Choice Modelling
Choice Modelling is a product of combinatorial mathematics, mathematical modelling and behavioural theory.
A single Choice Model has the power to make accurate numerical predictions about how people would behave in many trillions of possible scenarios.
The Nobel Prize in economics was awarded to Daniel McFadden in 2000 for his work on Choice Models.
Why know about it
Where the results from a traditional survey might report:
"75% of people rated your product as good or very good",
A Choice Model will report:
"by pricing your product at $59.99, packaging it in red and removing the two for one offer, 75% of people will buy it".
The details of how to build a choice model are complex. Nevertheless their use and application has accelerated over the past few years. In parallel, the cost of implementation has begun to drop as tools for managing them efficiently emerge.
A properly constructed model yields far more useful information than a quantitative survey does.
Choice Modelling is emerging from academia and gaining currency in the commercial world. Tools for constructing and managing Choice Models are being developed and more and more professionals without a background in advanced econometrics are benefitting from their power.
For these reasons, Choice Models will overtake traditional market research as the central tool in product planning and pricing.
When to use it
In any situation where you want to make robust predictions about how the price and features of a product or service affect its uptake.
- New product development
- Changes to existing product or services
- Optimising pricing and packaging of an existing product or service
- Brand and advertising positioning
Further Reading
- Introduction to Choice Modelling
dcm_basic.pdf (287K pdf) - Credit Card Model
credit_card.pdf (171K pdf)
External Links
- Wikipedia entry on Choice Modelling
http://en.wikipedia.org/wiki/Choice_modelling
- Kenneth Train's web site
http://elsa.berkeley.edu/~train/

