Keyword Research for Beginners: Find a Topic Before Buying a Tool

What keyword research for beginners actually means

Keyword research for beginners is writing down the exact questions your reader types into a search box, checking what already ranks for each one, and choosing the question you can answer better than anyone else. You need no paid tool to start: a notes file, a search engine and your own site data are enough.

This lesson is for you if you are building a small affiliate site in evenings and weekends, you cannot justify a monthly subscription yet, and guessing topics has already cost you weeks of writing that went nowhere. It is not a list of tools, and it will not promise rankings. It is the order of work that makes any later tool purchase a decision rather than a gamble.

By the end you will have finished four things, in this order:

• a list of ten specific reader questions, grouped by intent, with one article outline ready to write;
• one existing page improved against the promise its title makes, with a saved baseline;
• a named information gap you can explain better than the pages already ranking;
• a written upgrade trigger for paid research tools — or a clear decision to keep using free methods.

Work through the sections in order and keep the results in one document. Our free checklist follows the same four sections, so you can tick each completion check as you go. If you want the wider picture first, the companion lesson on keyword research with free tools only covers the same free methods from a different angle.

Build a list of ten specific questions

A volume score tells you how many people search a term; it does not tell you whether your page will help one of them. A specific question does, because it tells you what the page must deliver. Google's guidance is to think about the words a user might search for, and it notes that people who already know a topic use different terms than newcomers ^[1].

1. Pick one reader and one problem. Write it as a single sentence: "someone who just signed up for my email list cannot find the download." One reader, one problem, no "everyone".
2. Write it in your reader's words. Use the phrasing a beginner would type, not the vocabulary of the industry. Keep both versions when they differ, because a newcomer and an expert often search differently ^[1].
3. Turn that problem into ten questions. Aim for questions that could each become one page. Ten is a working list, not a target to pad: if you only have seven honest questions, search those seven and add more as you learn.
4. Search each question once and note the format. Write down whether the results are tutorials, comparisons, product pages or forum threads. Do it yourself rather than trusting a summary — you are also studying what the reader will see.
5. Group the ten by intent. Sort each question into one row of this table, using the formats you just noted.

   Intent grouping for your ten questions

   Intent	What the question looks like	What the reader wants	Your output
   Learn a step	"how do I deliver a file after signup"	Steps they can follow today	Tutorial with numbered steps
   Compare two things	"X vs Y for a beginner"	An honest difference, not a winner	Comparison with stated criteria
   Decide whether to buy	"is X worth it when you are starting"	Cost, limits and alternatives	Decision guide with a "skip it if" line
   Fix a problem	"signup email never arrives"	A fix in five minutes	Troubleshooting walkthrough

6. Pick one question you can answer with a real example. Not the biggest number in a tool — the one where you have notes, a mistake you made, or a test you can run this week.
7. Outline that article in five lines using this fill-in line as your template: Reader: ____ | Problem: ____ | Question: ____ | My example: ____ | First step they take: ____

Illustrative example: a fictional beginner site about home espresso. The question "why does my espresso taste sour" returned tutorials and forum threads. Sorted with the table, the ten questions split into six step-by-step questions, two comparisons, one buying question and one troubleshooting question. The writer chose the sour-shot question because they had their own notes from fixing the same problem, not because any number looked large.

Done when: ten questions are grouped by intent and one of them has a practical article outline you could write this week.

Improve one page against its promise

Every page already made a promise in its title. Google's content guidance asks whether the main heading gives a descriptive, helpful summary, and whether a reader leaves feeling they learned enough to reach their goal ^[2]. Checking one page against its own promise is usually faster than writing a new one, and it teaches you what your future pages need.

1. Write the promise down. Read your title, then finish this line in one sentence: "This page says it will ______." If you cannot finish it, the title is the problem.
2. Read the page as a beginner, on a phone, before editing. Mark the first place the actual answer appears. If it sits below the third screen of scrolling, move it up.
3. Make the body followable. Numbered steps, in order, each one checkable by the reader without asking you a question.
4. Click every link. Replace anything that is dead, and point internal links at the page that helps next. The plan in internal linking for a 20-page affiliate site is enough for a small site.
5. Check the page can be indexed. Search site:yourdomain.com/page-url to see whether Google has it ^[5]. If nothing returns, look up that single URL with the URL Inspection tool rather than reading the whole index report; sites with fewer than 500 pages probably do not need the Page indexing report at all ^[5].
6. Add one worked example. A screenshot you produced, a filled-in form, or a before-and-after you can describe plainly. Pages that show work are easier to trust than pages that summarise other pages ^[2].
7. Save a baseline. Note the page's clicks, impressions, click-through rate and average position for the last three months if the site has data ^[4]. New sites may show nothing yet, which is normal rather than a failure.

Illustrative example: a fictional page titled "how to deliver a free download after signup". The promise was a setup in four steps, but the answer appeared in the ninth paragraph, one link was broken, and there was no example. The writer moved the four steps directly under the heading, replaced the broken link, added a screenshot of their own folder naming, and wrote down the page's impressions and average position as a baseline. Google notes that changes can take hours to several months to show, so they plan to wait a few weeks before judging it ^[1].

Done when: one page has a clear answer near the top, usable steps, checked links and a saved baseline.

Find a gap you can explain better

Three pages answering the same question show you the shape of a topic — and where the help stops. Google's guidance asks whether content offers substantial value beyond the sources it draws on instead of copying or rewriting them ^[2]. Your job in this section is to find one practical question the existing pages leave unanswered, then plan the explanation that fills it.

1. Take the question you chose from your ten. Keep it narrow enough that a single page can answer it completely.
2. Open the top three results in separate tabs and note the date you looked. Pages change, and your notes should be datable.
3. For each page, note four things: the examples it uses, the limitations it admits, the setup instructions it gives, and the practical question it never answers.
4. Circle the unanswered question a beginner would hit in their first week. Not an advanced edge case — something that stops a new person finishing the task.
5. Check you can answer it honestly from your own experience, a test you can run, or an official document you can cite. Experience and clear sourcing are among the trust signals Google's guidance describes ^[2].
6. Draft your answer as plain text before you build the page. Do not copy phrasing or borrow the order of someone else's steps; write the sequence you would actually follow.
7. Log the gap in your notes using this line: Question: ____ | Pages compared: ____ | Unanswered: ____ | My answer adds: ____
8. Treat tool estimates as estimates. A difficulty or traffic figure is somebody's model, not a count of the visitors your page will get. The gap you found is the evidence; the score is only a hint.

Illustrative example: a fictional writer compares three pages about restoring a website backup. All three list backup plugins and describe scheduling. None explains how to test that a restore actually works before you need it. That is the gap: a short walkthrough of restoring to a spare folder, what broke, and how long it took. The writer runs the restore themselves, records what happened, and writes the page around that record rather than around the other three articles.

Done when: you can name a specific information gap in one sentence and say, in one more sentence, how your page will fill it.

Set a reason to upgrade

Paid research tools are useful when they replace a repeated task you already understand. Before that point, a subscription answers questions you have not asked yet. This section turns your own research behaviour into a written trigger, so the decision to pay — or not — happens by rule instead of by mood.

1. Start a research log in a plain document. One line per session: Date: ____ | Question: ____ | Result formats: ____ | Gap or decision: ____ | Minutes: ____
2. Review the log after two weeks. Which job did you repeat most often? Which one did you avoid because it was slow?
3. Use your own search data first. Search Console's Performance report shows clicks, impressions, click-through rate and average position, and can be grouped by query, page, country or device ^[4]. Check it about once a month, or when you change the site's content ^[3].
4. Write the trigger in one sentence, filling the blanks honestly: "I will consider paying for a research tool when ______ happens ______ times a week and I have published ______ pages using the free method."
5. If you trial a paid tool, put the cancellation date in your calendar the day you start, and complete one real research task with it before that date. A trial you never used is not evidence of anything.
6. Record the decision either way: keep, drop, or revisit in a stated number of weeks. Staying free is a decision, not a delay.

Illustrative example: a fictional marketer logs four research sessions a month: two question lists, one comparison of three competing pages, and one indexing check. After four weeks the log shows the same competitor comparison repeated three times at roughly forty minutes each. That repetition — not a headline number in a tool — is what fires her trigger. She has also published four pages with the free method, which is the second half of her written condition.

Done when: you have a written upgrade trigger or a recorded decision to keep using free methods for now.

Tool decision: free route, upgrade trigger and what to research

Start with what is free or already in your account. Manual searching covers every step above, and Search Console is free: it shows how Google crawls, indexes and serves your site ^[3]. A new site may have almost no search data yet, which is a reason to wait rather than a reason to buy.

Your upgrade trigger. Consider a paid research suite only when a research or auditing job you repeat takes long enough to justify an ongoing cost, and when you already have a publishing process that will use the output. If you cannot name the task the tool would do this week, keep using the free route. That is not a compromise; it is the correct order of operations.

Candidate tool categories to research.

• Your own search performance: the Performance report for queries, pages and click-through rate ^[4], plus the indexing reports for whether pages can be found ^[5].
• Manual topic-fit research: search the reader's question yourself and study the result formats ^[1].
• Paid research suites: Semrush and Ahrefs are names to research, and this lesson claims nothing more than that. Confirm the task, the access limits and the current terms on their own pages before you pay; we state no price or plan here.
• Content quality checks: Google's people-first questions make a free pre-publish checklist — original information, clear sourcing, and a reader who leaves satisfied ^[2].

On trusting a guide like this one. You are right to ask whether a site is just pushing whatever pays the most. The four tasks above work with free access, and nothing here depends on a purchase. Where a product is named, treat it as a lead to check, not a recommendation — this lesson describes no hands-on tool trials, and it should not be read as if it does. Check the date on any guide, including ours, and check the tool's own page on the day you sign up.

Next step. When your trigger fires, compare options on our keyword research tools comparison and the head-to-head Semrush vs Ahrefs decision page. If the site itself is still the weak link, start with SEO for affiliate marketing. Background reading that pairs with this lesson: matching search intent to the query, the Search Console reports worth checking weekly, what Google's helpful-content guidance means for small affiliate sites, and keyword research with free tools only.