Conjoint 101
The Business Objective
You have a number of potential features of your product or service, and you’re not sure how to best optimize the price-feature relationship to maximize market penetration or profits.
Beyond using conjoint to figure out how to best price your product/service, it can also be used to redesign or reposition your product/service, how to figure out which features to put in your new product, or how to measure your brand equity. What-if market simulators can help answer questions such as “From which competitors would we take the most share if we increased the length of our service contract?” and “What happens to our overall share if we come out with another product with lesser performance at a lower price?”
A Brief “How-To”
For traditional choice-based (discrete choice) methods you need 300 to 600 respondents who know about the type of products or services you are studying
The most common type of conjoint to use is called CBC (Choice-Based Conjoint) – it’s really good for pricing research and to see interactions between attributes
Attributes are the product/service features you want to study – try to keep the number down, but if you can’t there are ways to handle large numbers of attributes
Level – each attribute has various levels – say the attribute is color, the levels could be red, yellow, and blue – again, be reasonable with the number of levels to each attribute; try 2 to 5; they should be mutually exclusive and cover the full range of possibilities
Conjoint goes beyond describing things to predicting them
People (customers, users) evaluate how much they like something based on the value of its separate (yet conjoined) parts
Why Would A Product Manager Use Conjoint?
Asking buyers to rate the importance of a bunch of attributes often gives little differentiation
Buyers have to make difficult trade-offs and concessions in real life
We have a way to force respondents to make difficult trade-offs, just like real-world buyers have to do
When we force them to make these trade-offs, we get a much better picture of the values of the product attributes; we get better discrimination
Plus, you get a cool market simulator that lets you play with various product configurations to figure out which can lead to the best market share
Other things managers like to hear
You can measure Brand equity – for example, you can quantify the price premium one brand may command over another
You can conduct Strategic pricing – test price ranges or new products outside of current offerings
Definitions
Attribute: You might think of this as a “variable” – it’s a characteristic of the product or service – something like brand, price, color, length of contract, etc.
Level: A degree or amount of an attribute – for example, the attribute “brand” could have levels of “IBM”, “Apple”, and “Dell”. Length of contract could have “1 month”, “6 months”, and “1 year”.
ACA: Adaptive Conjoint Analysis
CBC: Choice-Based Conjoint; also known as “Discrete Choice Modeling (DCM)”, “Discrete Choice”, or “Choice Analysis”
CVA: Conjoint Value Analysis; also known as “full-profile conjoint”
Concept: an offering shown to a respondent – for example, “HP, 3 GHz, 1-year warranty” – also known as “alternative”, “profile”, “stimulus”, or “treatment”. “Profile” is probably the most commonly used word.
Design: All the attributes and the levels of each
Part Worth: component of desirability associated with a particular level of an attribute
Utility: total desirability of the product alternative; made up of the part-worths of its separate attributes; the preference for the overall product concept
Share of Preference: predictions of market share from conjoint market simulators (expressed as percentages summing to 100 percent across competing product alternatives); don’t confuse with real “market share” which relates to actual purchases
Limitations
A conjoint survey assumes perfect information by educating respondents about available brands and features; but in the real world, obscure brands have less chance of being purchased
Conjoint assumes all products are equally available
Some survey respondents may not have the interest, authority, or ability to purchase
Results show potential market acceptance – you still have to promote and distribute the product
Conclusions
The reality in marketing research is that most stakeholders want simple answers; most stakeholders are also fairly stats/research-phobic. Therefore, conjoint is less used than simpler, less valid descriptive survey techniques. However, if you’ve got a big decision to make, it deserves the time, effort, and extra validity conjoint affords. If you’ve ever got a “big-decision” project and you want to discuss whether conjoint might be appropriate, just contact us at contactus@elucidatenow.com – we’ll help you sort it out.