Originally published: August 20, 2021 · Last updated: August 17, 2026
Keyword research is not the process of collecting the largest possible spreadsheet of search terms. It is the process of understanding how people describe their needs, which of those needs your site can satisfy, and whether a new or existing page is the right place to answer them.
A useful keyword workflow therefore starts with audience and intent, uses search-volume data as evidence rather than an order, and ends with an editorial decision: create, improve, merge or ignore.
1. Start with the problems your site is meant to solve
Write down the main tasks your intended audience comes to the site for. For CreativWebTools, examples could include choosing a WordPress tool, solving a technical problem, improving search visibility, understanding a web concept or selecting a maintainable workflow.
This prevents keyword research from drifting into unrelated high-volume topics simply because a tool suggests them.
2. Build seed topics from real language
Use customer questions, emails, support requests, internal-site searches, comments, sales conversations and your own experience to identify the words people use naturally.
Do not normalize the language too early. A beginner may search for “WordPress page disappeared” while a technical writer thinks in terms of revisions, Trash or database recovery. Both perspectives are useful.
3. Use Search Console before buying more keyword data
If the site already has search visibility, Search Console is one of the most valuable sources of keyword evidence because it shows queries for which Google already displays your pages.
Look for:
- queries with growing impressions;
- questions that an existing page only partially answers;
- several queries sharing the same underlying intent;
- several pages appearing for the same query family;
- important terminology you are not using clearly enough.
This can produce update opportunities before you create another URL.
4. Expand ideas with research tools
Keyword tools can help discover variations and estimate demand. Google Keyword Planner can generate ideas from seed terms or websites and provide search estimates, though it is built primarily for Google Ads. Google Trends is useful for comparing relative interest, seasonality and regional patterns.
Third-party SEO platforms can add competitive estimates, difficulty metrics and SERP histories. Treat all such metrics as estimates, not facts about what your page will rank for.
5. Group terms by intent, not just by matching words
Several phrases may belong on one page if the visitor is trying to accomplish the same task. Conversely, two phrases that share most of their words can require different pages when the expected answer differs.
For example, “best WordPress backup plugin” suggests comparison, while “how to restore a WordPress backup” suggests a procedure. One large page might cover both, but separate pages can be justified if each intent needs substantial treatment.
6. Inspect the current search results
Search the important query families manually. Look at the types of pages Google currently shows: guides, product pages, videos, local results, forums, tools or comparisons.
Do not copy the top results. Use them to understand what searchers may expect and where existing results are weak, outdated or excessively generic.
7. Separate demand from relevance
A high-volume query can be a poor target if it is outside your expertise or attracts an audience with no relationship to the site’s purpose.
A smaller query can be strategically valuable when it closely matches the people you want to help and connects naturally with other useful content on the site.
This is why “traffic potential” and “business value” should be separate columns in a keyword plan.
8. Evaluate competition realistically
Keyword difficulty scores can help compare opportunities, but they compress many factors into one proprietary metric. A smaller site can still compete when it has better first-hand evidence, a narrower angle, a useful tool, a strong local position or content that serves an overlooked audience.
Ask whether you can make a page materially better, not merely whether a tool says the difficulty is 37 rather than 54.
9. Look for low-competition specificity
Broad terms often attract broad competition. Specific problems can be more realistic for smaller publishers.
Instead of targeting “WordPress SEO”, you might answer a concrete need such as how to prevent a custom post type from being missing in search, how to audit a sitemap after a migration or how to choose one SEO plugin without creating duplicate output.
The specific article can later become part of a broader content cluster.
10. Decide whether the topic needs a new page
Before creating a new URL, search your own site. A related article may already exist and simply need expansion.
Choose among four actions:
- Create: a distinct search intent deserves its own durable resource.
- Improve: the current page is the correct destination but incomplete.
- Merge: several weak pages compete for the same need.
- Ignore: the query is outside scope or not worth maintaining.
This step prevents keyword research from turning into uncontrolled content growth.
11. Map the topic into the site architecture
A valuable page should connect naturally with existing categories, guides, tools and related articles. Plan which pages should link to it and which resources it should link to.
If a proposed topic has no natural relationship to anything else on the site, that can be evidence that it does not belong in the editorial strategy.
12. Write for the task, not the exact phrase
Use the words your audience recognizes in the title, heading and body where they fit naturally. Then focus on explaining the subject completely and accurately.
Google’s Search Essentials recommends using words people would use to find your content in prominent places. That is different from repeating an exact keyword at a prescribed density.
13. Measure after publication
Keyword research does not end when the article goes live. Search Console can show which queries actually generate impressions and which page Google associates with them.
After enough data accumulates, you may discover that the page attracts a different but useful query family, that the title needs clarification or that two pages are competing for the same intent.
14. Keep a keyword map, not a keyword graveyard
A manageable editorial sheet might include:
- topic or intent;
- representative queries;
- audience;
- existing URL if any;
- evidence source;
- business or strategic value;
- difficulty or competition notes;
- decision: create, update, merge or reject;
- status and owner.
Remove ideas that no longer fit instead of letting the sheet become an endless list of obligations.
A practical keyword-research sequence
- Define audience problems.
- Collect seed language from real users.
- Review existing Search Console queries.
- Expand with Keyword Planner, Trends or SEO tools.
- Group phrases by intent.
- Inspect current search results.
- Score relevance and strategic value.
- Assess realistic competition.
- Check your own existing content.
- Choose create, improve, merge or ignore.
- Plan internal links and site position.
- Publish only when the page deserves to exist.
- Measure actual queries afterward.
The practical rule
Keyword research should reduce the number of unnecessary pages you publish, not increase it. Use search data to understand demand, then apply editorial judgment to decide where your site can provide a genuinely useful answer. The best keyword is not the one with the biggest number next to it. It is the search need your site is well placed to satisfy.
Official references: Google SEO Starter Guide, Google Search Console guidance, Google Keyword Planner guidance and Google Trends comparison guidance.