# Question tiers in AEO: buying, preparing and facts

> In short: Tiers group tracked questions by how close the person asking is to buying. Tier 1 is choosing ("best SIE prep course"), tier 2 is preparing ("how long to study for the SIE"), tier 3 is facts ("how many questions are on the SIE"). Buying questions carry the most weight and the highest aim.

Definition from Keats, published 2026-11-12, updated 2026-11-12. Canonical: https://getkeats.com/glossary/question-tiers

Some questions sell and some don't. Someone asking "which prep course should I buy" is choosing; someone asking "how many questions are on the exam" may never buy anything. Tiers keep the two from being added up as if they were the same.

Keats uses three:

| Tier | The person is | Example | What winning means | Aim | Weight |
|---|---|---|---|---|---|
| 1 | Choosing | "best SIE prep course" | Named, ideally first | Named in 80% of answers | 3 |
| 2 | Preparing | "how long to study for the SIE" | Named | Named in 60% | 2 |
| 3 | Looking up a fact | "how many questions are on the SIE" | Your page cited | Cited in 25% | 1 |

The aims differ because the outcomes differ. On a buying question you want your name in the answer, and [named first](https://getkeats.com/glossary/named-first) is better still. On a fact question the assistant rarely names a brand at all, so the win is being the page it cites.

The weights feed the [AI visibility score](https://getkeats.com/glossary/ai-visibility-score): a tier-1 answer counts three times a tier-3 one. A sensible set has 20 to 30 questions per thing you sell, at least ten of them in tier 1, written the way buyers ask them.


## Sources

- [Keats: llms.txt, the operations and workflows for agents](https://getkeats.com/llms.txt), read 2026-11-12