Segments

A segment in Consumr.AI is a definition of a consumer cohort: who these people are, what they care about, and what distinguishes them from other groups. Segments are the starting point for building AI twins and running research.

Segments are twins

The phrase you’ll hear internally: segments are twins. When you create a segment, you’re not creating a report or a label; you’re setting up the specification that will become an AI twin with memories and a persona. The segment definition is what gets translated into audience inputs, which generate reports, which assemble into a twin. The twin is the realized form of the segment.

A means to an end, not a key feature

Segments are not the product. They are a means to get to the product, which is the research. Clients used to traditional segmentation work will keep building segments indefinitely.

The right number for most use cases is 6 to 10 segments. That range typically covers 80–90% of the market while leaving the edge cases (outliers who don’t drive business decisions) aside. Going beyond that creates noise, not depth: too many segments answer similarly, and you end up reviewing twins that are functionally identical.

If clients push for 100 segments, that’s the moment to explain why stopping at 6–10 is the strategically sound call.

A business decision, not a research decision

The way a market is segmented reflects business objectives, not data outputs. Do you care most about age and life stage? Usage intensity? Price sensitivity? Platform affinity? Each framing produces a different set of segments, and all of them could be technically valid.

Established companies with dedicated marketing teams typically come to Consumr.AI with segments already defined. Smaller organizations may not, and Consumr.AI has tools to help them build definitions. But in either case, the marketing team (or business owner) decides what the segments mean before any research runs. The segments encode business intent.

Segments are audiences, not keywords

This is one of the most important nuances to communicate when onboarding clients. A segment is defined by interests, behaviors, and demographics, not by a search query or keyword set. Because segments describe audiences, they will naturally overlap. A “budget-conscious family” segment and a “value-seeking suburban shopper” segment will share many of the same underlying behaviors and interests.

That overlap also means that if you build too many granular segments, you’ll find that some share identical underlying data reports and therefore give nearly identical twin responses. The name will differ; the intelligence will not.

The “you have too many segments” objection has a data-level answer: go into the reports and you’ll often see that two segments’ reports are the same. See the Objections section for how to handle this in client conversations.