Keyword research is the work of finding out what people actually search for and matching your site’s content to that demand. It comes before ranking, before titles, before technical settings — because no matter how well-optimized a page is, if it targets the wrong query, it pulls in the wrong visitor.
The core misconception here is treating keyword selection as “listing whatever terms come to mind.” If your product is called “marketing automation” but customers are searching for “bulk email software,” the traffic gets lost somewhere in between.
Ahrefs’ keyword research guide frames the process around a balance of intent, volume, difficulty, and business value. Moz’s related section advises against producing content before validating demand. On the academic side, Broder’s classic search taxonomy showed as far back as 2002 that queries already split into three distinct intent groups.
Search Intent: The Purpose Behind the Query
Two queries on the same topic can carry very different intent. The common classification breaks into four groups:
- Informational: “what is cpm,” “how to get email consent.” The user wants to learn; they may be far from buying.
- Navigational: “search console login.” The user wants to reach a specific site; these queries are hard to intercept.
- Commercial investigation: “best crm software,” “x vs y comparison.” The decision process has already begun.
- Transactional: “crm pricing,” “request a demo.” The group with the highest intent to act.
Searching the target query on Google yourself is the cheapest way to read intent. Is the first page dominated by guide content, product pages, or comparison lists? The first page already tells you which page type Google rewards for that query. Trying to compete for a product-page intent with a blog post, or an informational intent with a product page, is the most common matching mistake.
The Volume, Difficulty, and Long-Tail Balance
Volume alone shouldn’t decide it
A broad term searched 10,000 times a month looks tempting; but high volume usually means high competition and unclear intent. Someone searching “CRM” — do they want a definition, software, or career information? It’s not clear. Someone searching “crm with accounting integration” 90 times a month has clear intent and a high likelihood of converting.
Long-tail queries add up to more
Long, specific queries (long-tail) have small individual volumes, but their totals outweigh broad terms. For a new site or one with low authority, the sensible order is usually: build visibility on long-tail queries first, then move toward more competitive terms as authority accumulates. How that authority builds up is the subject of What Is a Backlink?.
Read the real difficulty signal from actual pages
Tool difficulty scores are a useful starting point, but the real signal is the competitors on page one. If all ten top results are big brands with comprehensive guides, that query shouldn’t go on your short-term target list. If small sites, forums, or thin content show up instead, the door is open.
Where Do You Find the Data?
Three free sources are enough for most early-stage research:
- Search Console performance report: The queries your site already gets impressions for. This is the most valuable list, because it shows what topics Google already associates with your site. Queries that get impressions but not clicks are opportunities to strengthen existing content.
- Google Keyword Planner: Google’s official tool is built for advertising, but its volume ranges and related-term suggestions are useful for organic research too.
- Google’s own suggestions: Autocomplete, “Related searches,” and “People also ask” boxes are a free data source reflecting real user language.
Count conversations with customers or the sales team as a data source too. Writing down the five questions you hear most often can sometimes surface a more accurate topic than the 500 keywords a tool suggests.
Keyword Map Checklist
The most useful output of research isn’t a keyword list — it’s a keyword map table. For a new site, fill in the table in this order:
| Step | Question | Output |
|---|---|---|
| Competitor selection | Who’s actually ranking for these queries? | Organic competitor list |
| Core keyword | What’s the broadest phrasing of this topic that’s still meaningful? | Core topic cluster |
| Long-tail variation | How does a user search this topic more specifically? | Low-competition opportunities |
| Page intent | Does page one favor guides, products, categories, video, or local results? | Page type decision |
| User questions | What do “People Also Ask,” sales questions, and support requests reveal? | H2/H3 and FAQ candidates |
| Difficulty | How strong are the authority and content quality of page-one sites? | Short-, medium-, or long-term target |
| Target URL | Which page should answer this query group? | URL decision that reduces cannibalization |
This table especially streamlines the content calendar. If a query group is already served by an existing page, that page should be strengthened instead of publishing a new post. If the intent is different, a separate page should be created. Content calendars built without this distinction often end up covering the same topic under different titles.
Keyword-to-Page Matching and Cannibalization
The output of research isn’t a keyword list — it’s a mapping table: which query group, which page, which intent. Similar queries should be consolidated onto a single page. Writing three separate posts for “what is crm,” “what does crm do,” and “crm meaning” is unnecessary, because the intent is the same.
The opposite scenario also causes problems. When two separate pages target the same query, Google can’t decide which one to show, and the two pages end up weakening each other’s visibility — this is called keyword cannibalization. If you notice two URLs alternating in the performance report for the same query, consider either merging the pages or differentiating their intent.
One last distinction: this targeting decision is a different job from choosing a match type in Google Ads. On the ad side, “broad or exact match” is an instruction given to the bidding system; that mechanism is covered in What Are Keyword Match Types?. On the organic side, there’s no match type — Google matches whatever the page’s content actually covers.
Summary
Keyword research answers four questions: what is the user searching for, with what intent, how much competition exists for that query, and which page will answer it? Prioritize intent over volume, start with long-tail on a new site, and match each query group to a single page. Once the mapping table is ready, the next job is optimizing those pages — the subject of What Is On-Page SEO?.