Keyword research is simply figuring out the words people type into search — and it's the difference between writing content nobody looks for and content that quietly pulls in visitors for years. You don't need expensive tools to start; you need the right questions.
This guide covers how to find real keywords, read the intent behind them, judge whether you can rank, and group them into content that works.
What you need
- ✓A topic or product you want to be found for.
- ✓A keyword idea source (autocomplete, related searches, a keyword tool).
- ✓A simple sheet to collect and group the keywords.
Step-by-step
- 1
Start with seed topics
List the broad topics your audience cares about — the buckets, not the exact phrases. If you sell running shoes, seeds are 'running shoes', 'marathon training', 'foot pain'. These are the roots you'll branch from.
- 2
Expand with real search suggestions
Feed each seed into search autocomplete and 'people also ask'. Those suggestions are literally what real people type. A keyword generator or research tool multiplies this fast, surfacing questions, comparisons and long-tail phrases.
- 3
Read the search intent
For each keyword, ask what the searcher wants: to learn (informational), to buy (transactional), to find a specific site (navigational), or to compare. Match your page to the intent, or it won't rank no matter how good it is.
- 4
Judge competition and pick your battles
Broad head terms ('shoes') are fiercely competitive; specific long-tail phrases ('best trail running shoes for flat feet') are easier to rank for and often convert better. New sites should start long-tail and work up.
- 5
Group keywords into topics
Cluster related keywords that share intent — they belong on one page, not many thin ones. One strong page targeting a cluster beats ten shallow pages competing with each other.
- 6
Map keywords to content and track
Assign each cluster to a page (existing or new), write for the searcher, and check over time which keywords actually bring traffic — then double down on what works.
Examples
- Seed 'password' expands via autocomplete into 'how to create a strong password', 'password manager vs browser', 'is my password safe' — each a real, rankable intent.
- Clustering 'reduce image size', 'compress image', 'make image smaller' onto one page (they share intent) instead of three thin pages that would compete.
Tips
- →Autocomplete and 'people also ask' are free goldmines of real queries — mine them first.
- →Match the page to intent: a how-to won't rank for a buying query, and vice versa.
- →New sites win on long-tail; specificity beats volume until you have authority.
- →One page per cluster — don't split near-identical keywords into competing pages.
- →Track which keywords actually convert, not just which get impressions.
Common mistakes
- Chasing only high-volume head terms. They're the hardest to rank; start with specific long-tail phrases.
- Ignoring search intent. Match your page type to what the searcher wants, or it won't rank.
- Making a thin page per keyword. Cluster related keywords onto one strong page instead.
- Guessing keywords from your head. Use autocomplete and tools to find what people actually search, not what you assume.
Conclusion
Keyword research isn't guesswork — it's listening to what people already search. Start from seed topics, expand with real suggestions, read the intent, favour long-tail where you can win, and cluster keywords into focused pages. Do that and your content targets demand that actually exists.