2 min readRivo

What changes when your customers stop searching

A search engine returns a list and lets the customer choose. An assistant returns an answer and makes the choice on their behalf. That is a small difference in interface and a very large difference in what determines whether your brand is part of the outcome.

Ranking is positional. There is a query, a set of documents, and an ordering, and you can watch your position in that ordering move. Being recommended is generative: the model produces a description of a category and names the brands it can describe confidently within it. There is no list to be tenth on. You are either in the answer or you are not mentioned at all.

This is why a page that ranks well can still be invisible, and why ranking and AI citation overlap far less than people assume. The ranking signal lives in the index; the recommendation signal lives in what the model has absorbed about your brand (from your own content, but also from retailers, reviews, comparisons, forums and structured data you do not control).

What that means for measurement

The metric that matters is not a position. It is closer to a share: across the range of ways a real customer might describe their need, how often does your brand appear, in what terms, and against whom?

That question cannot be answered by checking a handful of prompts by hand. The phrasing space is enormous, answers vary between engines and between runs, and the interesting cases are the ones nobody thought to type. The research on variance is unambiguous that sampling design decides whether the number means anything.

Where to start

Three things are worth knowing before anything else:

  1. Presence. Across the descriptions your customers actually use, how often are you named at all?
  2. Perception. When you are named, what is said? Is it what you would have said?
  3. Displacement. Who appears when you do not, and what are they doing that the model finds easier to describe?

Those three are worth defining precisely before anything else, because each points at different work: presence and perception often turn on what the model retrieved versus what it already believed, while displacement is usually decided by pages you do not own.

The gap between those three answers and the story you tell about your brand is usually where the work is.