SEO Keywords: How to Research and Prioritize

SEO keywords are the foundation of every content decision your team makes. Poor keyword research means months of publishing pages that never rank, convert, or justify their budget. Good research makes each article compound into measurable organic traffic and pipeline.

This guide walks through the full keyword research workflow we run inside SoarTask — from seed lists to keyword clustering to a prioritization table you can hand to a writer today. It is the same process we use to plan SEO strategy and measure outcomes with SEO analytics. If you're new to the discipline, start with our primer on what is SEO first.

What are SEO keywords, really?

An SEO keyword is any phrase a person types into a search engine when they have a question, a problem, or an intent to act. In modern SEO content production, a "keyword" is usually a query string that carries meaning, context, and expected outcome, not just a single word. The job of keyword research is to find the queries your audience actually uses, estimate how valuable ranking for them would be, and decide which ones deserve a page.

Three things make keyword research harder than it looks:

• Search volume is a range, not a fact. Google Keyword Planner buckets volumes into broad ranges for most accounts, and third-party tools model their own estimates. Treat any single number as directional, not exact.
• Difficulty scores are proprietary. Every tool calculates "keyword difficulty" differently — some prioritize backlinks, others domain authority, and still others SERP features. A KD of 45 in one tool is not comparable to a KD of 45 in another.
• Intent is the real ranking factor. A high-volume keyword with the wrong intent will never convert, no matter how well you rank.

The keyword research workflow, step by step

We run keyword research as a repeatable, six-step workflow. Each step has a clear input, output, and owner so the work can be delegated and audited — which is the whole point of running SEO as operations rather than guesswork.

Step 1 — Build a seed keyword list

Seed keywords are the broad terms that describe your product, service, or topic area. They are the starting point, not the final list. You generate them from four sources:

1. Brainstorm with the revenue team. Ask sales and support how customers actually describe the problem. "CRM" becomes "customer tracking spreadsheet" and "sales pipeline tool."
2. Google Keyword Planner. Enter each seed and pull the suggested variations. This is free, directly from Google's data, and gives you a baseline volume range.
3. Competitor SERP scan. Pull the top 10 ranking pages for each seed and note the phrases in their titles, H1s, and URL slugs.
4. Customer interviews and support tickets. Real questions from real users surface long-tail terms no tool will suggest.

Aim for 30–80 seed keywords for a typical content vertical. You'll expand these into hundreds of candidate terms in the next step.

Step 2 — Expand into candidate keywords

Feed each seed into a keyword expansion tool (Google Keyword Planner, Ahrefs, Semrush, or SoarTask's keyword cluster tool) and collect every related query with a reported volume. At this stage you're collecting, not filtering — duplicates and irrelevant terms get removed later.

What to capture per keyword

For every candidate, record five fields so you can prioritize later without re-pulling data:

• Keyword — the exact query string
• Monthly volume — the tool's estimate (note the source)
• Difficulty — the tool's score (note the scale)
• SERP intent — informational, navigational, commercial, or transactional
• Current ranking URL — if you already have a page, note it

Step 3 — Classify search intent

Search intent is the single most under-weighted factor in keyword research. We use a four-class framework adapted from the search intent guidance in Google's Search Central documentation and Backlinko's search intent study:

Classify by reading the live SERP, not by guessing. Open the top 5 results for each keyword and note the dominant page type. If four of five results are listicles, the intent is commercial — even if the keyword "sounds" informational.

Step 4 — Weigh volume against difficulty

This is where most teams make a common mistake: they sort by volume, take the top 20, and ignore difficulty. The result is a content calendar full of terms they can't rank for. The honest approach is to balance the two and add a third dimension, business value.

Use this simple scoring heuristic:

Opportunity score = (Volume × Business value) ÷ (Difficulty + 1)

Business value is a 1–5 rating you assign based on how closely the keyword maps to a revenue action. A transactional term like "buy seo software" scores 5; an informational term like "history of search engines" scores 1, regardless of volume.

Step 5 — Cluster keywords into topics

Once you have a scored list, keyword clustering groups terms that share the same search intent and can be satisfied by a single page. Clustering is what turns a flat list of 400 keywords into a content plan of 30–60 pages.

Two keywords belong in the same cluster when:

• They share the same core intent (e.g., both informational, same sub-topic)
• The top-ranking pages for both queries overlap by 3+ URLs in the SERP
• One is a natural H2 or H3 of the other within a single article

You can cluster manually for small lists, but at scale use a tool. Our keyword cluster tool groups by SERP overlap so clusters reflect how Google actually relates queries, not just string similarity.

Cluster outputs

Each cluster becomes one content brief. The primary keyword (highest opportunity score) becomes the page's target. Secondary keywords become H2/H3 headings and supporting sections. This is also where you check keyword density to avoid both stuffing and under-optimization — aim for natural usage, not a fixed percentage.

Step 6 — Map clusters to content and prioritize

The final step converts clusters into a prioritized content calendar. Each cluster is assigned a priority (P1–P4), an owner, a content format, and a target URL. P1 clusters are high-opportunity, low-difficulty, and high business value — ship them first. P4 clusters are nice-to-have and get scheduled only after the P1–P3 backlog is cleared.

Keyword prioritization table (real example)

Below is a prioritization table from an actual SEO content plan built in SoarTask. Volume and difficulty are illustrative ranges pulled from Google Keyword Planner and a third-party tool — remember, treat them as directional.

The pattern here is clear: the P1 keywords are not the highest-volume terms. They're the terms where difficulty is low enough to win quickly and intent is clear enough to convert. High-volume terms like "keyword research" stay at P2 because the difficulty (71) means a new page won't rank for 6–12 months, and the informational intent means it won't drive signups directly.

Mapping keywords to content the right way

Once keywords are prioritized, mapping them to content is where strategy becomes execution. The rule we follow: one primary keyword per page, one intent per page. Trying to target multiple intents on one page dilutes relevance and confuses both Google and the reader.

The one-page-one-intent rule

Each page gets exactly one primary keyword (the cluster's highest-opportunity term). Secondary keywords from the same cluster become section headings. Keywords from a different intent class get their own page — even if they're closely related.

For example, "keyword research" (informational) and "keyword research tool" (commercial) belong on separate pages. The first is a guide; the second is a product or comparison page. Combining them forces you to serve two audiences in one URL, which is why hybrid pages rarely rank for either term.

Internal linking between clusters

Clusters that share a topic pillar should link to each other. The informational guide on "keyword research" links to the commercial "keyword research tool" page and vice versa. This passes relevance between pages and helps Google understand the topical relationship. SoarTask's internal link guardrails flag over-optimization automatically — we never let exact-match anchor text exceed 50% of inbound links to any URL.

Common keyword research mistakes to avoid

Trusting search volume too literally

Google Keyword Planner rounds volumes into buckets (e.g., 1K–10K) for most accounts, and third-party tools extrapolate from clickstream samples. A keyword showing "3,600" may actually receive 1,800 or 7,200 in a given month. Use volume to compare keywords relative to each other, never as a forecast of traffic you'll receive.

Ignoring keyword difficulty until it's too late

Difficulty is the cost of ranking. A keyword with volume 10,000 and difficulty 80 will likely take 12+ months and significant backlink investment to reach page 1. For most teams, that's a poor use of resources compared to five keywords with volume 800 and difficulty 20. Always read difficulty alongside volume.

Skipping intent classification

This is the mistake that wastes the most budget. Publishing a 2,000-word guide for a transactional keyword means you'll never rank — Google has decided that query wants product pages. Always open the SERP and verify intent before assigning a content format.

Key takeaways

• SEO keywords are queries with intent, not just strings. Research the intent, not just the volume.
• Seed keywords come from people, not just tools. Sales, support, and customer interviews surface terms no tool will suggest.
• Volume and difficulty must be read together. Opportunity = (Volume × Business value) ÷ (Difficulty + 1).
• Keyword clustering turns a flat list into a content plan. Group by shared SERP overlap and intent, not by string similarity.
• One page, one primary keyword, one intent. Never mix intent classes on a single URL.
• Prioritize by opportunity, not by volume. The P1 queue should be winnable and revenue-relevant, not just big.
• Treat all volume and difficulty numbers as directional. They're estimates, not facts — compare relatively, forecast conservatively.

Frequently asked questions

How many SEO keywords should I target per page?

One primary keyword per page, supported by 3–8 secondary keywords from the same cluster. The primary keyword drives the page's title, H1, and URL. Secondary keywords become H2 and H3 headings. Targeting multiple primary keywords on one page splits relevance and rarely ranks for any of them.

What is the difference between keyword research and keyword clustering?

Keyword research is the process of finding and evaluating candidate queries — volume, difficulty, and intent. Keyword clustering is the next step: grouping those candidates into sets that can each be satisfied by a single page. Research produces a list; clustering produces a content plan.

Is Google Keyword Planner enough for keyword research?

Google Keyword Planner is an excellent free starting point for seed expansion and volume ranges, but it doesn't show difficulty, SERP overlap, or intent classification. For production content planning, pair it with a tool that adds difficulty scoring and clustering — or use SoarTask's integrated keyword workflow, which pulls GSC data from your authorized connection so you're working with your own real query data, not just estimates.

How often should I redo keyword research?

Refresh your keyword list quarterly. Search behavior shifts, new competitors enter, and SERP features change. A keyword that was informational last quarter may shift commercial as Google tests new layouts. Re-pull volume and re-check intent for your top 50 keywords every 90 days, and run a full seed expansion twice a year.

What is a good keyword difficulty score to target?

It depends on your site's authority. For a new or low-authority domain, target difficulty under 20 for your first 20 articles. For an established domain with a strong backlink profile, difficulty 30–50 is realistic. Above 60 is generally a long-term play unless you have a topical authority advantage on that specific subject. Always benchmark against what your existing top-10 pages actually rank for.

Keyword research only creates value when it becomes published, ranked, and measured content. SoarTask turns the workflow above into a managed pipeline — seed, cluster, prioritize, brief, publish, and measure outcomes against GSC and GA4 data, all in one workspace. Start your free SoarTask workspace and run your next keyword research cycle with the same evidence-backed process our team uses every day.