Search intent is the reason behind a query — whether the person wants to learn something, compare options, buy, or reach a specific site. Getting it wrong is the most expensive avoidable mistake in content planning, because a page built for the wrong intent will not rank no matter how well it is written. A buying guide will not win a query where every result is a product page, and a product page will not win a query where every result explains a concept.
This classifier sorts a pasted keyword list into four intent buckets using modifier patterns, and then — the part that matters — shows you which pattern fired for each keyword. Arb Digital built it that way deliberately. Plenty of tools return an intent label with no explanation, which leaves you either trusting it blindly or ignoring it. Seeing that "best solar panel installers" was labelled commercial because it matched the modifier best lets you accept the call in a second, or overrule it just as fast.
What This Keyword Intent Classifier Does
Paste keywords, one per line, and each is tested against four ordered rule sets. Navigational patterns are checked first — brand names you supply, plus login, download, portal, official site and similar — because a branded query is navigational regardless of what else it contains. Transactional patterns come next: buy, order, price, quote, near me, for sale, discount, coupon. Then commercial: best, top, review, vs, compare, alternative. Anything left that matches question words or learning modifiers is informational, and anything matching nothing at all is reported as unmatched rather than being silently assigned.
The output gives you counts for each bucket, the dominant intent in the list, a proportional bar breakdown, and an audit table listing keyword, assigned intent, and the literal pattern that produced it. That last column is the whole point: it makes the classification arguable.
How to Use It
- Paste a real keyword list, not a curated one. The value is in seeing the actual balance of your target set, including the terms you would rather not think about.
- Add your brand and competitor names. Without this, branded queries scatter across the other three buckets and the informational count becomes meaningless.
- Read the audit table, not just the totals. Scan for calls that look wrong and note the pattern that caused them.
- Overrule the tool where the SERP disagrees. For any keyword that matters commercially, open the results page. What actually ranks is the definitive answer on intent; a modifier list is only a shortcut. Search Console's performance report is a useful cross-check, since it shows the queries already reaching each page.
- Group the confirmed buckets into page types before writing anything — guides for informational, comparison pages for commercial, product or service pages for transactional.
The Formula / How It's Calculated
There is no formula in the mathematical sense; this is ordered pattern matching, and the order is the interesting part. Each keyword is lowercased, then tested against the four rule sets in a fixed priority: navigational, then transactional, then commercial, then informational. The first set to match wins, and the matching token is recorded and displayed.
Priority is what stops absurd results. "Buy sunpower panels" contains a brand name and a transactional verb. Classified purely by brand it is navigational, which would be wrong — the person is buying, they have just already chosen a manufacturer. So the tool checks whether a transactional pattern also matched a branded query and promotes it to transactional when it does. That single exception handles a large share of real-world branded commercial demand.
"Solar panel reviews 2026" matches both review (commercial) and nothing transactional, so it lands in commercial. "Solar panel cost calculator" matches cost, which sits in the transactional set here, and reasonable people disagree about that — cost queries often want an explanation rather than a checkout. The tool shows you it fired on cost so you can move it yourself. And "solar panels" on its own matches nothing, so it is reported as unmatched: a bare head term whose intent is genuinely mixed and can only be settled by looking at the results page.
Why Modifier Matching Is Only Ever a First Pass
Modifiers are a proxy for intent, not intent itself. Three failure modes come up constantly, and knowing them makes the tool far more useful.
Head terms carry no modifier at all. "Running shoes" could want a category page, a buying guide, or an explanation of shoe types. The word alone contains no signal. These are exactly the keywords where a modifier tool is least useful and where checking the SERP is most necessary.
Intent drifts over time. A query that returned informational articles two years ago may now return product listings, because the market matured and the searcher population changed. Google's guidance on creating helpful content frames this in terms of serving the person rather than the pattern, and the person changes.
The same modifier means different things by industry. "Free" in software is a product tier and a strongly transactional signal. "Free" in a legal context often prefixes "free consultation", which is a lead-generation query. In education it frequently just means a free resource to read. One word list cannot resolve that, which is why every call in this tool is shown with its evidence.
What Each Bucket Should Actually Be Used For
Informational keywords justify guides, explainers, and definitions. They convert poorly on the visit and they build the topical coverage that supports everything else, so judge them on assisted value rather than direct conversion. If you are measuring whether your coverage of a topic is complete, the topical authority calculator is the natural next step.
Commercial keywords — best, review, versus, alternative — are the highest-leverage bucket for most businesses and the most consistently under-served. The searcher has a problem, knows solutions exist, and is choosing. Comparison pages, honest alternatives pages, and category buying guides live here.
Transactional keywords want a page where the action can be completed: product page, service page with pricing, quote form, booking. Sending this traffic to a blog post is the single most common intent mismatch, and it shows up as a page with strong rankings and no conversions.
Navigational keywords are people looking for a specific brand. Your own branded terms should be defended and measured separately — see the brand search share calculator — and competitor branded terms are usually best approached through comparison content rather than by trying to outrank a company for its own name.
Where Intent and Cannibalisation Meet
Once a list is classified, a specific failure becomes easy to spot: two of your own pages targeting the same keyword from different intents. A blog post explaining what a thing costs and a service page selling it will both chase "solar panel cost", and Google will pick one, often inconsistently, often the wrong one. Neither page performs as well as a single page built for the intent the results page actually rewards.
The fix is usually consolidation rather than differentiation — one page that resolves the query, with the secondary angle folded in as a section. Run the keyword cannibalization checker across the URLs that came out of the same bucket, because the classifier tells you what a keyword wants and the cannibalisation check tells you whether you have already answered it twice.
The reverse case is worth watching too. If a whole bucket comes back with no matching pages at all, you have found a structural gap rather than an optimisation problem, and the content gap calculator is the right tool for sizing it.
Boundary: Classifying Versus Grouping
This tool answers "what does the searcher want?" It does not group semantically related keywords into topic clusters — that is a different job, done by the keyword cluster generator, which finds which keywords belong on the same page. Nor does it estimate how hard a keyword is to rank for; the keyword difficulty checker covers that side.
Used in sequence, the three answer the three questions that decide a content plan: what does this searcher want, which keywords belong together on one page, and can we realistically win it. Intent first is the right order, because a difficulty score for a keyword you would answer with the wrong page type is a wasted calculation.
Arb Digital's SEO team maps keywords to intent and page type before a word is written, so transactional demand lands on pages where the visitor can actually act.
SEO Services Content Marketing ServicesCommon Mistakes to Avoid
- Trusting a modifier label on a head term that contains no modifier at all — those need a look at the actual results page.
- Sending transactional queries to blog posts, producing pages that rank respectably and convert almost nobody.
- Skipping the brand list, which scatters navigational queries across the other three buckets and inflates the informational count.
- Treating intent as permanent — results pages shift, and a query that wanted an article last year may want a product listing now.
- Applying one modifier dictionary across every industry, where words like "free", "cheap" and "template" carry genuinely different meanings.
Related Free Tools From Arb Digital
Group the classified terms with the keyword cluster generator, size the effort with the keyword difficulty checker, and check what the results page looks like before committing using the SERP feature opportunity calculator. Once pages exist, the SEO traffic forecast calculator turns rankings into expected clicks. The free online tools hub has the rest.
Frequently Asked Questions
Informational, commercial, transactional and navigational. Informational wants to learn, commercial is comparing options before deciding, transactional wants to complete an action, and navigational is looking for a specific site or brand. The boundary between commercial and transactional is the fuzziest of the three.
Good enough for a first pass on a large list, and unreliable on head terms with no modifiers. That is why this tool displays the pattern that fired for each keyword rather than just a label — it lets you check the reasoning instead of trusting the output. For commercially important keywords, the results page is the authority.
Google does not publish an intent taxonomy or an API that labels queries. The four-category model used here is a long-standing industry convention, not an official classification, and different practitioners split the categories differently.
Because it contained none of the modifier patterns — usually a bare head term like "running shoes" or "accountant". Unmatched is an honest answer rather than a failure: those queries have genuinely ambiguous intent and can only be settled by looking at what currently ranks.
Usually yes, because a page trying to explain a concept and sell a product at once tends to do neither convincingly. The exception is a strong commercial page that includes a short explanatory section for readers who need it, which serves the primary intent while covering the secondary one.
They sit on the boundary. Some want a checkout, many want an explanation of what drives the price. This tool treats them as transactional and tells you it did, so you can move them after checking whether the ranking pages are pricing pages or guides.
No. The classification runs in your browser using plain JavaScript. Nothing is uploaded, stored, or transmitted, which also means the list disappears when you close the page.
Intent labels produced here are heuristic estimates based on modifier patterns, not an official classification, and should be confirmed against live search results for any keyword that matters commercially.
