Check this page with an assistantOpens a chat asking it to summarise this article and name the evidence behind each claim.
Claude opens with the prompt on your clipboard: Anthropic does not support prefilled prompts on the web, and we would rather copy it than ship a button that drops it.
Why Most Keyword Advice Fails a Five-Person Business
Mainstream keyword advice is built for content teams with publishing quotas, so it optimizes for volume: big exports, big head terms, big dashboards. The queries that actually grow a five-person business are specific, local, and low volume, exactly the ones that tools commonly priced at $99 or more per month deprioritize and that free methods surface well. What you need is a short list of queries your customers really type, ranked by intent, not a 5,000-row spreadsheet.
Look at what a subscription tool is engineered to do: estimate search volume at scale and compute difficulty scores across hundreds of thousands of terms. Those features earn their price when you are prioritizing among thousands of candidate topics. A plumber, a bookkeeper, or a boutique agency is not doing that. You are choosing among maybe forty plausible queries, most of which the tools list as "10 to 100 searches" or no data at all, because local and niche demand sits below their measurement floor.
The failure mode this produces is predictable. The owner signs up for a trial, exports a thousand rows, sorts by volume, and starts chasing terms like "accounting software" that they will never rank for and that would not convert if they did. Meanwhile the query "bookkeeper for restaurants [city]" goes unclaimed by anyone. Free sources are not a budget compromise here. For this job, they are the better instrument.
The Free Sources, and What Each Is Actually For
Each free source answers a different question, and the workflow below draws on all of them. None requires a login except Search Console, which is worth the one-time setup on its own.
| Source | What it reveals | Time needed |
|---|---|---|
| Google Autocomplete | Real phrasings people type, including local variants you would not guess | 10 min |
| "People also ask" boxes | Question intents worth answering on-page | 15 min |
| Google Search Console | Queries you already almost rank for; requires setup, covered in our small business SEO checklist | 20 min |
| Competitors' page titles | What they believe converts, visible in one search | 15 min |
| Business Profile insights and your sales inbox | The words customers use unprompted | 15 min |
| Free tiers of keyword tools | Spot-checking rough volume ranges; optional | 10 min |
Times assume one focused pass; the total is the ninety minutes the workflow below is built around.
The 90-Minute Workflow, Step by Step
The workflow runs in six steps: seed a list from your services, expand it with autocomplete, mine the question boxes, check Search Console for queries within striking distance, sweep competitor titles, then score everything on intent and winnability. The output is a ranked list of 15 to 25 queries, each mapped to a page type. Block ninety minutes and run it in one sitting.
- Write the seed list.List every service you sell and every problem it solves, in your customers' words, until you have about ten terms. "Emergency drain clearing" and "blocked kitchen sink" both belong; internal jargon does not.
- Expand with autocomplete. Type each seed into Google slowly and record the suggestions, then repeat with modifiers appended: near me, cost, best, vs, and your town name. Ten seeds typically become forty to sixty candidate phrasings, all of them things people demonstrably type.
- Mine "People also ask." Search your strongest candidates and expand the question boxes; each click loads more. Copy every question a real buyer would ask. These become FAQ sections and guide topics rather than standalone pages, so keep them in a separate column.
- Check Search Console for striking distance. Open the Performance report, filter to the last three months, and sort queries by impressions. Anything sitting at average position 5 to 15 is a query Google already half-trusts you for, and improving an existing page usually moves it faster than anything else on this list.
- Sweep competitor titles.Search your top five candidates and read the page titles that rank. Titles are the one piece of a competitor's research you can see for free: the phrasing they chose is the phrasing they believe converts. Note any angle they all share and any obvious query none of them covers.
- Score every candidate. Rate each query 1 to 3 on intent, meaning would the searcher plausibly pay someone, and 1 to 3 on winnability, meaning who ranks now: if the first page is all national brands, score 1; if small sites and directories rank, score 3. Sort by the combined score.
Keep the top 15 to 25 and map each to a page type: service page, comparison, guide, or FAQ section. That mapping is the deliverable, and everything below the cut line goes into a parking lot for next quarter, not into your build queue.
Reading Intent: The Only Scoring That Matters
A query's value is a function of who is typing it and how close they are to paying, not how many people type it. "Emergency plumber [town]" beats "how do drains work" at one hundredth the volume, because one searcher is holding a phone over a flooding kitchen and the other is idly curious. Every scoring decision in this workflow reduces to that distinction.
In practice, queries sort into three tiers. Buy-now queries name a service plus an urgency or location signal: "same day bookkeeping cleanup." Comparing queries weigh options: "quickbooks vs hiring a bookkeeper." Researching queries seek understanding: "what does a bookkeeper actually do." All three have a role, but they are worth wildly different amounts per visit.
The rule for a small site: build for buy-now and comparing queries first, and add research queries only once those pages exist. Research content earns trust and feeds internal links, but it converts through the commercial pages, so publishing it first builds an audience with nowhere to land. If you want to validate intent quickly before committing to a page, a small ad test can settle it in days; our SEO vs Google Ads comparison covers how paid click data doubles as free intent research.
From Keyword List to Page Plan
The conversion from list to plan follows one rule: each surviving query gets exactly one page, and near-duplicate queries share one page. "Bookkeeper for restaurants" and "restaurant bookkeeping services" are the same intent and must not become two pages, because two pages competing for one query split your ranking signals. This single rule prevents the cannibalization that quietly caps most small sites.
The mapping logic is short. Transactional queries become service or pricing pages. Comparison queries become honest versus pages. Questions become FAQ sections on an existing page, or a guide when the answer genuinely needs a thousand words. Once the plan exists, work through it at whatever cadence you can sustain; our guides on writing SEO-friendly posts and how many posts per week you actually need cover the execution side.
Honest disclosure: this mapping and prioritization step is precisely what SearchHandled automates. The Growth Map reads your Search Console data and produces the ranked create-refresh-merge queue this article teaches you to build by hand. The manual version above works fine at small scale, and if it stops fitting in ninety minutes a quarter, that is the signal to automate it.
Questions Owners Ask About Finding Keywords
- How do I do keyword research for free?
Five free sources cover it. Google Autocomplete for real phrasings, the People also ask boxes for question intents, Google Search Console for queries you already almost rank for, competitor page titles for what converts in your market, and your own sales inbox for the words customers use unprompted. Score each candidate query on intent and winnability, then map every survivor to one page. The whole workflow takes about ninety minutes and repeats quarterly.
- What are long-tail keywords and why do they matter for small businesses?
Long-tail keywords are longer, more specific queries with low individual search volume, such as an equipment name plus a town plus the word cost, rather than a single broad head term. They matter for small businesses because the intent is clearer, the competition is thinner, and the searcher is usually much closer to paying. A small site can realistically win a long-tail query this quarter; it may never win the head term.
- How many keywords should a small business target?
Fifteen to twenty-five well-chosen queries is plenty for a first pass, with each query mapped to exactly one page. That is enough to cover your core services, your most common comparisons, and the questions buyers actually ask, without spreading a small site thin. Expand the list only after those pages exist and Search Console shows how they perform.
- Is Google Keyword Planner accurate without running ads?
Not very. Without an active campaign, Keyword Planner shows broad volume ranges such as 100 to 1K instead of precise counts, and it lumps similar phrases into one bucket. Use it for ideas and rough sanity checks, not precision. For a small business, whether a query is typed by people ready to pay matters far more than whether it shows 90 or 400 monthly searches.
