Predictive Research for Online Market and Product Testing
Sampling for PROMPT
We have seen with the Iowa University election forecasts that prediction markets do not require traditional sampling in the target segment. Rather, they thrive on knowledge from multiple perspectives.
Generally people will be able to contribute to an improved forecast if they either use a product (or a substitute) in the category or if they can observe others who do. The number of necessary participants is much smaller than for frequentist methods: prediction markets reach validity after 16 knowledgeable respondents, and after more than 30 there is no significant improvement from adding further participants.
Therefore, Prediki PROMPT is suited for a broader range of projects, including innovative and niche products or business-to-business innovations, whereas STM is mostly useful for incremental changes or re-launches of fast moving mass market consumer goods (FMCG).
When confidentiality requirements are high, Prediki PROMPT can give good results even if participation must be strictly limited to employees such as sales, customer care, marketing and PR staff. Best results are achieved when including in-house product users or channel partners.
Next: Context for PROMPT