Have you ever had a need to understand what specific product attribute or feature your customers (and, maybe prospects) prefer or find most important? If so, you may have come to the realization (after a lengthy, costly research study) that simply rating attributes or features on a five, seven or even 10-point Likert scale will not provide clear differentiation of preference or importance among competing attributes/features.
- Offers better preference/importance differentiation among attributes/features, compared to Likert-based attribute/feature ratings. This is especially important if budget constraints come into play – sample sizes may be reduced without necessarily compromising the significance or robustness of MaxDiff results.
- Eliminates all issues related to scale bias. Scale bias occurs when one respondent may rate an attribute lower than a second respondent would, even though the two respondents’ preference or importance inclinations are the same. Scale bias is normally a significant issue when the respondent sample is diverse across countries of origin or cultures.
- Allows for relative statements to be made about one attribute/feature compared to another. This is accomplished through the computation of shares of preference (or importance) for each attribute/feature. For example, If feature #1 has a share of preference of 30%, and feature #2 has a share of preference of 15%, it would be appropriate to claim that feature #1 is twice as preferable as feature #2 (30/15=2).
Below is an illustration of one choice set that might be part of comprehensive MaxDiff exercise consisting of multiple choice sets:
For more information on how to put MaxDiff to work for your company, click here.
