TL;DR
A Category Entry Points study maps the real situations that bring people into a category, using the mental-availability method of walking the seven W's of each occasion: when, where, with whom, doing what, how feeling, why, and how. This template covers spontaneous occasion recall with a breadth loop, an aided mop-up list, one deep-dive episode, and a light look at unmet occasions. Use the plain-text template at the end to run this study AI-moderated across any category.
What Category Entry Points research actually tells you
Most category research asks what people buy. A CEP study asks what situation brought them into the category in the first place.
The method, built on the Romaniuk mental-availability approach, focuses on breadth rather than depth. It maps as many distinct entry occasions as possible, each characterized by the same seven dimensions, so a brand can see the full spread of moments where the category gets entered, not just the most common one.
The output is a map of situations, not products. It tells you when and why someone reaches for a category, which is the foundation for building mental availability across the moments that matter, not just the ones a brand already owns.
What this template covers
Five questions across one section, structured around a strict sequence. The logic flows in four stages:
Warm up and frame the conversation around entry moments, without collecting data yet
Capture spontaneous occasion recall, looped repeatedly for breadth
Run an aided mop-up list to catch occasions not mentioned unaided
Deep-dive on one standout occasion, then close on unmet needs
Question two is the heart of the study. It opens on the most recent occasion, walks the seven W's, then loops repeatedly, asking for another occasion that was different again, continuing until the participant runs dry. This breadth loop is what separates a CEP study from a standard occasion question.
Question three is the aided mop-up, and it must run after spontaneous recall, never before. Question four then goes deep on a single standout occasion, walking all seven W's thoroughly rather than lightly. Question five closes on two distinct types of unmet need: occasions where nothing satisfied, and occasions where no suitable option was available at all.
How to set up this study for your category
A few stetup rules:
Swap the category throughout, but preserve the sequence exactly: spontaneous recall must always come before the aided list
Never introduce a brand, product, or tool name during the situational questions. This study maps triggers and occasions, not solutions
Build your aided list to be mutually distinct and reasonably exhaustive, covering situations across a full day or week
Where CEP research goes wrong
Three failure modes to watch for:
Showing the aided list too early. If situations are suggested before spontaneous recall, you lose the breadth of the reading and anchor participants to the researcher's frame instead of their own.
Stopping the breadth loop too soon. The value of this method is in the range of occasions surfaced. Keep looping for genuinely different situations until the participant clearly runs dry.
Conflating the two unmet-need mechanisms. "Considered but unsatisfying" and "no suitable option available" are distinct gaps with different implications. Always separate them rather than treating unmet need as one category.
Running Category Entry Points studies with Conveo
Conveo's AI moderator manages the breadth loop automatically, recognizing when an occasion has been well described and prompting for another that is genuinely different, without repeating occasions already covered.
The moderator holds the spontaneous-before-aided sequence strictly, so occasion breadth is never contaminated by early suggestion
Follow-up depth shifts automatically between the light aided scan and the deep single-episode reconstruction, matching each question's intent
Thematic analysis clusters occasions across participants by situation type, surfacing the full occasion map without manual coding
Want to see how this template runs inside Conveo?
