Keyword research: match search intent to useful pages

Choose the searches your business can answer well, then give each useful intent a clear destination. A longer keyword list is not a better content strategy.

Start with a customer decision

This workflow helps a business decide which existing pages to improve and which genuinely missing pages to create. Use it when a website describes products clearly to its owner but customers still struggle to find the right service. It also works before commissioning a batch of AI-written articles.

Write down one audience, the decision they face and the action your business can fulfil. “People moving a piano between Singapore apartments need to understand access requirements and request a quote” is more useful than “rank for moving.” Collect wording from enquiries, support tickets and sales conversations, removing personal information. These are topic leads, not measured search volumes.

Separate observed demand from ideas

Use the free Search Data workbench to analyse your Search Console query CSV in your browser. It groups observed average positions, highlights review opportunities and compares compatible periods. The files are not uploaded, and no live AI connection is needed.

Open the Search Console performance report for the correct property. Choose a reporting period, search type, country and device before reviewing queries and pages. Keep those settings with the export. Impressions show appearances associated with your property, not the total number of searches in the market.

Clicks divided by impressions gives CTR. Average position describes the highest-positioned result for the relevant property or grouping, averaged over its observations; it is not a universal live rank. A position change can reflect a changing mix of queries or devices. A missing query row does not prove zero demand: Google’s API documentation explicitly says it does not guarantee all rows.

For a new site without data, use customer language and manual result inspection to form hypotheses. Record the date, location and question you investigated. Do not relabel an AI-generated phrase as “high volume” or “low competition” without an actual measurement and methodology.

Build an intent-to-page map

  1. Group by the decision. “Piano moving cost” and “how much to move a piano” may need the same pricing explanation. “Can movers use the stairs?” needs access and safety details, which may fit a section on that page.
  2. Inspect the existing destination. Check whether its offer, title, main heading and examples answer the intent. A keyword appearing once is not evidence that the page is useful.
  3. Choose one main destination per intent. Reuse a relevant service page before creating near-duplicates. Distinct services or audiences can justify separate pages when their information is materially different.
  4. Record the evidence and next action. Keep the phrase, observed query data if available, proposed URL, missing information, owner and verification date in a simple sheet.

SingCompare’s keyword suggestions and page-placement observations help inspect coverage. They do not establish market demand, keyword difficulty or a measured Google ranking. Use them as inputs to this map.

Worked example: a moving service

Illustrative example, not observed business data: a mover’s export contains 200 impressions and four clicks for “piano moving cost” in a selected 28-day window. That is 2% CTR. The service page only says “contact us,” so the owner adds the real factors used in a quote: piano type, access, distance and any separately charged services. No invented price is necessary.

Illustrative decisions; no search-volume estimates
Customer intentDestinationUseful addition
Understand moving cost/piano-moving/Quote factors and a factual example
Check stair accessThe same service pageAccess checklist and exclusions
Prepare before collection/guides/prepare-piano/A specific preparation checklist

The first change improves a real decision. It does not establish that 2% is intrinsically poor or that rewriting the page will increase clicks.

Publish a small change and verify it

Update the selected page through your CMS; retain its previous copy or version history. Align the title and description with the information actually added. Use descriptive internal links from relevant existing pages. Avoid changing several important URLs merely to fit exact phrases.

Inspect the published page and rerun its diagnostic. Later, compare the same page and query group over comparable reporting windows, keeping country, device and search type consistent. Note seasonality, promotions and other releases. A changed metric is an observation; this before-and-after comparison alone cannot prove causation.

Avoid volume without usefulness

Common mistakes are writing a page for every wording variant, copying competitors’ headings without understanding the customer, treating average position as a fixed rank, and assuming every low-CTR query deserves a rewrite. A broad informational query may reasonably earn fewer clicks than a specific buying query.

Google’s helpful-content guidance asks whether content contributes original value. Use AI to organise your evidence and draft clear explanations; verify prices, capabilities and experience yourself. Review the map when the service changes, rather than publishing pages solely to meet a weekly quota.

Sources and editorial approach

Primary documentation reviewed on . Examples are explicitly illustrative; they are not customer results or measured search data. Workflows are SingCompare’s practical recommendations. Provider documentation remains the authority for its own rules.

  1. Google Search Console: Performance report overview and metrics
  2. Google Search Console API: Search Analytics query and row limitations
  3. Google Search Central: Creating helpful, reliable, people-first content

Continue reading

Write useful titles and descriptionsConnect the right pages with internal linksMeasure changes with the right evidence

Start with your website.

Get a free technical SEO and GEO readiness check, with findings you can inspect. No sign-in required.

Run free diagnostic