Keyword research sets the direction for almost everything else in an SEO engagement: which pages get built, what the content says, and which terms a site has any realistic chance of ranking for. Done poorly, it produces a list of high-volume terms a brand-new site will never rank for. Done well, it produces a roadmap of achievable, commercially relevant terms ranked by how much business value each one carries.
The process generally runs through these steps, in roughly this order:
- Pull raw keyword data from discovery tools (volume, difficulty, related terms)
- Run a competitor gap analysis to surface terms a client isn’t targeting yet
- Map each keyword to its underlying search intent
- Filter the list by volume and difficulty against the site’s current authority
- Layer in seasonal timing where relevant
- Group the surviving keywords into clusters mapped to site hierarchy
- Prioritize by commercial value, not volume alone
- Evaluate long-tail coverage to find the lowest-competition opportunities
Using Advanced Keyword Discovery Tools
Agencies start with dedicated keyword research platforms, most commonly Ahrefs, Semrush, or Google’s own Keyword Planner, to pull search volume, ranking difficulty, and related-term data for a client’s market. These tools draw from clickstream and search-engine data rather than guesswork, which matters because keyword volume is heavily skewed toward the low end: according to Ahrefs’ analysis of search data, 94.74% of all keywords receive 10 or fewer monthly searches in the United States. That single statistic is why keyword research can’t stop at a handful of obvious head terms; the bulk of addressable search volume lives in a long tail of smaller, more specific queries.
It also helps explain why search itself keeps shifting under researchers’ feet. Google has stated, and reaffirmed multiple times since first noting it in 2013, that roughly 15% of the searches it sees on any given day have never been searched before. A keyword list built once and never revisited misses that constant churn of new query phrasing.
Analyzing Competitor Ranking Gaps
Beyond a client’s own target terms, agencies run gap analyses that show which keywords competitors rank for that the client doesn’t. This usually surfaces opportunities the client hadn’t considered, often terms tied to a competitor’s specific product feature, service area, or content format. The deeper version of this work, including backlink and content-format comparison, is its own process, covered in our guide to how an SEO company performs competitor analysis.
Mapping Keywords to User Search Intent
Every keyword carries an implied intent, and matching content type to that intent is what separates rankings that convert from rankings that don’t. Research from Amra & Elma’s search intent study puts the current breakdown at four categories, a split that shifts from year to year as AI Overviews and shopping assistants change how people search:
| Search Intent | Share of Queries | What It Means for Content |
|---|---|---|
| Informational | ~57% | Blog posts, guides, and explainer content answering a question |
| Navigational | ~33% | Branded searches looking for a specific site or page |
| Commercial | ~16% | Comparison and "best of" content for buyers still evaluating options |
| Transactional | Under 1% | Product, service, and landing pages built to convert |
That distribution matters operationally: a site that only builds transactional landing pages is ignoring the majority of search volume, which is informational, while a site that only publishes blog content misses the smaller but higher-intent commercial and transactional searches that tend to convert at a much higher rate.
Filtering by Search Volume and Difficulty
Once a raw keyword list exists, agencies filter it against two variables: search volume and ranking difficulty, weighed against the site’s current domain strength. A newer site with limited backlink authority generally has a poor chance of ranking for highly competitive terms even with strong content, so agencies typically prioritize a mix: a smaller number of higher-volume “stretch” terms for the long term, alongside a larger set of lower-competition terms the site can realistically rank for within months rather than years. The exact difficulty threshold an agency uses shifts with the site’s existing authority and the competitiveness of the niche, which is why this filtering step is a judgment call informed by tool data rather than a fixed cutoff applied identically to every client.
To illustrate how this filtering plays out in practice, consider a hypothetical example rather than live tool output: a personal injury law firm might find that a broad term like “car accident lawyer” in a major metro carries very high search volume and very high difficulty, while a more specific variation like “what to do after a car accident” carries meaningfully lower volume but is realistically rankable for a newer site and still reflects strong commercial intent. Actual volume and difficulty numbers vary by market and change over time, so any real keyword plan should be built from a live pull in Ahrefs, Semrush, or Google Keyword Planner rather than figures quoted from an article.
Identifying Seasonal Keyword Opportunities
Search volume for many keyword categories swings with the calendar, tax preparation, holiday retail, HVAC repair, and agencies use historical trend data (commonly pulled from Google Trends alongside tool-level monthly volume history) to plan content ahead of demand rather than reacting to it. As a general practice, agencies aim to have seasonal content published and indexed well before the relevant demand window opens, since Google needs time to crawl, index, and build ranking signal for new pages; the exact lead time needed depends on how competitive the seasonal term is and how established the site already is, but waiting until the season starts is consistently too late.
Grouping Keywords by Website Hierarchy
Once prioritized, keywords get grouped into topic clusters and mapped to a planned page or section of the site, rather than treated as a flat list. This grouping step is what turns keyword research into an actual site plan: which terms belong on a service page, which belong in supporting blog content, and which terms are close enough in meaning that they should target the same page rather than create competing, near-duplicate pages. The structural side of this work, including how those clusters map to URL architecture and internal linking, is covered separately in our piece on how an SEO company optimizes website structure.
Prioritizing High-Conversion Terminology
Volume alone is a poor proxy for value. Agencies weight keyword priority by commercial intent signals, terms that include qualifiers like “cost,” “near me,” “best,” or a specific product/service name tend to convert at meaningfully higher rates than broad informational terms, even when their search volume is much lower. This is also where keyword research connects directly to a client’s actual revenue goals rather than staying a purely traffic-driven exercise, and it is part of why keyword prioritization should never happen in isolation from the client’s sales and conversion data.
Evaluating Long-Tail Keyword Potential
Long-tail keywords, the longer, more specific phrases that individually carry low search volume, account for roughly 70% of all search traffic according to Ahrefs’ research. A separate Backlinko analysis of more than 300 million keywords measured a different angle of the same pattern: 91.8% of all distinct search queries qualify as long-tail, even though each individual term draws only a trickle of volume on its own. Individually these terms are easy to overlook, but in aggregate they often represent the largest, least competitive opportunity available to a smaller site, since they face far less competition than the handful of head terms everyone is chasing. Building a keyword strategy that captures this long tail, rather than fixating only on the highest-volume terms, is frequently the difference between a site that grows steadily and one that stays stuck competing for terms it can’t yet win.
Good keyword research is never a one-time deliverable. Search behavior shifts, competitors publish new content, and tool data updates monthly, so agencies that treat the keyword list as a living document, revisited quarterly at minimum, consistently outperform those that build it once at the start of an engagement and never touch it again. Once that list informs a content plan, the longer-term execution of it, including refresh cycles and topic clusters, is covered in our guide to content strategies for long-term SEO growth.
