Hello
After running a CBC survey, the CBC-HB utilities for the price attribute are not at all what we expected. The exercise had 17 SKU's, 5 price levels (with conditional pricing) and a small-ish sample of 335 respondents.
Approximately 50% of the sample has a higher utility for the highest price than for the lowest price, and only 5 respondents follow the logic we expected - i. e. utilities decreasing as price increases.
In CBC-HB I used 200.000 iterations, and part worth coding for both attributes. All other settings where the default.
My questions are any at this point, but to keep it short I'll only ask a few:
1. One hypothesis is that price is not at all important for these SKUs (the values are indeed low, it is not a great deal of money). Can I somehow test this hypothesis ?
2. What would you do differently in CBC-HB? Should I try Linear coding for the price attribute? Any other settings / options I might try?
Any advice is welcome of course, it doesn't have to be a direct answer to my two questions.
Best regards,
Michal