what are we ranking for but which non-branded searches show real intent. A query carrying cost, near me, best, vs, review, repair or how to fix tells you far more than a bare keyword. Those modifiers separate someone browsing from someone comparing, troubleshooting or close to buying, and that should drive what page you build.
The Long Tail Shows What Customers Are Unsure About
Question-based searches accounted for 248,717 unique queries and 16,416,231 impressions: how, what, why, when, can, should, does. Individually these look like noise, a few clicks each, easy to skip in a keyword report. Read together, they show the concerns, blockers and decision points customers hit before they act.
The trap is turning every question into its own article, which produces thin, repetitive content. Group them into themes instead. Several cost questions point to one clear pricing page you can link to. Repeated is X better than Y queries point to comparison content. A cluster of problem searches points to troubleshooting advice that connects naturally to the relevant service.
The long tail is also where the customer’s real language lives: the phrases people use before they learn the technical terms or brand names. That wording is useful well beyond SEO, feeding landing pages, paid search copy, sales conversations and email.
How Intent Should Map to the Page You Build
This is the practical core of it. Before creating or reworking a page, ask what stage of the journey the query represents, then match the page type to it:
| The query signals | The page it usually needs |
| Informational (how, what, why) | A helpful guide or FAQ |
| Commercial (product, service, supplier) | A strong service or category page |
| Comparison (vs, best, review) | A balanced explanation of the options |
| Cost or price (cost, price, quote) | Clearer pricing guidance in one place |
| Local (near me, a place name) | Stronger location signals |
| Problem-led (broken, fix, not working) | Practical advice that links to the service |
Volume alone should not decide priority. A high-impression query with broad intent can be worth less than a low-volume query that shows a clear local need or a specific problem. Chasing impressions that look good in a report but carry no intent is the most common way SEO effort gets wasted.
What Search Console Cannot Tell You on Its Own
Search Console shows the queries used, how often the site appeared, organic clicks, CTR and average position. That is a strong view of search behaviour, but it is one channel and one moment in a journey that is rarely linear.
It cannot tell you whether someone saw a paid ad first, arrived via social, clicked an email, read a review, asked for a recommendation, or saw the brand offline. It cannot tell you whether an organic visitor became a lead or a sale unless you connect it to other tools. A branded search can look like an SEO result while the demand behind it was built by something else entirely.
None of that makes the data less useful. It means the commercial meaning has to be read in context, alongside GA4, CRM data, call tracking, form submissions, paid search and social. What Search Console gives you that nothing else does is a direct view of the language people use as they move through the search journey.
How to Read Your Own Search Console Data
You can run this same intent lens on your own site. You do not need 123 websites or a database, just one clean export and a spreadsheet. Here is the process I use.
- Export at query level. In Search Console, open the Performance report, set the widest date range it allows (16 months), and export the Queries tab with clicks, impressions, CTR and average position all included. Pull position too; you will need it later.
- Split branded from non-branded first. This is the only clean, exhaustive split, and it is the most important one to get right. Build a list of your brand terms, including common misspellings, spacing variants and any sub-brand or product names people use as shorthand. Flag every query containing one of those as branded, and treat everything else as non-branded. From here on, compare the two groups separately, because their CTRs are not comparable. Expect branded to run many times higher; in my data it was 21.39% against 0.89%.
- Apply modifier buckets over the non-branded pool. This is where intent shows up. Tag each query by the words attached to it, using match lists like the ones below. These buckets overlap on purpose, so a single query can sit in more than one, and they are not meant to add up to your total.
| Intent bucket | Words and patterns to match on |
| Cost and price | cost, price, prices, pricing, fee, fees, quote, cheap, cheapest, how much, per month |
| Comparison | vs, versus, best, top, review, reviews, compare, alternative, alternatives |
| Local | near me, local, “in [place]”, plus your town, city and region names |
| Problem-led | broken, not working, won’t, stopped, error, fault, faulty, fix, repair, issue, problem, troubleshoot, help |
| Question and research | how, what, why, when, can, should, does, do (as the first word) |
| Commercial | buy, hire, service, services, supplier, provider, installer, company, for sale, plus your core product and service nouns |
Note the problem-led list is deliberately broad. A narrow match list is exactly how that group gets undercounted, so err towards catching too much and refine afterwards.
- Read CTR within intent, against your own baseline. Do not judge every query against one target. A 0.7% CTR on a broad research question is fine; the same figure on a cost or local query sitting at a decent position probably is not. Compare each query to what is reasonable for its intent and its average position, which is why you exported position. The query worth investigating is the one with high impressions, a page-one position and a weak CTR, because that usually points to a title or intent mismatch rather than a ranking problem.
- Follow the intent, not the volume. Sort by intent value rather than impressions. A handful of cost, near me or specific-problem queries can matter more than a broad term with fifty times the visibility. Keep an eye on the broad, high-impression, low-position queries that drag your averages down and make the whole account look like it is failing, when it is really just visible for searches it was never placed to win.
- Turn the intent into pages. Use the mapping earlier in this piece. Once each meaningful query has an intent attached, the action follows: research questions become guides or FAQs, cost queries need clearer pricing in one place, comparison queries need a balanced options page, local queries need stronger location signals, and problem queries need practical advice that links to the service. Group similar queries into one strong page rather than one article each, or you will end up with the thin, repetitive content nobody links to.
Do this once and Search Console stops being a ranking report. It becomes a map of what your customers are trying to do before they ever reach you.
The Takeaway
Search Console is at its weakest as a list of keywords to chase and at its strongest as a read on intent. Branded search had a much stronger CTR but still did not guarantee a click. Non-branded search held most of the impressions but plenty of it was broad, competitive or unclear. Not every impression is an opportunity, not every query has commercial value, and not every branded search belongs to SEO.
Once you understand the intent behind the search, you can build content that meets people at the point they are most likely to convert, and report on the work honestly rather than celebrating visibility for its own sake.
