Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.
CONJOINT ANALYSIS<br />Prof Narayan Janakiraman<br />
Customer<br />Value<br />Customer Value Assessment Procedures<br />Inferential/Value-Based<br />Behavior-Based<br />Attitu...
Show one product concept and get overall “Purchase Intent” feedback
Also get product diagnostics
Conjoint Analysis
Show multiple concepts and ask for overall preference
Concepts differ on Attributes and levels within an attribute
Based on overall preference get “part-worths” for attributes and levels within an attribute</li></li></ul><li>A Survey<br ...
Survey Results<br />
Conjoint Analysis in Product Design<br />Should we offer our business travelers more room space or a fax machine in their ...
P&G and Disposable Diapers<br />Question: What value do consumers associate with two improved features in disposable diape...
Conjoint Analysis Assumption<br />Products can be defined by their individual attributes and levels within the attribute<b...
Eg. Packaged Soup<br />
Eg. Packaged Soup individual concept<br />
Eg. Packaged Soup Conjoint INPUTCards and Ratings<br />
Eg. Packaged Soup Conjoint OUPUT 1Part Worths<br />
Eg. Packaged Soup Conjoint OUPUT 1Part Worths<br />
Eg. Packaged Soup Conjoint OUPUT 2Importance<br />Weights of Attributes<br />Flavor45%<br />Calories25%<br />Salt Freeness...
How were the part-worths calculated and how was the importance determined?<br />How does one use<br />Part Worths?<br />Im...
The Conjoint Model<br />
Notebook computer example<br />1)Processingspeed:1.5GHzor2.5GHz<br />2)Harddrive:120GBor160GB<br />3)Memory:1GBor2GBRAM<br...
One respondent’s preference<br />
Input to computer system – dummy variable regression<br />
Part Worth Estimation<br />Regression of ranks vs the attributes<br />U = a + b1*Processor + b2*Hard Drive + b3*Memory<br ...
Forecast preferences to check accuracy<br />
Weightage and Relative Importance of Each Attribute<br />Processor Speed<br />= 57%<br />Hard Drive<br />= 29%<br />Memory...
Segment consumers based on preferences<br />Are there segments in terms of preferences?<br />Here preference is the “basis...
Eg. Packaged SoupWhich is the most important attribute & which is the best product to introduce?<br />
Conjoint Simulation - The Motivation<br />What share can the new brand obtain?<br />Where does this share will come from?<...
Conjoint Simulation - The Principle<br />Before introduction share: A=40%, B=60%. <br />After introduction share: A=20%,B=...
Other ways of getting responses<br />
Conjoint Study Process<br />Stage 1—Design the conjoint study:<br />Step 1.1:Select attributes relevant to the product or ...
Upcoming SlideShare
Loading in …5
×

Lecture9 conjoint analysis

3,842 views

Published on

  • Be the first to comment

Lecture9 conjoint analysis

  1. 1. CONJOINT ANALYSIS<br />Prof Narayan Janakiraman<br />
  2. 2. Customer<br />Value<br />Customer Value Assessment Procedures<br />Inferential/Value-Based<br />Behavior-Based<br />Attitude-Based<br />Indirect/(Decompositional Methods)<br /><ul><li>Conjoint analysis</li></ul>Direct Questions<br />Unconstrained<br />Constrained/Compositional Methods<br /><ul><li>Multiattribute value analysis</li></li></ul><li>Conjoint Analysis and Concept Testing<br /><ul><li>Concept testing
  3. 3. Show one product concept and get overall “Purchase Intent” feedback
  4. 4. Also get product diagnostics
  5. 5. Conjoint Analysis
  6. 6. Show multiple concepts and ask for overall preference
  7. 7. Concepts differ on Attributes and levels within an attribute
  8. 8. Based on overall preference get “part-worths” for attributes and levels within an attribute</li></li></ul><li>A Survey<br />Familiarity & usage of value assessment methods<br />58 industrial firms in the top 125 of the Fortune 500 list<br />16 market research firms from the top 40<br />
  9. 9. Survey Results<br />
  10. 10. Conjoint Analysis in Product Design<br />Should we offer our business travelers more room space or a fax machine in their room?<br />Given a target cost for a product, should we enhance product reliability or its performance?<br />Should we use a steel or aluminum casing to increase customer preference for the new equipment?<br />
  11. 11. P&G and Disposable Diapers<br />Question: What value do consumers associate with two improved features in disposable diapers:<br />Improved absorbency<br />Elastic waistband<br />
  12. 12. Conjoint Analysis Assumption<br />Products can be defined by their individual attributes and levels within the attribute<br />Consumer responses to the overall preference can be then partitioned to attributes<br />
  13. 13. Eg. Packaged Soup<br />
  14. 14. Eg. Packaged Soup individual concept<br />
  15. 15. Eg. Packaged Soup Conjoint INPUTCards and Ratings<br />
  16. 16. Eg. Packaged Soup Conjoint OUPUT 1Part Worths<br />
  17. 17. Eg. Packaged Soup Conjoint OUPUT 1Part Worths<br />
  18. 18. Eg. Packaged Soup Conjoint OUPUT 2Importance<br />Weights of Attributes<br />Flavor45%<br />Calories25%<br />Salt Freeness22%<br />Price8%<br />
  19. 19. How were the part-worths calculated and how was the importance determined?<br />How does one use<br />Part Worths?<br />Importance?<br />The Black box<br />
  20. 20. The Conjoint Model<br />
  21. 21. Notebook computer example<br />1)Processingspeed:1.5GHzor2.5GHz<br />2)Harddrive:120GBor160GB<br />3)Memory:1GBor2GBRAM<br />Thereare8differentcombinationsofnotebook-definedasproductprofiles:<br />
  22. 22. One respondent’s preference<br />
  23. 23. Input to computer system – dummy variable regression<br />
  24. 24. Part Worth Estimation<br />Regression of ranks vs the attributes<br />U = a + b1*Processor + b2*Hard Drive + b3*Memory<br />The intution<br />
  25. 25. Forecast preferences to check accuracy<br />
  26. 26. Weightage and Relative Importance of Each Attribute<br />Processor Speed<br />= 57%<br />Hard Drive<br />= 29%<br />Memory<br />= 14%<br />
  27. 27. Segment consumers based on preferences<br />Are there segments in terms of preferences?<br />Here preference is the “basis” and “age” could be the descriptor<br />
  28. 28. Eg. Packaged SoupWhich is the most important attribute & which is the best product to introduce?<br />
  29. 29. Conjoint Simulation - The Motivation<br />What share can the new brand obtain?<br />Where does this share will come from?<br />
  30. 30. Conjoint Simulation - The Principle<br />Before introduction share: A=40%, B=60%. <br />After introduction share: A=20%,B=50%, and New=30%. <br />
  31. 31. Other ways of getting responses<br />
  32. 32. Conjoint Study Process<br />Stage 1—Design the conjoint study:<br />Step 1.1:Select attributes relevant to the product or service category,<br />Step 1.2:Select levels for each attribute, and<br />Step 1.3:Develop the product bundles to be evaluated.<br />Stage 2—Obtain data from a sample of respondents:<br />Step 2.1:Design a data-collection procedure, and<br />Step 2.2:Select a computation method for obtaining part-worth functions.<br />Stage 3—Evaluate product design options:<br />Step 3.1:Segment customers based on their part-worth functions,<br />Step 3.2:Design market simulations, and<br />Step 3.3:Select choice rule.<br />
  33. 33. 29<br />Attributes Should Be…<br />Determinant<br />Easily measured and communicated<br />Controllable by the company<br />Realistic<br />Such that there will be preferences for some levels over others<br />Compensatory<br />As a set, sufficient to define the choice situation<br />Without built-in redundancies<br />
  34. 34. 30<br />How Many Levels per Attribute?<br /><ul><li>Levels and range should be meaningful, informative, and realistic to consumers and producers
  35. 35. Avoiding absurd configurations
  36. 36. Marginal increases in levels can greatly increase respondent’s task</li></li></ul><li>31<br />Which Data Collection Method?<br />Full profile: Show complete list of attributes<br />Limited to 6-7 attributes<br />Pair-wise: Show pairs of attributes in matrix; each cell rated from most to least preferred<br />Lacks realism<br />Inconsistent responses likely<br />
  37. 37. Designing a Frozen Pizza – Paired Comparison Approach<br />1. Crust2. Type of Cheese3. Price<br />PanRomano$ 9.99<br /> ThinMixed cheese$ 8.99<br /> ThickMozzeralla$ 7.99<br />4. Topping5. Amount of Cheese<br /> Pineapple2 oz.<br /> Veggie4 oz.<br /> Sausage6 oz.<br /> Pepperoni<br />A total of 324 (3 * 4 * 3 * 3 * 3) different pizzas can be developed from these options!<br />

×