32 0 8MB
Getting Started with Conjoint Analysi Strategies for Product Design and Pricing Research Second Edition
Bryan K. Orme -*' •* !
\
v*
'\ M L-
/'
,A V !/ r^
Research Publishers
"-' v.-/ /-
V
Getting Started with Conjoint Analysis: Strategies for Product Design and Pricing Research Second Edition by Bryan K. Orme Copyright © 2010 by Research Publishers LLC
Contents
All rights reserved. No part of this work may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, mechanical, electronic, photocopying, recording, or otherwise, without the prior written permission of the publisher. f ^ .-•;') Publisher Research Publishers LLC 660 John Nolen Drive Madison, WI53713-1420 USA http://www.research-publishers.com
Foreword
^ } ! X
$ X
X X X X X
Exhibit 2.2. Credit card design plan degree of precision. For example, it would be difficult to distinguish the separate effects of brand and interest rate if Visa always appeared with the lowest interest rate. Would the preference for such a concept be due to the desirability of Visa or the low interest rate?
2.3 2.2
:^
^ N?
How much do you like this credit card offer? Use a scale from Oto 10, where 0 *= not at all and 10 = very much. Write your answer in the blank box.
20%lhterest cmdlt limit
9
2.3 Credit Card Survey
X X
How Conjoint Analysis Works
Credit Card Survey
Now that I have given an introduction to conjoint analysis and design plans, please complete a survey as if you were choosing a new credit card. Drawing upon the design plan, exhibit 2.3 presents nine credit card offers to be evaluated. For each credit card offer, you are asked," "How much do you like this credit card offer?" Use a ten-point scale to score each offer, where 0 means "not at all" and 10 means "very much." After you have completed the credit card survey, rank the credit card offers in the final exercise (exhibit 2.5 on page 17). I hope you found the credit card survey interesting. You probably also found it a bit challenging. To evaluate the cards, you probably developed a strategy. Perhaps you decided early on which attribute was most important to you, and you based your decision on that aspect. Perhaps each of the attributes carried about equal weight in your decision, and you needed to decide how much of one you were willing to give up for the other. You probably did not find an offer that was exactly what you wanted in every way. So it is with real purchase decisions. Consumers make these kinds of trade-offs every day.
10
How Conjoint
Analysis
Show how much you like each of the nine credit card offers below by writing your answers in the blank boxes. Use a scale from 0 to 10, where 0 = not at all and 10 - very much. Discover 20% interest $5,000 credit limit
MasterCard 10% interest $2,500 credit limit
Utilities
11
Now you are about to learn why conjoint analysis is so useful to managers and market research analysts. Based on the credit card ratings you provided, you will compute a weight or part-worth utility for each of the attribute levels. The set of weights would account for your overall credit card ratings. After all, you probably were loosely applying some sort of unconscious scoring mechanism or system of preference weights. Conjoint analysis seeks to uncover that system of preference weights. Human decision making is undoubtedly complex, and a simple set of weights can never fully capture the complexities. Conjoint analysis tends to do a very good job despite the model simplifications.
3
MasterCard 20% interest $7,500 credit limit
5
2.4 Conjoint Analysis
visa 15% interest $7,500 credit iimit
6
p..
Discover 15% interest $2,500 credit limit
rv
Discover 10% interest S7.500 credit limit
2.4 Conjoint Analysis Utilities Because each attribute level appeared exactly once with every other level in the study, there is a simple way to estimate attribute level utilities (also known as partworths). Of course, conjoint studies in the real world are rarely so straightforward. I have constructed this example so that the utility estimation may be done with a hand calculator. For this simple illustration, the utility estimate for each level is the average score for cards that include that level. Instructions for the calculations are provided in exhibit 2.4. Please compute the utility scores for each level now. If all has gone well, the utility scores should make intuitive sense to you. The higher the score for a level, the more desirable that level. It often helps to visualize data or to plot utility scores on a line chart. Please plot your part-worth utility values for each attribute in figure 2.1, drawing lines connecting the points on each graph.
2.5 Importance Scores
Visa 10% interest $5,000 credit iimit
MasterCard 15% interest i.OOO credit limit
9 Visa 20% interest $2,500 credit limit
Some researchers calculate an importance score for each attribute. An importance score reflects the effect each attribute has upon product choice, given the range of levels we included in the questionnaire. The calculations are straightforward, and you will again be able to use a hand calculator. Use figure 2.2 to work on your calculations. Refer to chapter 8 (page 80) or the glossary (page 171) for examples of importance score calculations. After you have finished calculating importance scores, you can plot them on the bar chart in figure 2.3.
2.6 Conjoint Analysis as a Predictive Model of Choice
Exhibit 2.3. Credit card survey
Charts of utilities and importance scores are useful, but a what-if market simulator that can be built using conjoint results is the most valuable tool for managers. A market simulator uses the utility scores to predict which product alternatives respondents would choose within competitive scenarios. The predictions can be made not only for the few product alternatives that were actually shown to respondents, but also for the often thousands or more potential combinations that were not shown.
How Conjoint Analysis Works
IZ
Record your survey responses in the blank boxes for the numbered cards. The rating for each credit card offer is recorded in three groupings: once for the brand, for the interest rate, and for the credit limit.
conjoint nnuiyzt-i H.J >• - —•---