Keyword research means finding out which words and questions your customers use when they look for what you offer. The best source is the questions customers already ask you. You add queries from Google Search Console and suggestions from Google itself. Then you work out what the searcher wants for each question and match every question to one page on your website.
This guide works without paid software. Each step says what you do and what you end up with.
The plan at a glance
- Your offerWrite down what you sell and to whom, in your customers’ words.
- Customer questionsCollect the questions customers already ask you.
- Search ConsoleSee which searches your site already appears for.
- Google itselfUse suggestions, follow-up questions and the pages at the top.
- Search intentWork out what the searcher wants and which kind of page fits.
- One page per questionGroup the questions and match each group to one page.
- AI assistantsWrite down the longer questions people ask an assistant.
Step 1: What do you sell and to whom?
List your services or product groups. For each line, note who it is meant for and what problem it solves for that customer. Use your customers’ words, not your internal product names. An installer might sell a “hybrid heat pump system” while the customer searches for “heat pump alongside gas boiler”. You will not find a keyword like that in your own catalogue, but you will in your inbox.
Result: a list of topics you want to be found for.
Step 2: Which questions do customers already ask you?
The questions customers ask before they buy are often exactly what other people type into Google. Ask your customer service and sales teams what they hear most often and what customers hesitate about. Doubts about price, delivery time, suitability or the difference between two options are valuable, because good answers to them are hard to find.
Where to find customer questions
- Emails and chats with customers
- Notes from phone calls and quote meetings
- Questions that come with returns and complaints
- Reviews, including those of other businesses in your sector
- Questions your sales staff or engineers answer again and again
- Forums, trade groups and product Q&A sections
- Comments under videos about your field
Forums and video comments show how people describe a problem before they know which product they need. Those questions come from the research phase, before someone has a brand in mind. They are often long, specific searches; see long-tail keyword.
Result: a rough list of questions in your customers’ own words.
Step 3: Which searches do you already appear for in Google?
In the Performance report in Google Search Console, the Queries tab shows the searches for which your site appeared in the results. For each query you see impressions, clicks, the click-through rate and the average position.
Pay particular attention to:
- queries with many impressions and few clicks. Google finds your page relevant, but the title or the answer does not persuade;
- queries for which an unexpected page appears. You may be missing a page that answers that question directly;
- questions starting with who, what, how or why. They show which explanations people are looking for.
Google notes in that documentation that a query appearing in your list does not mean you will see your site when you run the same search yourself. Results differ by person and place.
Result: queries you are already visible for, and chances to improve existing pages.
Step 4: What does Google itself show?
Type your topics into Google and watch what happens:
- Suggestions as you type. According to Google Search Help, these predictions are based on searches people have actually made. Google warns that they are not simply the most popular searches, so you cannot read search volume from them.
- People also ask. This box of expandable questions shows which follow-up questions belong with a topic.
- Related searches at the bottom of the results page.
- The pages that rank at the top now. Are they product pages, comparisons or explanations? That says something about what Google thinks the searcher wants.
replace e-bike battery
- bikeshop-example.co.uk › advice › batteryWhen should you replace your e-bike battery?How to recognise a battery that is wearing out and what to look for in a new one.
- batteryshop-example.co.uk › batteriesE-bike batteriesBatteries for common motor systems, with notes on compatibility.
- cycling-forum-example.co.uk › maintenanceReplace or refurbish a battery?Cyclists share their experience of replacing and refurbishing.
How much do people search?
Keyword Planner in Google Ads gives an estimated number of monthly searches per keyword, plus new keyword ideas. You need a Google Ads account in which you have entered billing information. For keywords with very low volume, the tool sometimes returns no result.
Use volume to sort, not to cut. A question that few people search for but that comes up just before a purchase can bring in more than a broad term with high volume.
Result: variants and follow-up questions you had not thought of yourself, and an idea of which topics people search for most.
Step 5: What does the searcher want with each question?
Search intent is what someone wants to achieve with a search. Two searches about the same product can call for very different pages.
| Intent | Example | Page that fits |
|---|---|---|
| Find out | “how long does an e-bike battery last” | Explainer or resource article |
| Compare | “e-bike mid-drive or front hub motor” | Comparison or buying guide |
| Buy or book | “buy 500 Wh e-bike battery” | Category or product page |
| Local | “e-bike repair Zwolle” | Service page with location details |
For comparison questions, read on in writing comparisons and buying guides; for buying questions, in optimising category pages.
Result: every question has a label for the kind of page it needs.
Step 6: How do you match each question to one page?
Put questions that want to know the same thing together. Each group gets one page: an existing page you improve, or a new one. That stops two pages on your site from competing for the same search.
A simple overview looks like this:
| Main question | Variants | Intent | Page | Status |
|---|---|---|---|---|
| Which e-bike battery fits my bike? | replace battery, compatible battery | Buy | /batteries | Improve existing page |
| How long does a battery last? | battery lifespan, battery charge cycles | Find out | New article | To be written |
Questions that can be answered briefly and belong with an existing page often fit in an FAQ on that page. How to set one up is covered in an FAQ that helps customers, Google and AI assistants.
Result: a plan that says, for each page, which question it has to answer.
Step 7: Which questions do customers ask AI assistants?
Some of your customers ask an AI assistant instead of typing a search term. Such a question is usually longer and includes the person’s situation. Take a bike shop in Zwolle, a city in the east of the Netherlands, as an example.
The same need, two ways of asking
| Search in Google | Question to an AI assistant | |
|---|---|---|
| Example | e-bike hire Zwolle | I want to spend a weekend cycling around Overijssel and I don’t have an e-bike. Where can I hire one in Zwolle and what should I bring? |
| Length | A few words | One or more sentences |
| The person’s situation | Usually not mentioned | Plans, wishes and constraints are included |
| What you use it for | Deciding a page’s title and topic | Checking that your page gives the full answer |
Write down a longer version like that for your most important main questions. You use the list in two ways: to check that your page gives the full answer, and later to see whether and how your business appears in AI answers. How that measuring works is explained in how do you measure your visibility in AI answers? How to write pages AI assistants can use is covered in writing content that AI assistants can cite.
Then keep the list up to date. Your offer changes and customers come up with new questions. Look in Google Search Console again regularly and ask your customer service team what is new. Add questions to your overview and mark which pages need an update.
Result: a fixed list of questions for your content and for later checks.
How does Outseen do this?
Outseen starts an SEO or GEO project with the questions around your offer. We research which questions customers ask, which pages already answer them and where information is missing. A writer then works from your briefing and sources, a reviewer checks the text, and you give feedback before anything is published. More on the approach is on the SEO page.
What does this mean for your business?
You do not need software or an agency for a first round. Set aside an afternoon, collect the questions from your email and customer conversations of the past few months and put them next to the queries in Search Console. The questions that keep coming back and do not yet have their own answer on your site are your first pages or FAQ items.
The overview of main questions and pages then becomes your work plan. It shows which page you improve first, which one needs to be new and which questions you will later use to check how AI assistants talk about your business.