Question Research for AEO: Find What Customers Ask

Your own records are the most valuable source. Customers put these questions to you because they were trying to understand your business, and their wording has not been filtered. Every list bought from outside is a guess; your records are an observation.

Ranking the sources by value speeds the work up. Look inside first and outside second, and use the outside sources to confirm and widen what you found inside.

• Sales calls, quotation emails and appointment notes
• Support records, WhatsApp and phone messages, replies to reviews
• Phrases typed into your on-site search box
• Autocomplete suggestions from search engines
• Question boxes such as People Also Ask
• Industry forums, community groups and social media comments

Open the last few months of correspondence and copy only the question sentences into a separate list. Take the question, not the answer, and leave it as the customer wrote it. Typos, abbreviations and half sentences are information too, because people type and speak to search boxes and voice assistants in the same way.

Ask your sales team one thing: which objection do you meet most often in a conversation? An objection is usually a hidden question. Behind the sentence 'it seemed expensive' there is often the question 'what does this price include?'. Picture a flooring manufacturer: what its dealers keep asking is not the price list but which product suits a damp subfloor. That question shows up in the records, and if it is missing from the site it is missing from answer engines too.

Autocomplete suggestions and question boxes show which questions people tend to ask around your subject. They are a hint, not a measurement: what these boxes show changes over time and from person to person, so do not treat one day's view as settled. Try the same subject with several different phrasings and mark the questions that keep coming back.

Forums and community groups give you the feeling behind a question. People drop the industry vocabulary there and state their worry plainly. Picture a spa group: suggestions may show a short phrase such as 'spa prices', while a forum thread asks 'I am going for the first time, what should I bring and will anyone look at me?'. The second sentence is a far better question to write an answer for.

• Search your subject with three or four different wording patterns
• Mark recurring worries, not only recurring questions
• Write the source next to every question: record, suggestion or forum
• Do not decide from a single day's screenshot

Once you hold a few dozen raw questions, group them by the customer's path to a decision rather than by topic. Someone asking 'what is it?' about a product is in a different place from someone asking 'which one should I choose?', and needs a different shape of answer.

When the grouping is done, put one main question at the head of each group; the rest become variants of it. A page that tries to answer the questions of several groups loses its focus, so plan a separate answer page, or a separate page section, for each main question.

• Definition questions: what is this, what is it for
• Comparison questions: which one, what is the difference
• Decision questions: is it right for me, when is it needed
• Process questions: how is it done, how long does it take
• Trust questions: who does it, what does it cover, what if it goes wrong

You do not have to weigh every question equally. Look at three measures: how often the question arrives, whether its answer directly affects your business, and whether your site already gives a clear answer today.

Put at the top the questions that arrive often, sit close to a sale and have no answer on the site. Where an answer exists but is vague, rewriting it usually costs less effort than opening a new page from scratch. When you write the answer, the site's article on how to write quotable content will help you build sentences that can be lifted out whole.

• Frequency: how often it turns up in your records
• Impact: how close the answer sits to the buying decision
• Gap: whether the site gives a clear answer today
• Effort: a new page, or a fix to an existing one

The first mistake is inventing the questions from your own assumptions. A list written at a desk under the heading 'customers will wonder about this' usually holds the questions the business wants to answer, not the ones customers ask. The second mistake is cleaning the question up: once the customer's words are translated into industry terms, the page drifts away from the phrase people actually use.

The third mistake is building the list once and leaving it. Customer questions move with products, seasons and competitors. A short monthly refresh is enough: add the new questions, close the ones that now have answers. The fourth mistake is assuming answer engines will pick you. No research guarantees that; it only secures the first condition of being selectable, which is answering the right question.

If you have no customer records and the business is new, a heavy research set-up is pointless; noting down the questions of your first ten customers by hand is enough.

There are also subjects where the answer is already single and fixed: opening hours, address, contact details. These need a clear and current page, not research. Keep research for the questions that have more than one possible answer and leave the customer undecided.