Keyword research is where SEO strategy starts. It's the process of figuring out what questions your customers type into Google, how much traffic each question could bring you, how hard it'd be to rank for, and what content each one deserves. Skip this step and you'll spend months writing pages nobody searches for. Do it well and you'll have a two-year content roadmap ranked by business value. This page walks through the whole workflow end-to-end.
Most people think keyword research is about "finding keywords." That's backwards. You're not looking for keywords. You're listening to your customers. Every query is a person, typing into a box, hoping to find what they need. Your job is to hear what they're asking and build the answers.
That reframing changes everything. Instead of asking "what keyword has high volume," you start asking "what do my customers actually struggle with, and what words do they use to describe it." The answers are very different. The second question is where real SEO strategy lives.
The workflow is the same whether you're working on a 50-page site or a 5,000-page one. What changes is how long you spend on each step. For a new blog, this is a one-week sprint. For an enterprise site, each phase might be its own two-week project.
Start with 5 to 15 core topic terms you already know. What you sell, the language your customers use, the job-to-be-done, the broader category. These aren't your final keywords. They're the starting input for the expansion step.
Plug each seed into a keyword tool (Ahrefs, SEMrush, Google Keyword Planner) and pull related queries. Scrape People Also Ask questions. Grab competitor ranking keywords. By the end of this step you should have 500 to 2,000 candidate queries.
Go through the list and drop what's irrelevant, too broad, or obviously brand-dominated. This is painful but necessary. Expect to remove 50 to 70% of the raw pool.
Group related queries that should target the same page. "Best insurance CRM" and "top insurance CRM software" are the same cluster. Run SERP-similarity clustering to let Google's own behavior tell you what belongs together.
For each cluster, estimate volume, difficulty, and business value. Combine into a priority score. You'll rank everything from quick wins to long-term bets.
Assign each cluster to a specific page (new or existing), a specific format (guide, comparison, product), and a specific content brief. The output is a content calendar that could keep a team busy for months.
Every keyword decision boils down to three numbers. Most opportunities score well on two of three. Chasing perfect scores on all three wastes time.
High volume and low difficulty is a popularity contest with no winners. Low volume and high value is a niche goldmine. High difficulty and high value but low volume is brand-building. Balance all three, but weight by what your business needs most right now.
A good keyword research deliverable is a single spreadsheet. One row per keyword. One tab per cluster summary. Columns include: the query text, which cluster it belongs to, whether it's primary or supporting, monthly volume, difficulty score, intent, current ranking (if any), target URL (existing or new), suggested format, priority, and notes.
Don't overcomplicate the output. The goal is to go from "thousand keywords in a messy tool export" to "here are the 15 pages we should build this quarter, in order." If your deliverable doesn't answer that question clearly, it's not done.
If you've never done keyword research for your site, block a week. Seed, expand, filter, cluster, score, plan. The first time is slow. Every time after is faster because you learn which shortcuts are safe. The output pays off for years, not weeks.
Next: search intent, the most underrated lever in SEO.