Skip to content
AEO Canon · the reference for answer-engine optimization
AEOCanon

Lesson 2 of 6

Measure Per Engine, Not a Blend

Engines overlap only about 11%, so a blended average hides where you stand. This lesson measures per engine.

Learning objectives

  • ▸Explain why blending misleads.
  • ▸Report share of voice per engine.
  • ▸Use a weighted index only as a headline.

The lesson

Read the full lesson →Why Per-Engine Measurement Beats a Blended AverageMeasure AI visibility per engine, not as one blended average. Profound found only ~11% citation overlap across engines, so a single "AI visibility" number averages together separate universes — hiding where you actually win or lose. Here's why per-engine tracking is the only honest read.5 min read

Key takeaways

  • ▸Profound found about 11% cross-engine citation overlap.
  • ▸A blend can hide a one-engine collapse.
  • ▸Decide from the per-engine view.

Knowledge check

Knowledge check

0 / 2

  1. 1. Why measure per engine?

  2. 2. When is a blended number acceptable?