Confectionery concept screening

Breadth without losing depth

A leading European confectionery brand was evaluating twenty-four candidate concepts for a major line extension. The industry standard for this problem - sequential testing through external research partners, three or four concepts per round - would have required six months and produced findings whose usefulness degraded with every round of delay, as market context shifted under them. Internal teams were stuck in a familiar pattern: concept pruning by committee, gut feeling, the loudest voice; consumer evidence trailing months behind the decisions that needed it.

The constraint was not the quality of the research method. It was the architecture of the process. Traditional concept testing was built for a world in which consumer evaluation was expensive, slow, and capacity-limited. The design of a modern line extension - parallel, iterative, compressed - has outgrown that architecture. The question was whether a depth-psychological read could hold across twenty-four parallel concepts without losing the fidelity that makes it worth doing in the first place.

How we worked it

We built four consumer simulations, each anchored in a distinct psychological position within the brand's core audience. The critical move was the anchoring. Each simulation was constructed on the foundation of an eighteen-person depth-psychological study with real consumers in this category - not on demographic data, not on segmentation output, not on a language model's generalized notion of what a chocolate buyer sounds like. The simulations inherited the contradictions, defenses, and emotional patterns that real people brought into the interview room. Their psychological texture was inherited from human beings.

Twenty-four concepts ran in parallel through two iterative rounds. The simulations participated in structured depth interviews and small-group dynamics, functioning inside each round as both evaluators and collaborators - surfacing weaknesses in each concept, proposing structural refinements, ranking ideas by psychological resonance. Between rounds, the weakest concepts were eliminated and the strongest were sharpened against the simulations' feedback. The screening was not a sort. It was a conversation between twenty-four concepts and the psyche of the category, conducted at a pace the category had never previously allowed.

What emerged

Four concepts - validated, sharpened, and ready for development. The compression was substantial (weeks rather than months), but the result that mattered was the next one. A follow-up depth-psychological validation study with real consumers confirmed the simulations' findings. The simulated readings aligned with the human readings on the concepts that won, the concepts that lost, and - most significantly - the specific reasons why each concept worked or failed. The instrument had not produced a plausible approximation of a real consumer read. It had produced a consonant one.

That final detail is the one worth naming directly. The value of this engagement was not that it replaced traditional research. It was that it rendered the sequence - months of staged testing followed by pilot research followed by validation - unnecessary, because the inner work had already been done against a psychology grounded in real human material. The brand moved from twenty-four candidate ideas to four launch-ready concepts in the time a single traditional testing round would have taken.