Marketing Research: What’s Your Part Worth?

 

Using conjoint analysis, companies can now answer that question.

 

Most products consist of a variety of features and benefits.  The perennial question is: how many features to add to your product? Guess wrong and you lose sales — either by over-engineering your brand or not giving customers enough of a reason to by your brand.  Part worths, the outcome of conjoint analysis, tell firms how much consumers are willing to spend for each feature they’re considering adding to their brand.  Part worths also tell a firm if consumers are willing to exchange one feature for an

other.

How Conjoint Analysis Works

1. A brainstorming session (or sessions) develop ideas for new products — hopefully building on customer needs with the idea of solving a customer problem.
2. Ideas are related to evaluate fit with the firm’s competitive advantages, its market opportunities,
   etc.
3. Top innovative ideas are fleshed out so they have a degree of concreteness — identifying features and benefits consumers are most likely to want in the new product.
4. Bundles are built — each bundle differs from the next bundle by one feature.  These bundles don’t have to be physically created, this is simply a tool for consumers to envision a new product.  The bundles do not have to be feasible either.
5. Data is collected on how members of the target market rank each bundle.
6. Data are analyzed using SPSS, SAS or some other statistical package.  The statistical package creates output containing the part worth of each feature of the bundle.
7. Now, is the time to re-evaluate the bundles since only those that are physically feasible can be created.

Marketing Research – an Example

Lets say you were an automobile company interested in creating a new car.  You might brainstorm features based on existing options, as well as market research with consumers to identify unmet needs in existing cars.  Remember, at this point we suspend our need for reality and worry about how to create the features after we determine which features consumers want most.  Let’s say you end up with this list:

• Electric power
• Gas power – high MPGs
• Hybrid drive
• Autopilot car
• Leather seating
• Cloth seating
• heated seating
• Burl wood finish
• Teak finish
• Pearl inlaid finish
• … the limit is your imagination

Now, we create bundles

1. Electric, leather, teak

2. Electric, leather, Burl

3. Electric, leather pearl inlay

4. Electric, cloth, teak

5. Electric, cloth, Burl

6. Electric, cloth, Pearl inlay

7. Electric, heated seating, teak

8. Electric, heated seating, Burl

9. Electric, heated seating, pearl inlay

So, with 3 features, 3 levels of each feature you would have 27 bundles.  Each bundle would also need a price associated with it — and this price needs to reflect a realistic price based on the cost of providing the features contained in the bundle.

Problems Implementing Conjoint Analysis

1. Conjoint designs are relatively complex and may be difficult to implement if there are a large number of proposed features, levels of features, or both.
2. The cognitive load on consumers evaluating the bundles is high.  For this reason, often individual consumers may only rank a subset of the total number of bundles.  If not accomplished correctly, the resulting data is useless.
3. Consumers often have difficulty envisioning products or features and assess a value to each.  This is especially difficult for features they’ve never encountered before or that they haven’t used.  Features based on technology they don’t understand can also cause them to reject options as being unrealistic.
4. Consumers have a hard time evaluating their response to features situated within a social context.  For instance, what is the value of having a feature similar to one valued by members of their social group?  What is the emotional worth of having a feature desired by your family? etc.