The Amazing Bandit MaxDiff

Traditional MaxDiff is good for studies involving up to about 30-40 total items. Once you get to 40 or more items, it becomes hard to get enough information from each respondent on each item. An innovative and adaptive approach can help! If the goal is to identify the top item (or handful of items) for the population, it doesn’t make sense to ask about the less important items as much as the most important items. Bandit MaxDiff learns from past respondents to oversample the more preferred items for the next respondents. Sample size savings can be 4x or greater compared to standard MaxDiff. Stop throwing your money away on data collection and embrace this new intelligent approach for using MaxDiff with large lists of items—even 100 or more!