Header image

Home>accomplishments>Choice modeling
Developed state-of-the-art Choice Modeling technique and analysis software

Product development teams were continually faced with tradeoff decisions and no good way to evaluate the impact of the tradeoff. Would it be better to include a specific feature and charge more for it? Include it as an optional add-on? Reduce the performance in one area to allow for greater performance in another?

This type of tradeoff question can be easily modeled if the right market data is collected and the right analysis tools are provided.

This need led to us developing a customized choice modeling capability that has been used by over a dozen operating divisions to help refine and optimize their products.

We used a blend of self-explicated and discreet choice conjoint techniques to provide the most accurate predictive capabilities possible.

The technique was used to

  • predict future revenue for several different HP product lines
  • refine the designs for new products to maximize their sales potential
  • track customer and brand loyalty over a seven-year period
  • provide extensive "What-If' scenario predictions to anticipate the impact of competitive moves, product cannibalization, and unforeseen market changes

Some examples of the accuracy of the technique are:

  • predicted the revenue impact of a proposed price decrease in Japan to within 3% of the actual impact
  • predicted that 90% of the revenue from a proposed new Power Meter design would come from cannibalization of existing products; came up with slightly altered design that would still meet the release schedule but get 35% of its sales from competitive products
  • showed that a new Signal Generator should actually be split into a family of 3 products; sales analysis one year after introduction showed the predicted sales for each product to be within 5% of actual sales
return to all Accomplishments