Design. Analyse.
Simulate the choice.
What would they choose if the offer changed, and which product should we put in market?
01 / Before fieldwork
Design
Attributes, levels, screens and versions. Combinations you forbid never appear, and the sample size names the rule that binds it. Already have a design? Upload your own, and the AI maps its columns for you to review.
Every design comes with a design validation report on efficiency, level balance and precision, so you know it works before a single respondent sees it.
See it in a demo02 / After fieldwork
Analysis
Fielded results go in as an SPSS (.sav) file, and the AI maps the columns for you to review. No data yet? A synthetic panel answers your design, so the whole workflow runs before fielding. Hierarchical Bayes gives each respondent their own utilities, and GPU acceleration makes it one of the fastest ways to run HB. Every run reports all diagnostic checks:
- Model fit (RLH)
- Holdout hit rate
- Parameter recovery on a synthetic panel
- Sampler acceptance rate
- Counts and an aggregate MNL benchmark beside the model
03 / Explore the possibilities
Simulator
Build products that were never shown and watch every option on the shelf respond. Test prices, search for the best configuration, size cannibalization, read willingness to pay, cut by segment, and hand the whole simulator to your client.
See it in a demoGive them a simulator, not a slide.
One CBC product: the design, Hierarchical Bayes analysis, and a simulator the brand team can keep using.