Conjoint Simulator: Demand Varies by Feature
Which features are the most profitable to deliver?
How much value do consumers place on different features? This can include features that are currently in the marketplace or hypotheses about new features you think might move the needle.
Context matters
The amount of demand for a product/services is dependent on how much consumers like it (utility) and the price, but also on the other options available to choose from. Consumers may prefer milk chocolate over dark 10:1, but in a store filled with milk chocolate, a dark alternative will get more purchases than utility alone would dictate. Differentiation that is so important in the marketplace also shows up in conjoint simulators.
Conjoint requirements
A conjoint simulator has one requirement that is too often overlooked. That is each features must have exactly one level per product. For example if the chocolate bar feature Format/Shape has the levels: Bar, Cup, Squares, Candy coated and Mini chunks, each product in the simulator has to have one of those levels - not zero, not two.
Sometimes it makes sense that some products won't have a feature level. In that case, be sure to include "none" or "blank" as a level within that feature for the choice exercise.
Also, sometimes it makes sense that multiple levels of a feature are possible. You can plan for this by making combo-features like (caramel and peanut butter). It is also possible to split the levels into multiple features. Otherwise, there is no fix on the back-end to include more or less than one level per feature.
Heterogeneity - not just a fun word for parties
Heterogeneity just means that people are different or they like different things. I like milk chocolate a lot more than dark or white chocolate. My cousin, the chocoholic, likes them all equally. Others might like dark more than the rest.
When there are difference in preference across people, the heterogeneity is vital for multi-product portfolio that captures as much profit as possible.
Automation
Changing features one at a time and recording the new demand forecast can be so boring you are asking for human error. Using automation tools to flip through each feature and record the outcome saves time and reduces errors.
