How to do keyword research for beginners (without drowning in data)
A reader emailed me last month with a screenshot of her keyword spreadsheet. Forty-seven tabs. Twelve thousand rows. She'd been building it for five weeks and had published exactly zero articles. That's the trap nobody warns you about: keyword research feels like progress, so beginners keep researching instead of writing.
Here's the uncomfortable truth I've landed on after years of doing this on my own sites and for clients: keyword research for beginners isn't a data problem, it's a decision problem. You don't need more keywords. You need a way to pick three of them and move on. Everything below is built around that idea.
Key Takeaways
- Start with 5-10 "seed" phrases that describe your topic in plain language, not with a tool.
- The metrics that matter early on are search intent and competition, not search volume.
- Long-tail keywords (4+ words) are where a new site can realistically rank in the first months.
- A keyword is only "good" if you can write something genuinely useful about it this week.
- Free tools are enough to start. Paid tools save time later, not at the beginning.
- Re-check your list every quarter, because AI summaries have changed what a click is worth.
What keyword research actually is (and what it isn't)
Strip away the dashboards and keyword research is just listening. You're figuring out the exact words people type or say when they want something you can provide, then deciding which of those words you have a fair shot at answering better than everyone else.
What it isn't: a list of high-volume terms you sprinkle into text. I made that mistake early. I once stuffed a page with a term that had tens of thousands of monthly searches, ranked on page four for about three weeks, and got eleven visitors total. Eleven. The term was too broad, the intent was wrong, and the page answered a question nobody was actually asking in that phrasing.
What is keyword research in SEO?
In SEO specifically, keyword research is the step that tells you which pages to create and how to title them. It connects three things: what people search, what your pages can realistically compete for, and what your site is actually about. Skip it and you're guessing. Do it badly and you'll write excellent content for queries with no audience.
What is keyword research in digital marketing?
Zoom out and the same activity feeds ads, social posts, YouTube titles and email subject lines. The queries you find for organic search often double as the exact phrasing that works in a paid campaign or a short-form video hook. One research session, several channels. That's the practical value, and it's why I stopped treating it as an SEO-only chore.
Why the old beginner advice is quietly breaking
Search results have changed shape. A meaningful share of queries now get answered directly at the top of the page, so a ranking doesn't guarantee a visit the way it did a few years ago. I watched one of my pages sit at position two for a definition-style query and collect roughly 60% fewer clicks than it did eighteen months earlier, with the position barely moving.
What does that mean for someone starting now?
- Chase queries where the answer needs context, examples or a decision — not a one-line definition.
- Conversational, long-form queries tend to survive better than short factual ones.
- Treat visibility in AI assistants as a bonus, not the plan. They pull from pages that already rank and are structured clearly.
- Social search deserves a slot in your process now. People search TikTok and Instagram the way they used to search Google, especially for anything visual or hands-on.
The wider point: pick keywords where a human still needs to click to get value.
A beginner workflow that actually ends
Six steps. I've compressed this from a version that used to have fourteen, most of which were procrastination in disguise.
Step 1: Write your seeds without a tool open
Close the browser. Write down five to ten phrases that describe what you do, in the words a customer would use. If you sell handmade leather bags, your seeds might be "handmade leather bag", "full grain leather tote", "leather bag care". No volume data, no difficulty scores. Just language.
This step matters because tools only return variations of what you feed them. Garbage seeds in, garbage list out.
Step 2: Expand with free sources first
Google's own autocomplete and the "related searches" block at the bottom of a results page are still the fastest free expansion method. Type your seed, add a space, and read what pops up. Then check the tools.
| Tool | Best for | Cost reality |
|---|---|---|
| Google Keyword Planner | Broad volume ranges, ad-focused ideas | Free with a Google Ads account; volumes are banded, not precise |
| Semrush | Competitor keyword gaps, difficulty scores | Paid; the free tier caps you at a handful of searches per day |
| Ahrefs | Backlink-informed keyword ideas | Paid; strong data, heavier learning curve |
| Answer platforms and forums | Real phrasing, real problems | Free, and underrated |
My honest take: for your first three months, free sources plus a paid tool on a trial basis will get you further than a subscription you don't know how to read yet.
Step 3: Read the intent before you read the numbers
Search any candidate keyword and look at what ranks. If the top results are all product pages and you're planning a blog post, you're fighting the page type, and you'll lose. If they're all listicles, write a listicle or don't bother.
Intent comes in a few flavours worth naming:
- Informational — the person wants to understand something.
- Commercial — they're comparing options before buying.
- Transactional — they're ready to act.
- Navigational — they want a specific brand or page.
Match your page type to the dominant intent and half your ranking battle is already won.
Step 4: Prioritise, then cut hard
This is the step beginners skip, and it's the one that matters. Rank your list on two axes: how badly you want it, and how likely you are to win it. Then only keep the upper-right quadrant.
I use a blunt rule. If a keyword is more than four words long, has a difficulty score I can plausibly beat with a good page, and I can write something genuinely original about it, it goes on the shortlist. Everything else goes in a "someday" file I almost never open.
Quick wins versus long-tail, in plain terms: a long-tail phrase might bring in twenty visits a month. Ten of those phrases bring two hundred. That's how a new site builds momentum — not by chasing one big term.
Step 5: Map keywords to pages
One primary keyword per page. Related variations live on the same page. That's it. The most common beginner error I see is two or three pages targeting near-identical queries, which splits your own signal and helps nobody. Pick a winner, merge the rest.
Step 6: Write, publish, and measure
The research isn't finished until something is live. Set a reminder for six weeks out and check whether the page has started appearing for its target query. If it hasn't, the problem is usually intent mismatch or thin content — not the keyword itself.
The mistakes I made so you don't have to
Let me be specific, because vague warnings don't stick.
- Chasing volume. My first "big" keyword had massive search numbers and no realistic path to page one. I wasted most of a quarter on it.
- Ignoring intent. I wrote a thoughtful explanation piece for a query where everyone wanted a comparison table. It never ranked.
- Cannibalising my own pages. Two articles, one query, neither ranking. Merging them fixed it in about a week.
- Researching forever. The spreadsheet with forty-seven tabs produced zero published articles. That's the failure mode I see most often.
Notice the pattern? None of these were tool problems. They were judgement problems, and judgement only improves by publishing and watching what happens.
How often should you revisit your keyword list?
Quarterly is plenty for most beginners. Monthly is overkill unless you're in a fast-moving niche. What I actually do: a quick pass every three months to check which target pages have stalled, and whether any new conversational queries have appeared around my topic. Fifteen minutes, not a weekend.
Do I need a paid keyword tool to start?
No. Google Keyword Planner, autocomplete, related searches and forum threads will carry you through your first several months. Upgrade when the manual work becomes the bottleneck, not before.
How many keywords should I target on one page?
One primary query, plus closely related variations that share the same intent. If you find yourself wanting two clearly different intents on one page, split it into two pages.
What most beginner guides still leave out
Two things, and they're the ones I'd want to know from day one.
First, prioritisation. Every guide tells you how to find keywords. Almost none tell you how to choose between them, which is the actual skill. A list of two hundred keywords is worthless without a rule for cutting it to five.
Second, the click economy has shifted. Definitions and quick facts increasingly get absorbed into the results page or an AI answer. The queries that still reward a well-written page are the ones where someone needs reasoning, a real example, or a decision. Build your list around those and you're future-proofing without really trying.
The part nobody tells you
You will pick a keyword, write the best page you can, and watch it do nothing for two months. Then it moves. Then it moves again. That delay is normal and it's the reason most beginners quit — not because their research was wrong, but because they stopped before the compounding started.
So here's the question worth sitting with: if you only had time to research five keywords this week and publish one page, which one would you pick? Answer that honestly and you've already done the hard part. The rest is just typing.
