Conjoint analysis, another technique used in market research, was developed in the early 1970’s to gauge how buyers value different features and components of a product/service. MedPanel has many years of experience employing this method, working closely with Joe Curry, founder of Sawtooth Technologies and expert in choice-based conjoint analysis. Together, we are able to deliver powerful research results that translate readily to effective business solutions for our clients.
What is Conjoint Analysis?
Oftentimes, price is not the only product attribute a client wishes to configure or evaluate. After all, there are many other features that factor into buyer decisions. Say, for example, that you manufacture televisions, and you wish to streamline your production with an option that better satisfies your market. You are unsure, however, what screen format you should use: LCD, Plasma, or LED? You’re debating between a resolution of 720p versus 1080p, too. And you’re torn, about the refresh rate as well: Should you produce a TV that boasts 60Hz, 120Hz, or 240Hz?
Including all the most high-end technologies in your TV would be great, but doing so is likely to increase production costs and lower profits. Worst yet, it may price the product out of the target consumers’ reach. Ultimately, how do you decide which features to include, and which to forego?
Using conjoint analysis, a robust statistical tool, MedPanel is able to provide answers for such quandaries. By presenting market research participants with a number of different product profiles, each with a different combination of the individual attributes in question, MedPanel is able to evaluate the tradeoffs consumers are most willing to make. Because forcing difficult tradeoffs reveals the qualities buyers truly value, conjoint analysis enables us to identify the unique added worth of each individual feature. Taken together, the results of the probe will point to the combination of elements that maximizes the product’s overall appeal.
When we use Van Westendorp versus when we use Conjoint Analysis:
Van Westerndorp is the preferred methodology when trying to assess the pricing of a product for which other features have already been configured. It is useful, as well, in instances where a product is entirely new, with little or no data to pre-indicate a competitive pricing range.
If, however, price is just one of multiple factors of interest with respect to the product at hand, Conjoint Analysis is the more suitable approach. By simulating various market scenarios, this method requires respondents to make decisions and tradeoffs that effectively reveal their priorities. It is also worth mentioning that conjoint analysis does not come in just one “flavor.” Many types of conjoint analysis have been adapted since the technique’s inception in the early 70’s, as shown in the diagram below:
How we incorporate Conjoint Analysis in our research:
We vary the individual features being examined (i.e. the independent variables) to create a number of product designs or concepts. This way, each concept represents a different combination of features. We then ask respondents to rate or rank those product concepts against one another. Based on their responses, we are able to determine the unique value (or utility) of each feature specified in the survey. Essentially, rather than asking buyers directly if they prefer one feature over another (say, high resolution over high refresh rate), we offer realistic package choices that represent tradeoffs from which their preferences can be inferred.
An example of Conjoint Analysis using “Prepaid Debit Cards:”
Step 1: Specify attributes of interests, and the options within each one
|Make Purchases||At Stores that Accept Debit Cards in US|
At Stores that Accept Debit Cards World Wide
|Pay Bills Electronically||No|
|Get Cash||Bank Teller Only|
Bank Teller or Merchant
Bank Teller, Merchant or ATM
By Email, Phone or Text
|Balance Inquiry||ATM Only|
ATM, Phone or Online
|Deposits are FDIC Insured||No|
Step 2: Create a spectrum of products using different combinations of attribute levels, starting with one that has all the least attractive options (e.g. “Worst Card”) and ending with one that has all the most attractive options (e.g. “Best Card”)
Neither extreme is viable – viable alternatives are somewhere in between – but where?
Step 3: Force people to make tradeoffs by indicating preference. Repeat the process 10-12 times, varying the attributes level shown. Then, collect this preference data.
Step 4: Quantify the preference data collected in Step 3 and express in terms of utility (the value deemed by buyers).
Step 5: Utility values can then be further analyzed to predict choices individuals would make in various hypothetical market scenarios.
Predictive Example 1:
Predictive Example 2:
Step 6: These predictions can then be generalized to a larger group of individuals.
- The example shown in Steps 1-6 represents a Choice-Based Conjoint (CBC) Analysis. Working with us will allow you to utilize this approach and other types of Conjoint Analysis as well as You should add an example of how conjoint (in the example we used CBC) can predict an individual’s choice after you show the Recipient 1’s utilities. Working with us will not only let you do conjoint analysis, but also ACA, CBC, MaxDiff and ACBC.
The table below provides a list of some of the most commonly employed forms of Conjoint Analysis and a brief description of each:
|Choice-Based Conjoint Analysis||CBC||CBC is the most popular conjoint-related technique in use today. Respondents are shown multiple product concepts (and an optional “None” alternative) and are asked which one they would choose.|
|Adaptive Choice Analysis||ACA||ACA describes a survey style in which the computer interview customizes the experience for each respondent. It is designed for situations in which the number of attributes exceeds what can be reasonably done with more traditional methods (such as CBC).|
|Adaptive Choice-Based Conjoint Analysis||ACBC||ACBC is a new approach to preference modeling that leverages the best aspects of CBC (Choice-Based Conjoint) and ACA (Adaptive Conjoint Analysis). ACA focuses on the attributes that are most relevant to the respondent and avoids information overload by focusing on just a few attributes at a time.|
|Best/Worst Conjoint (a.k.a. Best/Worst Scaling)||MaxDiff||MaxDiff is an approach for obtaining preference/importance scores for multiple items. It is relatively easy to use and applicable to a wider variety of research situations.|