Answer Engine Optimization

Chapter 5. Measuring Performance

I’m just going to be real straight with you here. Measuring AEO performance right now is kind of a mess. Actually, that’s being generous. It’s a total mess.

In traditional SEO, we had it so good, and honestly, I don’t think most of us appreciated it at the time. Google Search Console told you exactly which queries drove impressions and clicks. Analytics showed you the traffic. Rank trackers showed you the positions. You could build a decent dashboard in an afternoon and feel reasonably confident that the numbers meant what you thought they meant. It was clean. It made sense. You could point to a chart in a meeting and say, “This is working,” and people believed you because the data backed it up.

AEO? Not even close.

I’ve literally asked ChatGPT the same question five times in one day and gotten five different sets of cited sources. Five! Same question, same day, five different answers about who the “best” sources are. Perplexity passes pretty clean referral data (bless them for that), but ChatGPT’s attribution is spotty at best, Claude barely shows up in most analytics setups at all, and Google AI Overviews is cannibalizing clicks in ways that are genuinely hard to untangle from normal organic fluctuations. And here’s the real kicker. The platforms that are most transparent about citations (Perplexity) represent the smallest slice of total query volume, while the ones with the most volume (Google, ChatGPT) are the absolute hardest to measure.