Have an idea?

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

Calibration - overestimatation issue

I am doing CBC allocation study with doctors to assess preference for new product (line extension) and impact of new market entrant. Doctor will allocate next x patients across scenarios. Results will be used in forecast uptake of new product and brand as a whole. I tried searching forum and technical papers but still could not find a concrete way to deal with overestimation problem in new product. The only thing I was able to find is from 2004 conference.  Is there any more materials/methodology that you can refer to tackle the issue of overestimation of preference share?

P.S. I have already considered traditional approaches like analogues, diffusion model to adjust the share for awareness.
asked May 12, 2014 by Mike

1 Answer

0 votes
Best answer

I see these a lot and I have a possibility for you to consider.  I assume in answering that you're as yet uncertain about what attribute levels will and won't pertain by the time of product launch so that you're using CBC to test multiple levels of multiple attributes.

One thing I often do in these kinds of studies is to ask two standard, 5-point scale, purchase intent ratings questions about the new product, one for  a well-featured version of the new product and one for a relatively impoverished version of the new product.  You can simulate shares for the new product using your CBC utilities, of course.  You can also apply some common deflators of purchase intent overstatement.  This gives you a target  to use in calibrating your CBC utilities - in fact, you now have deflated purchase intent for both strong and weak versions of the new product to use in your calibration/data fusion.   

There are some other things I often do in these kinds of studies, but it sounds like you're already planning those.  Good luck.
answered May 13, 2014 by Keith Chrzan Platinum Sawtooth Software, Inc. (90,475 points)

Unfortunately, due to nature of new product and this being pharmaceutical drug, a distinction between well-features vs. relatively impoverished version is not feasible. Outside conjoint, we do show them product file and get answers on improvement level vs. current offering, allocation of patients across products and when this would become standard therapy. Would really love to pick your brain on this. Any ideas? Would love to call you.