In this article
The fastest way to find SEO keywords that convert is to stop starting in a keyword tool. Open your analytics, list the five pages that already generate revenue, and reverse-engineer the language those buyers used. Volume is a vanity metric. A term with 90 searches a month and a credit card behind it beats a 40,000-volume head term every single time — and it usually ranks in a quarter of the effort.
Here's the process I use with clients, from first-party data through SERP validation to a scoring model you can run in a spreadsheet.
Start With the Pages That Already Make Money
Before touching a research platform, pull a report of your top converting landing pages from the last 12 months. In GA4 that's Reports > Engagement > Landing page, with a key event applied as the secondary metric. Sort by conversions, not sessions. The picture is often uncomfortable — one B2B SaaS client discovered that 61% of their demo requests came from three comparison pages they'd never optimised, while the blog they'd spent two years building drove almost none.
Those three pages are your seed list.
Now cross-reference them in Google Search Console. Filter by page, look at every query bringing impressions, and note which ones sit in positions 4-15. Those are queries Google already associates with a page that converts. Pushing a position-9 term to position 3 is cheaper than creating anything new, and the conversion behaviour is already proven.
Pay attention to the shape of the winning queries. Are they problem-led ("invoice reconciliation is slow")? Solution-led ("automated reconciliation software")? Brand-comparison ("Xero vs QuickBooks")? That shape tells you which family of terms to expand into. If your converters are all comparison queries, chasing top-of-funnel definitions will waste six months of production budget.
One practical note: exclude branded queries from this analysis or they'll swamp everything. People searching your brand name convert at 5-10x the rate of non-brand traffic, and including them makes every other insight invisible.
What Makes a Keyword Actually Convert?
A converting keyword signals that the searcher has a budget, a deadline, or a decision to make. It usually contains a solution category, a comparison, a price signal, or a location. Informational queries build trust but rarely close. The reliable markers are commercial modifiers, plural product nouns, and a SERP filled with pricing pages rather than definitions.
Beyond that 60-second answer, intent lives on a spectrum. I sort every candidate into four buckets: transactional ("buy", "pricing", "free trial"), commercial investigation ("best", "vs", "alternatives", "review"), problem-aware ("how to fix", "why is my"), and purely informational ("what is").
Commercial investigation is the sweet spot for most content teams. Transactional terms are usually dominated by your own product pages and by Amazon or aggregators. Problem-aware terms convert surprisingly well when the problem is expensive — "payroll tax penalty notice" converts better for an accounting firm than "best payroll software", because the searcher is panicking and has a deadline.
Job titles are another underrated signal. "CRM for real estate agents" converts at multiples of "CRM software" because the qualifier tells you the searcher has self-identified as a buyer in a specific segment. Anything that narrows the audience narrows the intent too.
What nobody tells you: intent shifts over time. "AI writing tool" was informational in 2022 and is now overwhelmingly transactional. Re-check your assumptions on core terms once a year, because the SERP quietly reclassifies them and your page stops matching.
Mine Your First-Party Data Before Any Tool
Your CRM contains keyword research that no subscription can sell you. Export the free-text fields — "how did you hear about us", the notes your sales reps type after discovery calls, the subject lines of inbound emails. Buyers describe their problems in language that rarely matches internal product vocabulary, and that gap is where uncontested search volume hides.
Three sources I check on every engagement:
- Internal site search — set up as a GA4 event, this shows exactly what visitors couldn't find. High-exit search terms are content gaps with proven demand.
- Support tickets — filter for tickets raised in the first 14 days after purchase. Those questions are the objections that almost stopped the sale.
- Sales call recordings — Gong, Fireflies, or even Zoom transcripts. Search for the phrase "we were looking for" and read the next twenty words.
An ecommerce client selling workwear kept seeing "flame retardant" in tickets while their whole catalogue said "FR rated". One terminology change across category pages and meta titles produced a 34% lift in non-brand clicks to those pages within two months. Nothing else changed.
Reddit and niche forums serve the same purpose when you're launching into a market with no data of your own. Search site:reddit.com "your category" recommendations and read the threads properly. The phrasing people use when asking strangers for advice is the phrasing they later type into Google. Our complete guide to keyword research goes deeper on structuring these raw inputs into a usable map.
Stack Modifiers to Expand Commercial Terms Fast
Once you have seed terms, modifier stacking is the quickest mechanical way to find SEO keywords with buying intent. Put your core noun in column A of a spreadsheet, then combine it with four separate modifier lists to generate every possible phrase.
List one, commercial: best, top, cheap, affordable, premium, professional. List two, comparison: vs, alternative to, versus, compared, better than. List three, qualifier: for small business, for beginners, for freelancers, for enterprise, near me. List four, stage: pricing, cost, review, demo, trial, template, checklist.
A 10-term seed list against 24 modifiers gives you 240 candidates in about four minutes. Paste them into Keyword Planner or Ahrefs Keyword Generator in batches, keep anything with real volume, and bin the rest.
Cost-per-click is your shortcut for scoring what's left. Advertisers bid on terms that convert; that's the whole business model. Any keyword carrying a CPC above your category average has been validated with someone else's ad spend. In legal or insurance verticals you'll see $50+ clicks, which tells you exactly where the money sits even if monthly volume looks tiny.
Watch for a trap here. Tools round low-volume terms down to zero or lump them into "10-100" buckets, so genuinely valuable long-tail phrases look dead. I've seen a "0 volume" term drive 40 qualified leads a year. If the phrase makes commercial sense and a competitor has a page for it, build the page. Modern SEO tools for keyword research handle the clustering, but judgement still beats the export.
Read the SERP Before You Commit a Single Word
The results page is the only opinion that matters. Google has already tested thousands of page types against each query, and what ranks is what satisfies. If page one for your target term is nine listicles and a Reddit thread, publishing a product page there is throwing budget away.
Open an incognito window and check five things.
- Page type — commercial pages, blog posts, forums, or video? Match it or don't compete.
- Ads — four shopping ads plus three text ads means proven commercial value.
- SERP features — an AI Overview or a featured snippet at the top can absorb the clicks. Informational terms are losing more clicks than commercial ones here.
- Domain profile — if every result is a DR 85 publisher and you're a DR 30 site, park the term.
- Freshness — dates from 2019 on page one signal a neglected SERP you can take.
Ahrefs' widely cited study found that over 90% of pages get no organic traffic from Google at all, and mismatched intent is a big part of why. People write the page they want to write instead of the page the SERP demands.
Here's the gotcha that costs teams months: a term can look perfect on volume, difficulty and CPC, and still be unwinnable because the SERP is locked to marketplaces. "Buy running shoes" is Amazon, Nike and Foot Locker. No blog post is displacing them. Check first, write second.
How Do You Find Keywords Competitors Are Converting On?
Look at where they spend money, not where they rank. Pull a competitor's paid keyword report in Semrush or Ahrefs and filter for terms they've bid on continuously for six months or more. Sustained ad spend on a specific phrase is the clearest public evidence that the keyword produces revenue for a business like yours.
That single filter — duration of bidding — separates real signals from tests. Anyone can run a two-week experiment. Nobody funds a losing keyword for half a year.
Next, examine which pages they've built dedicated landing pages for. If a competitor has a standalone page for "expense management software for nonprofits", someone in that company did the maths on that segment. Their internal linking tells you more: pages receiving the most internal links are the ones the business considers most valuable.
Check their sitemap directly too. Add /sitemap.xml to their domain and scan the URL slugs. You'll often spot a whole templated directory — comparison pages, integration pages, location pages — that never surfaces in a standard keyword export because individual pages get low traffic. Collectively they can drive a third of a company's pipeline.
Content gap analysis closes the loop. Feed three competitor domains and your own into a gap report, filter to terms where at least two of them rank in the top 10 and you rank nowhere, then sort by CPC rather than volume. That sorted list is your production queue. An AI-assisted keyword research workflow can cluster those gaps by intent in minutes instead of an afternoon.
Score, Prioritise and Validate With Real Money
Volume-based prioritisation is why so many content calendars underperform. Replace it with a conversion-weighted score. I use four inputs, each rated 1-5, then multiply rather than add so a zero-value dimension kills the term outright.
Intent fit: does the searcher need what we sell, today? Winnability: can we realistically reach the top five given our authority? Business value: does this map to a high-margin product or just a lead magnet? Effort: inverse score — a page needing original research scores lower than one we can produce in a day.
Multiply the four and you get a priority number between 1 and 625. Anything above 200 goes into the next sprint. It sounds crude. It stops arguments in planning meetings, which is the actual point.
Then validate the top ten with paid search before writing. Run a £300-£500 exact-match campaign for two weeks pointing at your closest existing page. You'll learn the real conversion rate, the real cost per lead, and whether the traffic is qualified — for less than the cost of one commissioned article. I've killed keywords that scored 400 on the model because the paid test produced twelve clicks and zero form fills.
Revisit the scores quarterly. Rankings shift, competitors publish, and your own authority grows, which moves the winnability input. Pair this with a routine technical review using solid SEO audit tools so a crawl issue isn't quietly suppressing a page you already invested in.
Converting keyword research isn't about finding more terms. It's about finding fewer, better ones and proving they're worth the production cost before you spend it. Start with your own converting pages, mine the language your buyers already use, sanity-check every candidate against the live SERP, and validate the shortlist with a small ad budget. Do that for one quarter and your content calendar will look nothing like it did before — shorter, sharper, and tied to revenue rather than impressions.
Frequently Asked Questions
How many keywords should one page target?
One primary keyword and three to eight close variants that share the same intent. Google understands synonyms well, so a page about "project management software for agencies" naturally ranks for "agency project management tools" too. Splitting near-identical terms across separate pages causes cannibalisation, where neither page ranks properly. Cluster first, then decide how many URLs you genuinely need.
Is search volume still worth checking in 2026?
Yes, but treat it as a rough range rather than a forecast. Keyword Planner buckets and third-party estimates both drift by 30% or more, and zero-click behaviour means impressions no longer translate cleanly into visits. Use volume to separate "a market exists" from "nobody searches this", then rely on CPC, SERP composition and your own conversion data for the real prioritisation.
How long before a commercial keyword starts converting?
On a site with existing authority, long-tail commercial pages often produce leads within six to ten weeks. Competitive head terms typically take six to twelve months and need links. The honest answer depends on your domain strength — check where your existing pages rank for similar-difficulty terms and use that as your baseline instead of a generic industry timeline.
Should I outsource keyword research or do it in-house?
Keep it in-house if you have direct access to sales calls, CRM data and support tickets, because that context is where converting terms actually come from. Outsource when you're entering an unfamiliar market or lack tool subscriptions. A hybrid works best: an agency builds the initial map, your team refines it quarterly using first-party conversion data.