On Tues August 24th, 2010, SurveyAnalytics teamed up with Chris Robson, Chief Scientist and Co-Founder of Parametric Marketing to present : Learn How to Use MaxDiff Analysis Create Products People Want To Buy
We had 80 guests in attendance. In our session Chris Robson covered the following:
What is MaxDiff?
-Why all the fuss?
-Problems with Ratings Scales, and why MaxDiff is better
Maximum Difference Scaling (MaxDiff) is a way of evaluating the importance (or preference) of a number of alternatives.It is a discrete choice technique: respondents are asked to make simple best/worst choices. It has the advantage that it is very simple for the respondent, but gives extremely rich information to the researcher
Q: When is it more appropriate to use Max-Diff instead of Discrete choice Conjoint? Conjoint is also about making trade-offs.
A: MaxDiff and Conjoint are related discrete-choice methods. Conjoint works best when the features to be tested can be structured into attributes with mutually-exclusive levels – e.g. the ‘attribute’ color has ‘levels’ red, blue or green. Also, Conjoint can be used to test price sensitivity. MaxDiff works well when there is a ‘laundry list’ of unalike features.
Q: What is your opinion on the question type, “Constant Sum” or “Continuous Sum” it is sometimes called. With this question type you can “Rank” the attributes, but are able to see to a certain degree how much more important something is over another. It seems like the end result of max-diff is similar. Are they comparable in effectiveness at all?
A: One of the main problems with ‘Constant Sum’ questions is that they get unwieldy for the respondent when there are more than a few options to consider. When there are many options respondents start to use different strategies, like “Give all points to best item” or “Spread thinly over many items”, which gives problems for analysis.
Q: Where can we find the paper he is talking about?
A: I recommend: Chrzan, K. and N. Golovashkina (2006) “An empirical test of six stated importance measures,” International Journal of Market Research, 48, 717-40.
A simpler overview of the problems with importance scales by the same authors can be found at: https://www.maritz.com/White-Papers/Research/Testing-Alternatives-to-Importance-Ratings.aspx
Q: Is share of preference similar to TURF?
A: TURF analysis can be used with MaxDiff results to identify optimum feature configurations.
Q: Is there a risk of forcing a difference between 2 attributes when they are actually equal?
A: It is wise to avoid including two features that are essentially the same. If, however, there are real differences forcing a choice is a good thing.
About the Presenters:
Co-founder Parametric Marketing LLC and serves as Chief Scientist, leading selection and development of all analysis capabilities.
Chris Robson co-founded Parametric and serves as Chief Scientist, leading selection and development of all analysis capabilities. He is a seasoned executive with extensive experience in both marketing and technology. Prior to co-founding Parametric, Chris was Vice President of Engineering for WebCriteria, Inc. (acquired by Coremetrics, Inc.), and was responsible for the development of Site Analyst, described by Forrester as “a smart – and unique – choice for Web Managers struggling to define success metrics”. Prior to that, he spent thirteen years with Hewlett-Packard in a variety of senior management roles in R&D, Marketing, Business Development and Advanced Research. He started his career as a mathematician building complex statistical models of radar systems. Chris is a native of England and holds an honours degree in Mathematics from Brunel University of West London.
Esther LaVielle is a Senior Account Manager at QuestionPro and Survey Analytics, which was started in 2002 in Seattle and is now one of the fastest growing private companies in the US. Prior to her adventure at QuestionPro she spent 3 years as a Qualitative Project Manager at the Gilmore Research Group.