Message and concept testing
Build a MaxDiff or conjoint design, add screeners and follow-up questions, and analyze the results in the MX8 Labs Research Platform.
Use best-and-worst choices to rank competing messages. Generate task sets, check item exposure and export the survey block for programming.
Open the MaxDiff design tool →
MaxDiff tasks and results, with comparison bars and uncertainty.
See MaxDiff design and reportingCompare combinations of features, price or other attributes to understand trade-offs. Define the levels, generate task sets and add the exported design to your questionnaire.
Open the conjoint design tool →
From attributes and task design to part-worth results in the platform.
See conjoint design and reportingCombine the choice tasks with ratings or monadic testing where appropriate, then ask respondents to explain their reactions.
Compare MaxDiff utilities and modeled preference shares, or conjoint part-worth scores and attribute importance, across audience segments.
Generate a study from your concepts and questions in a Word document. Review monadic or sequential monadic design, rotation, quotas, and validation before launch; setup time depends on complexity.
Fielded across 70+ integrated panels with real-time fraud detection and attention checks before a response is accepted. Weighting and significance testing run automatically as data lands.
Compare concept rankings overall and by segment, with significance results and AI-coded open-ended reactions. Your researchers check the coding and AI-generated report drafts, then recommend which ideas to develop.
For a product idea, start with questions like these. Use the same wording and response scales for each concept you compare, and adapt the questions when testing a service, message, or pack.
Ask respondents to explain their ratings so you can understand the reasons behind them. For a concrete message comparison, read our message testing guide.
Concept testing puts an idea — a product, a claim, a pack, a piece of creative — in front of real consumers before you invest in building or launching it, and measures appeal, relevance, uniqueness, and purchase intent against alternatives or a benchmark. It helps you compare options by segment and understand the reasons respondents give for their preferences.
In monadic testing, each respondent evaluates one concept, so their ratings do not depend on concepts shown earlier in the survey. Sequential monadic testing asks each respondent to evaluate several concepts. It uses fewer respondents for the same number of ratings per concept, but earlier concepts can influence later answers. MX8 Labs supports both designs, with rotation and quotas. Choose the design around your comparisons and available sample.
A typical core: overall appeal, relevance, uniqueness or differentiation, believability, purchase or usage intent, and open-ended likes and dislikes — plus category-specific measures like price sensitivity (Van Westendorp or Gabor-Granger) where the decision needs them. Open-ended responses are coded into themes across languages, so you can compare how often respondents mention each reason.
Start with the number of concepts, the segments you need to compare, and the smallest difference that would affect your decision. Early screening can use a smaller sample than a final comparison. Use the sample size calculator to plan each group, and check audience feasibility before fielding.
Compare concepts with real respondents and review the reasons behind their preferences.