|
Hu, C. (1994). Measuring meeting planner's preferences of hotel site selection: A hybrid conjoint analysis. Unpublished M.S. Thesis, Purdue University, West Lafayette, IN.
ABSTRACT
Hu, Clark. M.S., Purdue University, December 1994. Measuring Meeting Planner’s Preferences of Hotel Site Selection: A Hybrid Conjoint Analysis Approach.
Major Professor: Stephen J. Hiemstra, Ph.D.
This thesis addresses the need for measuring the meeting planner’s preferences of hotel site selection to help hotel management understand what their customers want in the meeting business. Traditional attitude (preference) measurements rely heavily on the expectancy-value models. These models fail to detect the respondents’ tradeoffs in their decision-making process. This thesis also attempts to demonstrate a better preference model – hybrid conjoint analysis – to overcome the drawbacks of traditional models and provide marketing information for further pricing strategy.
A group of 500 professional meeting planners was randomly selected as the research sample from the MPI® (Meeting Planner International) 1993~1994 Directory. Six important attributes and twenty-two levels involved in meeting planning decisions were identified and incorporated into questionnaire design. The meeting planner data were collected in three stages: self-explicated desirability rating, self-explicated importance weight allocation, and profile evaluation. Three preference models, self-explicated, regular conjoint, and hybrid conjoint models were tested by OLS (Ordinary Least Squares) regression in this study to develop an aggregate utility (preference) function across meeting planner data. Two cross-validation indicators, first-choice hit rate and correlation coefficients, were calculated to evaluate both internal validity and external validity for all three models. A probabilistic model was employed to simulate the market shares for five hypothetical hotel profiles to answer three hypothetical “what-if” scenarios. Finally, both own-price elasticity and cross-price elasticities were calculated to provide marketing information for pricing strategy.
The author found that the “price (room rate) range” is the most important attribute in meeting planners’ considerations of hotel selection. “Location” is the second most important attribute, followed by “functional properties of meeting rooms,” “hotel guestroom comfort,” “food and beverage function,” and “hotel conference planning procedure.” The most preferred level for each of the six attributes was also identified. The most preferred price range is “between $51 and $100” and the best location is “resort area.” It was discovered that the hybrid conjoint model predicts meeting planner’s preferences better than the other two models in terms of their R2, internal validation, and external validation. The study also suggested that there is no difference between one-stage and two-stage hybrid conjoint analysis based on their R2 and external validation. The author successfully demonstrated some applications of the hybrid conjoint analysis. They included market simulation (competitive analysis), own- and cross-price elasticity estimations.
TABLE OF CONTENTS
LIST OF TABLES ---- viii
LIST OF FIGURES ---- x
ABSTRACT ---- xii
CHAPTER 1 - INTRODUCTION ---- 1
1.1 The Meeting Industry ---- 2
1.1.1 The Lucrative Market ---- 2
1.1.2 The Demand Side and Supply Side ---- 3
1.2 Problem Statement ---- 8
1.3 Rationales of Using (Hybrid) Conjoint Analysis ---- 11
1.4 Research Objectives ---- 12
1.5 Research Assumptions ---- 13
1.6 Research Hypotheses ---- 14
CHAPTER 2 - LITERATURE REVIEW ---- 15
2.1 Utility Functions ---- 16
2.1.1 Two Cautions ---- 16
2.1.2 The Utility Models ---- 17
Vector Model ---- 17
Ideal-point Model ---- 19
Part-worth Model ---- 21
2.2 Conjoint Analysis ---- 24
2.2.1 Nature of the Conjoint Analysis ---- 24
2.2.2 Evolution and Classification of Conjoint Analysis ---- 25
2.2.3 Fractional Factorial Designs ---- 28
2.2.4 Balanced Incomplete Block (BIB) design ---- 33
2.2.5 Self-Explicated Model ---- 35
2.2.6 Regular (Traditional) Conjoint Model ---- 37
2.2.7 Hybrid Conjoint Model ---- 41
2.2.8 Model Validation ---- 45
2.2.9 Choice Simulation ---- 51
First Choice Model ---- 51
Share of Preference Model ---- 52
Logistic Model with Correction of Product Similarity ---- 55
Purchase Likelihood Model ---- 55
2.3 Conjoint Applications ---- 56
2.3.1 New Product/Concept Identification ---- 56
2.3.2 Market Segmentation ---- 57
2.3.3 Competitive (“What-If” Questions) Analysis ---- 59
2.3.4 Price Elasticity Estimation ---- 68
2.4 Conjoint Research in the Hospitality Industry ---- 74
CHAPTER 3 - RESEARCH METHODOLOGY ---- 76
3.1 The Sample ---- 77
3.2 Identification of Attributes in Hotel Site-Selection Decision ---- 77
3.2.1 The Pilot Test ---- 79
3.2.2 Sample Selection ---- 79
3.2.3 The Procedures ---- 80
3.3 Questionnaire Design ---- 81
3.3.1 Profile Configuration ---- 81
3.3.2 Random Blocking Samples ---- 83
3.4 Data Collection ---- 83
3.5 Data Processing and Analysis ---- 85
3.5.1 Data Format ---- 85
3.5.2 Computer Software ---- 85
3.5.3 Model Construction ---- 86
3.5.4 Dummy Variable Coding ---- 90
3.5.5 Model Validation ---- 93
3.5.6 Market (Choice) Simulation ---- 94
3.5.7 Price Elasticity Calculation ---- 98
CHAPTER 4 - RESULTS ---- 100
4.1 The Pilot Study ---- 101
4.2 Data Collection ---- 104
4.3 Model Construction and Comparison ---- 105
4.3.1 Self-Explicated Model ---- 105
4.3.2 Regular Conjoint Model ---- 107
4.3.3 Hybrid Conjoint Model (One Stage) ---- 109
4.3.4 Comparison Between Two Hybrid Conjoint Models (One- and Two-Stage Models) ---- 112
4.4 Model Validation ---- 115
4.5 Market Simulation (Competitive Analysis) ---- 117
4.6 Price Elasticity ---- 119
CHAPTER 5 - SUMMARY AND CONCLUSIONS ---- 122
5.1 The Best model ---- 123
5.2 The Most Important Attribute ---- 125
5.3 The Most Preferred Levels ---- 127
5.4 The Most Preferred Hotel Profile ---- 135
5.5 The Competitive Analysis (Market Simulation) ---- 137
5.6 The Price Elasticity ---- 140
5.6.1 Price and Preference ---- 140
5.6.2 Own-Price Elasticity ---- 142
5.6.3 Cross-Price Elasticity ---- 147
CHAPTER 6 - RECOMMENDATIONS ---- 149
6.1 For Hotel Management ---- 150
6.2 Limitation of the Study ---- 152
6.2.1 More Accurate Preference Predictions ---- 152
6.2.2 Cost-Effective Concerns of Conjoint Research ---- 153
BIBLIOGRAPHY ---- 155
APPENDICES ---- 167
Appendix A. Conjoint Software ---- 168
A-1. Commercial and Academic Conjoint Software ---- 168
A-2. The General Features of Commercial Conjoint Software ---- 170
A-3. Current Software Offerings in the Market ---- 172
Appendix B. Orthogonal Plans for Asymmetrical Factorial Designs ---- 181
Appendix C. Pilot Study ---- 182
C-1. Pilot Cover Letters and Questionnaire ---- 182
C-2. PILOT59 DATA A (Pilot Test Data) ---- 186
C-3. Pilot Test SAS Program (PILOT SAS A) ---- 187
C-4. Attribute Level Development ---- 188
Appendix D. Questionnaire Design ---- 189
D-1. Fractional Factorial Design Program - DESIGN6.DES (A Conjoint Designer’s Program) ---- 189
D-2. An Example of Non-Holdout Questionnaire ---- 191
D-3. An Example of Holdout Questionnaire ---- 200
Appendix E. Cover Letters for Mail Survey ---- 206
E-1. The First Cover Letter ---- 206
E-2. The Reminding Cover Letter ---- 207
Appendix F. SAS Programs ---- 208
F-1. RANDOM SAS A ---- 208
F-2. RANDOM1 SAS A ---- 209
F-3. DW136 DATA A (Partial Self-Explicated data) ---- 210
F-4. EV136 DATA A (Partial Profile Evaluation Data) ---- 211
F-5. DUMMY SAS A (Reference Dummy Identification) ---- 212
F-6 DUMMY LISTING A (Reference Dummy Identification) ---- 215
F-7. DYMCODE SAS A (Dummy Variable Assignment) ---- 216
F-8. DYMCODE LISTING A (Dummy Variable Assignment) ---- 218
F-9. RC SAS A (Regular Conjoint Program) ---- 219
F-10. SFX SAS A (Self-Explicated Program) ---- 221
F-11. HYB SAS A (Hybrid Conjoint Program) ---- 223
F-12. ATTCOMP SAS A (Attribute Comparison Program) ---- 226
F-13. ATTCOMP LISTING A (Attribute Comparison SAS Output).....227
Appendix G. Validation Programs ---- 230
G-1. RCINT SAS A (Regular Conjoint Internal Validation) ---- 230
G-2. SFXINT SAS A (Self-Explicated Internal Validation) ---- 232
G-3. HYBINT SAS A (Hybrid Conjoint Internal Validation) ---- 235
G-4. RCEXT SAS A (Regular Conjoint External Validation) ---- 238
G-5. SFXEXT SAS A (Self-Explicated External Validation) ---- 240
G-6. HYBEXT SAS A (Hybrid Conjoint External Validation)...............243
Appendix H. Choice Simulation (Competitive Analysis) Program: HYBSIM SAS A ---- 246
Appendix I. Elasticity Estimation ---- 249
I-1. Elasticity Estimation (ELAST SAS A) ---- 249
I-2. Own-Elasticity Estimation (ELAST LISTING A) ---- 252
I-3. Cross-Elasticity Estimation (CRSELAST SAS A) ---- 253
I-4. Cross-Elasticity Estimation (CRSELAST LISTING A) ---- 257
Appendix J. Partial Output of SAS Programs ---- 258
J-1. SFX LISTING A (Partial) ---- 258
J-2. RC LISTING A (Partial) ---- 259
J-3. HYB LISTING A (Partial) ---- 260
J-4. HYBCOMP SAS A Program ---- 263
J-5. SFXINT LISTING A: Internal Cross Validation for Self-Explicated Model (Partial Output) ---- 267
J-6. RCINT LISTING A: Internal Cross Validation for Regular Conjoint Model (Partial Output) ---- 270
J-7. HYBINT LISTING A: Internal Cross Validation for Hybrid Conjoint Model (Partial Output) ---- 271
J-8. SFXEXT LISTING A: External Cross Validation for Self-Explicated Model (Partial Output) ---- 272
J-9. RCEXT LISTING A: External Cross Validation for Regular Conjoint Model (Partial Output) ---- 274
J-10. HYBEXT LISTING A: External Cross Validation for Hybrid Conjoint Model (Partial Output) ---- 276
Appendix K: BEST SAS A: (Best Profile Identification) ---- 278
LIST OF TABLES
Table 1.1 Annual attendance and room nights used by different types of meetings held ---- 3
Table 2.1 Different steps and alternative methods in a conjoint analysis ---- 19
Table 2.2 Comparisons of six factors on three preference models ---- 34
Table 2.3 An example of three choice validation models ---- 41
Table 3.1 Dummy variables’ labels were translated to attribute levels ---- 77
Table 4.1 The pilot study results ---- 85
Table 4.2 The distribution of valid response rates ---- 86
Table 4.3 The results of Self-Explicated model ---- 88
Table 4.4 The results of Regular Conjoint model ---- 89
Table 4.5 The results of Hybrid Conjoint model ---- 91
Table 4.6 Comparison of two hybrid models ---- 92
Table 4.7 External validation of two hybrid conjoint models ---- 93
Table 4.8 Results of the internal and external validation ---- 94
Table 4.9 The results of market simulation ---- 95
Table 4.10 Prices in elasticity calculation ---- 96
Table 4.12 Price elasticities for Hotel H in different price ranges ---- 97
Table 4.13 Market shares in different price conditions ---- 97
Table 5.1 An external validation comparison ---- 98
Table 5.2 Relative importance of six attributes ---- 101
Table 5.3 The most preferred level in each attribute (Self-explicated model) ---- 103
Table 5.4 The most preferred level in each attribute (Regular conjoint model) ---- 104
Table 5.5 The most preferred level in each attribute (Hybrid conjoint model) ---- 105
Table 5.6 Adjusted market shares by common factor ---- 109
Table 5.7 Basic information: the Hotel H’s prices and preferences ---- 112
Table 5.8 Cross elasticities for each hotel profile ---- 116
LIST OF FIGURES
Figure 1.1 The demand side and supply side of the meeting industry ---- 4
Figure 2.1 The relationship between preference utility and the ith level of attribute j ---- 13
Figure 2.2 Two concepts in the ideal-point function. ---- 15
Figure 2.3 The part-worth model ---- 17
Figure 2.4 A Taxonomy of utility estimation methods ---- 20
Figure 2.5 The major classes of fractional factorial designs. ---- 23
Figure 2.6 Underlying concept of competitive analysis in conjoint analysis study ---- 50
Figure 2.7 Relationship between prices and the first-choice counts ---- 53
Figure 2.8 Relationship between price level and market share (by first choice model) ---- 54
Figure 2.9 Relationship between price range and market share (by probabilistic model) ---- 56
Figure 2.10 Price elasticity concept in conjoint study ---- 59
Figure 2.11 Own-price elasticity (Hotel H) ---- 60
Figure 3.1 Flow chart shows the profile configuration procedures ---- 68
Figure 5.1 Internal (upper) and external (bottom) validation of three preference models ---- 99
Figure 5.2 Attribute relative importance ---- 101
Figure 5.3 Utilities of 22 Levels (derived from the Self-explicated model) ---- 106
Figure 5.4 Utilities of 22 Levels (derived from the Regular Conjoint model) ---- 106
Figure 5.5 Utilities of 22 levels (derived from the HCA model) ---- 107
Figure 5.6 Market shares in hypothetical scenario: base case ---- 110
Figure 5.7 Market shares in hypothetical scenario #1 ---- 110
Figure 5.8 Market shares in hypothetical scenario #2 ---- 111
Figure 5.9 The relationship between price and preference of the Hotel H ---- 112
Figure 5.10 Market share distribution when Hotel H changes its price levels ---- 114
Figure 5.11 Hotel H’s own-price elasticity (in conjoint study) ---- 114
Figure 5.12 Hotel H’s own-price elasticity ---- 115
|
