Main area
Agency
on-page-seo

Keyword Research — Methodology and Tools from Seed to Mapping

Redaktion ·

Keyword research often gets reduced to a single action: open a tool, read off the search volume, pick the biggest number. That’s the fastest route to content that doesn’t rank and converts no one. Good keyword research is a five-step process — seed, expansion, evaluation, clustering, mapping — and search volume is just one of several dimensions. This article walks through the methodology and shows where the typical traps lie.

The process in five steps

Before the details, the big picture. Keyword research runs in a fixed order:

  1. Seed keywords — define the topic anchors
  2. Expansion — turn the seeds into hundreds of variants
  3. Evaluation — filter by volume, difficulty, and competition
  4. Clustering — group keywords by search intent
  5. Mapping — distribute clusters onto concrete page types

Each step builds on the previous one. Skip the clustering and go straight from the tool list to content, and you produce cannibalization and thematic chaos.

Step 1: Seed keywords

Seeds are the central terms that describe your business — not targets, but starting points. For an SEO agency that might be “SEO,” “search engine optimization,” “keyword research,” “local SEO.” Three to ten is plenty.

Where do they come from? From your head, from your product and service pages, from your customers’ vocabulary. The key: think like the searcher, not like the company. A plumbing firm sells “hydraulic balancing” — the customer searches “heating won’t get warm.”

Step 2: Expansion

Now you turn those few seeds into a large list. The sources:

  • Keyword tools take a seed and return hundreds of related variants along with volume and difficulty.
  • Google Autosuggest — the suggestions that appear as you type are real, frequent search queries.
  • People Also Ask / related searches — the question boxes and the block at the bottom of the page reveal related intents.
  • Competitor analysis — which keywords rank for the competitors you keep seeing in the SERPs?

At the end of this step you have a raw, unsorted list, often with several hundred terms. That’s intentional — filtering happens in the next step.

Step 3: Evaluate the metrics

Three numbers decide whether a keyword is worth it:

Search volume — how often a term is searched per month. Tools like the Keyword Planner provide estimates here, with the Planner often showing ranges (e.g. “1,000–10,000”) rather than an exact figure. Volume tells you reach, nothing more.

Keyword difficulty (KD) — how hard it is to rank on page 1, estimated from the authority of the competing pages. As a rough guide common tools use (as of 2026): KD 0–30 is feasible for younger domains, 30–60 requires real authority, 60+ is very hard. The scale is tool-specific and only an estimate — not a hard truth.

CPC as a competition proxy — the click price from the ad context reveals how commercially valuable a keyword is. A high CPC means advertisers genuinely pay for this traffic, so the keyword has commercial substance. For SEO that’s a useful signal of competition and purchase intent, not a direct ranking factor.

Step 4: Cluster by search intent

This is where usable research splits from amateur work. Instead of chasing individual keywords, you group them by the underlying intent. The four classic intent types:

  • Informational — the searcher wants to learn (“what is keyword research”)
  • Navigational — they’re looking for a specific site (“Search Console login”)
  • Commercial — they’re researching before a decision (“best SEO tools 2026”)
  • Transactional — they’re ready to act (“buy SEO tool”)

Terms with the same intent that answer the same question belong in the same cluster — and therefore on the same page. “Keyword research,” “how do I do keyword research,” and “keyword research guide” are one cluster, not three pages. Build a separate page for every variant and you cannibalize yourself. More on this in the lexikon article Understanding Search Intent.

Step 5: Map to page types

Finally you distribute the clusters onto page types. Intent decides:

  • Informational → blog posts, guides, lexikon
  • Commercial → comparison pages, “best X” lists, case studies
  • Transactional → product, service, and landing pages
  • Navigational → homepage, brand pages

This turns the keyword list into a content plan: every cluster gets its home, and you know before writing what a page is supposed to achieve.

Long-tail vs. head

Head keywords are short, generic, high-volume (“shoes”). They look tempting but are fiercely contested and vague in intent. Long-tail keywords are longer and more specific (“waterproof women’s hiking boots size 39”). Individually they have little volume; in aggregate they make up the bulk of all searches.

The crucial point: long-tail converts better. Someone searching “waterproof women’s hiking boots size 39” is close to buying. Someone searching “shoes” doesn’t even know what they want yet. For most sites — especially younger ones — long-tail is the realistic entry point: low difficulty, clear intent, high conversion.

The tools

Google Keyword Planner — free with a Google Ads account. Two functions: discover new keywords (enter a seed or URL, get suggestions) and pull volume plus forecasts for existing lists. Provides volume, average click prices, and a competition indicator. Important limitation: without an active campaign, the Planner shows volume only in broad ranges.

Google Search Console — the gold source, and free at that. The performance report shows the actual search queries users already find your page with, including impressions and clicks. No third-party tool knows better what you’re genuinely found for. If you already have an indexed site, start here — not with estimated volumes.

Third-party tools (Semrush, Ahrefs, Sistrix, etc.) — provide the largest databases for expansion, difficulty scores, and competitor analysis. Paid, but the standard for serious research. Their KD and volume figures are estimates and vary across tools — useful as a guide, not as absolute truth.

Common mistakes

Chasing volume only. Picking the biggest keyword with the highest difficulty is the classic error. A new site never ranks for it. Better: feasible difficulty plus clear intent.

Ignoring intent. Targeting a blog post at a transactional keyword (or vice versa) doesn’t work. Google reveals the intent through the SERP — look at which page type currently ranks and deliver the same.

Overlooking Search Console. Many buy expensive tools and ignore the free gold source with real data about their own site.

Creating cannibalization. Building several pages for the same cluster instead of one strong one. That dilutes rankings rather than concentrating them.

FAQ

Do I need paid tools for keyword research? To get started, no. Google Keyword Planner and Search Console are free and cover discovery and real query data. Third-party tools like Semrush or Ahrefs offer larger databases and better difficulty scores — worthwhile for serious, ongoing research, but not a required entry point.

Is search volume the most important criterion? No, and that’s the most common beginner mistake. Volume only tells you potential reach. Whether a keyword fits you is decided by keyword difficulty (can I rank at all?) and above all by search intent (does the searcher want what I offer?).

Why is Search Console the best source? Because it shows real data, not estimates: the actual search queries users find your page with, including impressions and clicks. No third party can model that more accurately for your specific site. The prerequisite is an already indexed website.

What does keyword difficulty mean exactly? An estimate of how hard it is to rank on page 1 for a term — derived from the authority of the currently ranking pages. The scale is tool-specific (often 0–100) and only a guide, not a guaranteed statement about your chances.

How many keywords should one page target? A page targets a cluster of related keywords with the same intent, not a single one. A main term plus its variants and questions. Building a separate page for every single variant leads to cannibalization.

Themenuebersicht

On-Page SEOKeyword-Recherche