Conjoint Analysis for Product Managers and Innovators- Valutrics
What’s in your toolbox for conducting consumer and market research? Does it include Conjoint Analysis?
- the types of information Conjoint Analysis can provide, such as pricing specifics,
- when to use Conjoint, and
- the specific steps for using Conjoint.
What do we use Conjoint Analysis for?
Conjoint Analysis is a powerful market research tool. It is a highly quantitative, sophisticated tool, and we use it to predict what are people going to do in the future. We present scenarios to them and see what they do. It relies on the old human adage, “You don’t know how important something is until it’s gone.” So what Conjoint does, is it offers people things and then takes them away. Then you find out which things were missed the most.
When in the NPD process is Conjoint Analysis helpful?
It usually applies in the middle to the late part of the design process, after features have been identified but before they have been selected. An example is where the developers say, “Hey, we could put 10 different features on this new product but the marketing people and the finance people said we can only afford three.” Well, which three are you going to select? That’s what Conjoint does so well. It helps you pick. It helps you use customers to pick which of those three are going to make you the most money and sell the newest products.
What are the steps performed to conduct a Conjoint Analysis?
The steps are…
- Identify all potential product features
- Select a subset of features (around 7) to evaluate
- Create scenarios incorporating combinations of the selected features
- Create a questionnaire/survey with scenarios
- Collect customer preferences for the scenarios
- Perform the statistical analysis
- Summarize your findings
Can you take us through a practical example?
It is much easier to explain the steps using an example. I like bicycles so I often use a bike example. Let’s assume that we are Trek or Specialized and we’re designing the bike for the next season. The engineers have come up with three features they could add to the new bike design. They need to pick the best one or two and figure out how much people are willing to pay. The features are: (1) an improved full-suspension frame, (2) a folding frame, and (3) a lightweight high-performing frame. We want to exam four price points also – $300, $600, $900, and $1200.
An Experimental Design tool, present in some statistical toolkits, can help build the scenarios. Below are 16 scenarios. Brian provided this to me and I added the BuyRating for each based on my personal opinion of how appealing the scenario was. The scale is: 1=Definitely Not Buy, 2=Probably Not Buy, 3=Might Buy, 4=Probably Buy, 5=Definitely Buy.
Brian applied statistical analysis using regression analysis, which can be run in Excel or dedicated statistic systems.
Brian’s analysis of the data allowed him to conclude that… “Chad is willing to spend $333 extra to get the lightweight frame (and cut 10lbs). Willing to pay $250 to get suspension. BUT, we’d need to pay him $83 to take a fold up bike.”
He also examined my price sensitivity based on my preferences for each scenario, finding… “Chad is VERY willing to buy any $300 bike. But really has trouble considering bikes at $600 or above. If it is $600, it had better have the Lightweight frame and Suspension! We can even predict his score for such a bike. It would be “4”, Probably Buy.”