⏰ The 30-second version
- GEO doesn't start with writing — it starts with building a question list. If you don't know what they're asking, you have nothing to answer.
- There are three question warehouses — your reviews, customer inquiries, and search autocomplete. All free, all real data.
- The frame that fills them in without gaps has five axes — selection criteria / comparisons / side effects & cautions / how to use / price.
- The target is 30. Sounds like a lot, but it's just five axes × six each.
Warehouse 1 — Reviews: your customers are already writing the answers
A review is an answer to a question. Behind "less dry and chalky than I expected" sits the question "Is there a chicken breast that isn't dry?" Skim just 50 of your own reviews and ten questions surface immediately. Competitor reviews are even better — especially negative reviews, which show exactly what customers worried about before buying, which is exactly what they'll ask AI.
Warehouse 2 — Inquiries: the purest question, verbatim
Customer-support inquiries are questions in the customer's own words. "Is this safe during pregnancy?" "Can I put it in the washing machine?" Don't polish these — drop them into your question list almost as-is. Customers ask AI the same way.
Warehouse 3 — Autocomplete: real demand, straight from the search box
Type a category keyword into Google's search box and autocomplete follows. These are words people actually searched. Type "probiotics" and you get "probiotics when," "probiotics side effects," and so on. Anything in autocomplete is a pre-validated question.
Filling the blanks with five axes
Place the questions you gathered from the three warehouses onto the five axes, and the empty axes reveal themselves.
The five question axes (six each = 30)
- [ ] Selection criteria — what should I look at to choose?
- [ ] Comparison — how are A and B different?
- [ ] Side effects & cautions — should certain people avoid this?
- [ ] How to use — when and how do I use it?
- [ ] Price — is this a fair price?
In my experience, sellers gather "selection criteria" and "how to use" well, but the comparison and price axes often sit empty. The more sensitive a question feels — the kind you'd rather avoid — the more your customer asks AI instead. If you don't answer it, someone else will.
Once you have 30
That list is your content roadmap. Measure how many of them your pages answer (self-check) and you get your priorities too. For reference, Bulti's free audit automates this whole process — feed it your product keywords and it builds 30 category-tailored questions and measures your coverage.
FAQ
Q. Do I really need 30 questions? Can't I start with 5 core ones?
You can start with 5. But customer situations vary, so the long tail of questions runs deep. Around 30, you'll always uncover a question you didn't know about — and that's where the competition is empty.
Q. Don't questions change with the season or trends?
They do. So a question list isn't a one-and-done — it's best to re-skim autocomplete each quarter and refresh it.
Q. I gathered the questions but don't have time to write.
You don't have to write them all. Start with the axes closest to a buying decision (comparison, selection criteria), one post at a time. One question per post is the basic unit.
Originally published at en.bultibulti.com. We run a free GEO audit there — it checks one of your pages against 30 buyer questions, no signup.
Top comments (0)