Keyword research is deciding which searches you want to win before you spend effort trying to win them. Skip it and you end up with the classic small business site: pages competing for terms no one searches, or terms dominated by national brands you cannot outrank. This is a practical method that works without expensive tools, sized for a small business with limited hours.
The principle: winnable and valuable
Every keyword sits somewhere on two axes: how valuable the searcher is to you, and how realistic it is that you can rank. "Insurance" is valuable and unwinnable. "What color is a stop sign" is winnable and worthless. Your targets live in the overlap: specific enough to win, commercial enough to matter. For small businesses, that overlap is bigger than you think, because most competitors never look for it.
Step 1: harvest candidates from real sources
Start with language customers actually use, not what you would search:
- Your inbox and call log. The questions people ask before buying are searches, phrased exactly how your market phrases them.
- Google's own suggestions. Type your service and let autocomplete finish it. Check the "People also ask" boxes and related searches at the bottom of results.
- Search Console, if your site has history. The queries report shows terms you already appear for, including ones you never targeted. Terms ranking at positions 8 through 20 are pre-qualified wins. See our Search Console guide for where to look.
- Competitor sites. The page titles of whoever outranks you are their keyword map, published in public.
Collect fifty to a hundred candidates in a spreadsheet. Do not judge yet.
Step 2: sort by intent
Intent, what the searcher wants to happen next, matters more than volume. Sort your list into three buckets:
- Ready to buy: "emergency electrician near me", "water heater replacement cost [town]". These map to service and location pages.
- Researching a decision: "tankless vs tank water heater", "how much does rewiring cost". These map to guides and comparison posts.
- Learning: "why does my breaker keep tripping". These map to helpful articles that build authority and catch future customers early.
A healthy plan targets all three, weighted toward buying intent early, because those pages pay first.
Step 3: judge winnability with your eyes
You do not need a difficulty score from a paid tool; the results page tells you directly. Search each shortlisted term and look at who ranks:
- If page one is national brands and household names, pass for now.
- If it is other local businesses with mediocre pages, that term is winnable.
- If it is directories and aggregator sites, a genuinely good specific page often beats them.
- If Google shows a map pack, local optimization on your Business Profile matters as much as the page.
Step 4: pick and assign, one term per page
Choose a primary keyword per page you plan to create or optimize, and a few natural variations to weave in. The rule that prevents most self-inflicted problems: no two pages targeting the same primary term. Two of your own pages competing for one query split your strength and usually both lose. Keep the map in a simple spreadsheet: page, primary term, supporting terms, status.
Step 5: schedule by payback
Order the work by how fast it returns: existing pages that already rank on page two first (optimize using our on-page checklist), missing service and location pages second, research-intent guides third, learning content as the long game. One properly built page a week beats a burst of ten and six months of silence.
A worked example: one trade, one afternoon
An HVAC contractor in a mid-size town runs the method. Harvest: the call log yields "water heater making banging noise", "is a water heater flush worth it", "water heater service cost"; autocomplete adds "water heater service near me" and "tankless water heater installation [town]"; Search Console shows the site already at position 11 for "water heater repair [town]" with no page really targeting it.
Sort: "water heater repair [town]" and "tankless water heater installation [town]" are buying intent; "water heater service cost" and "is a water heater flush worth it" are deciding intent; "water heater making banging noise" is learning intent. Winnability check: page one for the repair term is two local rivals with thin pages and a directory, clearly winnable; the installation term shows national brands, park it for now.
The plan that falls out: optimize the existing repair page first (it is already at 11), build a proper cost guide next, then the banging noise article, then revisit the installation term when the site is stronger. Four targets, sequenced by payback, chosen on evidence. That is the whole method working as intended.
What about volume numbers?
Free tools give rough ranges and paid tools give estimates; both are useful for comparing terms, not for precision. For local terms especially, tools understate volume badly. A term your customers demonstrably use (it came from your call log, autocomplete confirms it) is worth a page even when a tool claims near-zero volume. Trust evidence from your market over a database's guess.
Revisit quarterly
Keyword research is not a one-time ceremony. Each quarter, skim Search Console for new query opportunities, check which targets moved, and adjust. Fifteen minutes, not a weekend.
This research is the first deliverable of our SEO content service, where it becomes briefs and finished pages. And if you want to know which terms your site could realistically win right now, the free audit includes exactly that assessment.