Have an idea?

Visit Sawtooth Software Feedback to share your ideas on how we can improve our products.

Analysis of ACBC with constructed lists

I conducted a survey using ACBC and constructed lists. I wanted to know whether in a brand architecture, the product-, family- or the company brand is more important for the customer than the others. Since there are multiple product brands and not all customers are buying the same, I used constructed lists and unverified perl.
Now I want to run the analysis and get single utility scores for every brand (product brand a, product brand b, product brand c, family brand, company brand), and of course these scores are also minus values since every customer used a different brand.
Is there a possibility to combine the values of brand a, brand b, brand c to one value for the product brand?
asked Jul 23, 2017 by anonymous

1 Answer

0 votes
I'm not sure of a reliable way you could do this.  The levels of the attribute receive their values based on relative preference, so it feels pretty weird to try to add things together and feel confident that the number that comes out is providing you with something easily to interpret.  For example, if I had 3 varieties of bread Brand A and 3 varieties of bread Brand B, it might be interesting to show the averages for the brand as a whole.  But to me it gets a lot more complicated if I have 4 varieties of Brand A, 2 varieties of Brand B, 1 variety of Brand C, and some of the Brand A varieties are niche products that not many people buy.  What exactly is the average of those 4 varieties telling me?
answered Jul 24, 2017 by Brian McEwan Gold Sawtooth Software, Inc. (38,865 points)