METRICS
Average Position

IN ONE SENTENCE
Average position is a brand's mean placement across the AI answers that mentioned it; answers with no mention are excluded from the calculation.
AI answers usually return a shortlist. Average position measures where a brand normally sits in it. It is computed only across answers that mentioned the brand — an answer with no mention has no position and must not be scored as last.
OUR POSITION
Below five samples the number is not usable and must carry its sample size. Reading a mean built from one or two samples as a trend is the most common failure with this metric.
How it is calculated
Take the position only in answers that named the brand and average them. The denominator is mentions, not total answers.
Decide the counting rule up front: list order, or order of appearance in prose. Answers with subheadings and answers that are plain lists must be read the same way, or two analysts will produce two numbers.
Always publish the sample size with the figure. Write 'average position 2.0 (n=2)', never 'average position 2.0'.
It is not search ranking
Search ranking is a page's slot in a result list. Average position is a brand's order of appearance inside generated text — there is no stable notion of 'page one' and no click-through curve behind it.
The payoff differs too: in search, position converts to clicks; in an AI answer, being named may complete the discovery step without any click at all.
When it is worth watching
Once mention rate is off the floor and the question becomes whether the brand is the default choice or a runner-up. While mention rate is near zero, work on mention rate.
Read it per platform. The same brand can sit in very different slots across platforms, and a blended number hides that.
Data behind this page
0.1
Monthly clicks per keyword ranking below position 10 in this category
Source:Our own keyword-level analysis of 4,074 non-branded keywords for a leading site in this category (13.1 for positions 1–3, 3.4 for 4–10),2026-06
Sources
- [1]GEO: Generative Engine Optimization.Aggarwal et al., KDD 2024.2024
Updated 2026-08-10