Study context
A leading snack brand developed and tested eight distinct new concept ideas with our AI synthetic panel to identify high-potential innovations. They compared the scores for Purchase Interest, Newness & Uniqueness, and Meets a Need to trial benchmarks. Our newer Virality and Overall Score was useful to discriminate between top-performers.
Process steps
- Concept generation – Eight snack formulations developed by our tool from trend signals that emerged over the past 6 months.
- AI synthetic panel – Each concept scored on five metrics by the digital twin — the equivalent of N=1,000 modeled shoppers — in under five minutes.
- Tangible takeaways – Team reviewed the scores on our dashboard, and was immediately able to shortlist the top two ideas for development.
Key findings
By metric:
- Overall Score: the top concept scored 78% Top-2 Box (“TTB”) on our 5-point (Likert) scale. The #2 performing concept scored 72% TTB.
- Purchase Interest: ranged from 63–70% TTB across all five concepts.
- New & Unique: the best idea scored 80% TTB indicating a high degree of novelty.
- Solves a Need: scores averaged 75% TTB, indicating strong problem fit.
- Virality: the highest concept reached 72% TTB, suggesting solid peer-to-peer sharing potential.
Impact for the brand
The tool conceived and validated two high-potential concepts several months earlier than they’d get from their usual process. This gave them confidence to focus resources on the two top ideas and move to the formulation stage months ahead of time. Having later conducted an in-home usage test on the prototypes, they plan to bring both to market in early 2026.
Quote from Insights Lead
“This process was fast …(and) gave us clear direction on which flavors to pilot, eliminating guesswork and accelerating alignment across departments.”