Tutoring Conjoint Analysis with Examples

Conjoint Analysis has become an essential part of every marketer’s tool kit today. Most MBA students encounter Conjoint Analysis in the MBA program early in their first marketing course. GraduateTutor provides tutoring on Conjoint Analysis for MBA students. Rather than start with the theory, we teach with Conjoint Analysis by providing examples so students truly understand the concept and it’s application.

What is Conjoint Analysis

Essentially, Conjoint Analysis breaks down the value assigned by consumers to different product attributes. For example, the value attributed to a car can be broken down to value attributed to the country of origin, price, horsepower, torque, fuel efficiency, etc. The attributes used for a TV could be price, screen size, # of speakers, electricity rating, etc.

Regression & Conjoint Analysis

Conjoint Analysis uses the OLS regression method to tease out the value or utility of various product features and attributes individually. If you are not familiar with the OLS regression method, you can read about simple linear regression, multiple regression and how to interpret regression output here.

Key Terms in Conjoint Analysis

We start our conjoint analysis tutoring by understanding what is conjoint analysis and then define the key terms used in conjoint analysis.

Utility/Utils/Preference Score

Customers assign value to different products or product combinations. The value assigned can be captured by an imaginary score referred to as a utility. These are also referred to as utils/preference scores.

Profiles Or Variants

Companies offer different combinations or versions of products. These versions or combinations are called profiles or variants.

Source: Toyota.com retrieved on 12/10/2020

Attributes

A product is comprised of different features offered as part of the product. For example, a car’s attributes include country of origin, price, horsepower, torque, fuel efficiency, etc.

Levels

Each attribute can have different levels. For example, the price attribute or the products price can be $100 (low), $150 (medium), or $200 (high) and the fuel efficiency attribute’s levels can be 28 mpg, 32 mpg and 35 mpg.

Partworths

Partworths can be understood as the value of each part of a product’s profile. Or one part’s worth of the whole. This is measured in utils. For example, a car’s engine type may be worth 20% of the car’s value to some customers. Whereas its brand could be worth 50% of the car’s value.

Steps in Conducting a Conjoint Analysis

We tutor conjoint analysis using examples so it is easier for students to understand and retain the concepts as well as see its application. The steps involved in doing a conjoint analysis is outlined below.

  1. Choose Attributes: We first want to identify the key attributes that provide value to a customer. If we take a smartwatch as an example, attributes can be price, brand, health features, cellular reception, battery life, etc.
  2. Set levels for each attribute: For each of the attributes selected above, we need to set different levels or options that can be offered to customers or selected by customers. For example, the price attribute can be $100, $250, $500, $1,000. Battery life can be 24 hours and 28 hours. Note: Avoid vague levels. For example “higher price” can be interpreted by different individuals differently. Also note that the larger the distance between the levels, the more difficult it will be to interpret a value in between. In the example here, we can interpret a price of $150 (between $100 and $250) more easily than a price of $750 (between $500 and $1,000).
  3. Select Product Profiles: The combination of attribute levels (features) defines a product profile. For example, a low priced smartwatch from Samsung without health features, 24 hours battery life and WiFi connectivity is one profile. A high-priced smartwatch from Apple with health features, 48 hours of battery life and cellular connectivity is another profile.
  4. Possible profiles: The total number of profiles (combinations of attribute levels) can be computed by multiplying the number of levels in each attribute combined. For example, if we chose to have 5 attributes with 4 levels each, the total number of profiles possible is 4x4x4x4x4=1024 profiles. This is clearly too many profiles for consumers to review. In addition, more customer samples will be required to have statistically significant results. Therefore we need to carefully select the product profiles that are realistic to review.
  5. Execute the Consumer Survey: You can now conduct a customer survey armed with the profiles. Each customer must provide the utility provided by each of the profiles to him/her. The utility or preference rating usually ranges from 0-10 or 0-100 indicating that 0 is the lowest value. Each of the profile features must be marked and understood by the survey respondents. Note that there are many ways to execute this survey and appropriate conjoint model must be chosen.
  6. Conjoint Analysis Baseline Product: When performing the regression analysis, a baseline profile is first set up. All the other products are set up in reference to this baseline. If your Conjoint Analysis software does not automatically set up the regression variables, you will have to manipulate the data to make it regression ready.
  7. Perform the Regression Analysis: Conjoint Analysis uses the standard OLS regression method MBA students encountered in their statistics classes. The coefficients of the regression are used to tease out the value or utility of various product features and attributes. If you are not familiar with the OLS regression method, you can read about simple linear regression, multiple regression and how to interpret regression output here.
  8. Interpreting the Regression Analysis in Conjoint Analysis: The coefficients of the different levels provide fodder for the different interpretations and applications of Conjoint Analysis.

Applications/Uses of Conjoint Analysis

There are many ways conjoint analysis is put to use. Here are some of them:

Importance of Specific Attributes and Features

Product managers and marketers need to know which features of a product customers value and consider important. Conjoint Analysis can help you understand the value placed on each of the attributes and levels studied.

The range of each attribute’s levels gives the importance of each attribute. The percentage share of each attributes range indicates that the attribute’s importance to each customer. We can get the percentage share of each attribute’s range by taking the range of each attribute’s levels and dividing by the sum of the range of each attribute’s levels.

Estimating Market Share of Potential Products

It would be very helpful for managers and marketers to know what market share can be gained with specific features and pricing when evaluating new products. Conjoint analysis can help estimate a potential products market share based on its features and price levels.

We use the regression coefficients to estimate the utility of each product available in the market. From that utility, we can predict each customer’s purchasing behavior and then estimate the market share of a product.

Willingness to Pay (WTP)

Conjoint Analysis can help marketers and product managers understand the willingness to pay for different features and combinations. Since price is an attribute, we first estimate the $/utility. With the estimated $/value, we estimate a willingness to pay for different features and product profiles.

Evaluating Tradeoffs

Conjoint Analysis can help marketers and product managers understand the tradeoffs customers make for different levels of attributes and different attributes themselves.

Other Applications of Conjoint Analysis

Conjoint analysis can also be used in other situations such as segment wise market share, understanding individual differences, demand forecasting in established markets, etc.

Conjoint Analysis Software

Once you have understood the principles and processes involved in Conjoint Analysis, you can do Conjoint Analysis in Microsoft Excel, R, Python or any other marketing software packages available in the market. Appropriate skill is required in setting up the attributes, levels, conducting the surveys, and interpreting the output.

Conjoint Analysis: Origins

The name conjoint comes from co-joining. It originated from the Anglo-French word ‘conjoindre’. While Conjoint twins maybe what comes to mind for most people, ‘working together’ or ‘united’ is the more appropriate meaning of the word. This meaning is reflected in the usage “the product was the result of conjoint efforts of many teams”.

Another theory on the origins of conjoint analysis says that the word conjoint is simply a combination of considering (two ore more factors) jointly! The con from considering and join from jointly form conjoint in conjoint analysis! Because we say, customers are unable to or are not honest when we ask about the value of features or preferences, we tease out their preferences by evaluating their choices when given two different products. This process considers multiple decisions jointly and hence conjoint analysis.

This word Conjoint Analysis is used in the world of marketing because we believe that product preference and resulting market share is the coming together of many parts or factors. Conjoint Analysis is the process of breaking down the value of a product in a consumer’s mind.

Conjoint Analysis Case Studies

The following HBR case studies will help you understand and apply conjoint analysis to real life situations better.

Other Conjoint Analysis Resources

Do let us know if we can offer you tutoring assistance with your Conjoint Analysis.

Sign Up Now