Core Concepts: Conjoint and Discrete Choice Modelling for Pricing Research
Steven Forth is a Managing Partner at Ibbaka. See his Skill Profile on Ibbaka Talent.
Definition: Conjoint and Discrete Choice Modelling
Techniques used in market research to discover the contribution of different offer attributes (including) price to a preference or purchase decision.
Conjoint and Discrete Choice Modelling for Pricing Research
One of the tools used in pricing research is Discrete Choice Modelling. You have probably been subjected to a discrete choice experiment, shown two or three, possibly more alternatives, and asked to choose the one you prefer. It is a discrete choice because you have to make an either/or decision rather than rank the alternatives in order or to rate each of the alternatives. These experiments have been popular in consumer pricing work for a long time and are increasingly being used in B2B pricing.
A well designed Discrete Choice Experiment is very useful in understanding how to combine function groups into packages or bundles and to get a feeling for the relative price level for each package. Let’s say you have five different sets of functionality (ideally a functionality set will contribute to one or two connected value paths). How would you combine these different function sets into packages that will appeal to specific target segments? You could build 120 different packages from five function sets (factorial five or 5!) which is likely far more packages than you want to offer (you would just confuse buyers). The Discrete Choice Experiment will help you figure out which combinations are most attractive to buyers.
What are the different types of Conjoint and Discrete Type Modelling?
There are many different ways to design and analyze this sort of forced choice. The most common is conjoint analysis. Conjoint is so popular that it is sometimes used to refer to all sorts of choice experiments. It is more accurate to use conjoint for experiments in which several attributes (which can include price) are presented together. The choices are conjoined in that all of the relevant product attributes are presented together (like conjoint twins) rather than independently. This is a much better approximation of the real world and more informative when designing packaging. Ibbaka generally uses the discrete choice form of conjoint, in which people are asked to choose one alternative (or to reject all alternatives) rather than to rate them on a preference scale or put them into an order. This is closer to the actual buying process where one either buys or not.
The importance of part worth utilities to packaging and pricing work
When using discrete choice modelling you are likely to be presented with the ‘part worth utility’ of each attribute. The conjoint model assumes that each attribute contributes something to the total utility and that these can be added together to give the total utility (the combination with the highest utility is the one chosen). The part worths always add up to 100%.
Conjoint.ly, the platform Ibbaka uses for its discrete choice research, has good visualizations of part worth utilities (one example is shown below). The Conjoint.ly website has a good explanation of this.
When to use Conjoint and Discrete Choice Modelling?
Discrete Choice Modelling is most useful for established categories where people have a good understanding of the problem they are trying to solve and the possible solutions. It is product focused and not customer or value focussed.
It is most useful in deciding what attributes to combine into a package or bundle.
Can Discrete Choice Modelling be used to estimate value?
No! Conjoint and discrete choice studies will not tell you how much value a solution offers and is not a substitute for a good value analysis and for building a value model. Price design based in this type of study is likely to miss the real opportunities to frame value and to connect value and price.
Can Discrete Choice Modelling be used to estimate Willingness to Pay?
No! Willingness to Pay (WTP) is highly malleable and depends on how value is presented and the state of the buyer’s business. WTP is best thought of as something that is shaped by marketing, sales and pricing and not something that exists ‘out there’ waiting to be discovered.
Can Discrete Choice Modelling be used to estimate Price Elasticity?
Used carefully, conjoint and discrete choice modelling can give important insights into cross price elasticity (the degree to which buyers will change vendors based on price changes). This is important to understand when factoring the impact of competitor pricing. It is less likely to give accurate estimates of price elasticity of demand (the degree to which demand changes in response to price changes). The conditions under which these experiments are conducted are too isolated from actual market conditions that determine price elasticity.
Want to learn more about Conjoint and Discrete Choice Modelling?
If you prefer to read books, here are two options.
To get started read Getting Started With Conjoint Analysis by Bryan K. Orme. Bryan is the CEO of Sawtooth Software, one of the leading vendors of advanced tools in this space. Want to go deeper, try Becoming an Expert in Conjoint Analysis: Choice Modelling for Pros by Bryan Orme and Keith Chrzan.
More Reading For You
Core Concepts for Pricing and Customer Value Management
Conjoint and Discreet Choice Modelling for Pricing Research (this post)
Coming soon …
Community Value Driver
Connecting Value and Pricing Models
Cross Price Elasticity
Customer Value Journey
Customer Value Management
Economic Value Driver
Interactions of Cross Price Elasticity and Price Elasticity of Demand
Package Design
Pocket Price Waterfall
Price Elasticity of Demand
Pricing Design
Pricing Model
Value Based Market Segmentation
Value Ratio
Value to Customer (V2C)