sample="quota" bates="675094380" isource="bw" decade="1980" class="ui" date="19821216" John Morton Company A Brand Image Study of the Cigarette Industry Phase I ATTRIBUTE DEVELOPMENT AND SELECTION Brown & Williamson Tobacco Corporation December 16, 1982 P GOALS AND OBJECTIVES TO DEVELOP A UNIVERSAL SET OF ATTRIBUTES, BOTH PRODUCT AND SMOKER, THAT CAN DEFINE, DESCRIBE AND DISCRIMINATE BETWEEN IMAGES OF CIGARETTES, BRAND FAMILIES AND BRAND STYLES. TO SELECT A SUBSET OF ATTRIBUTES THAT DISTINGUISHES MOST CLEARLY AMONG EXISTING PRODUCTS AND THAT MOST DISTINCTLY DEFINES INDIVIDUAL PRODUCTS, ESPECIALLY BROWN & WILLIAMSON'S. BOTH PRODUCT AND SMOKER ATTRIBUTES MUST BE INCLUDED. P METHODOLOGICAL SUMMARY Qualitative Phase Merging of all currently available studies - 1970 Trade-Off Study - 1976 Image Study - 1982 Image Study - German Study Generation of new attributes via group and one-on-one interviews Quantitative Phase Field study of 1000 moderate to heavy smokers Brand images were described using large number of candidate attributes Numerical analysis of attributes - Differentation by Preference - Differentation by Family - Differentation by Style - Ratings of Brown & Williamson brands - Correlations among attributes (summarized as factor analysis) P PERSONAL INTERVIEWS 2 Group Interviews - Moderated by Georgia Bender - Average of 9 smokers per group - One Group Male, one Female - Duration: 2 hours 20 One-on-One Interviews - Moderated by Georgia Bender Research Staff - Interviews split half male - half female - Duration: 1 hour Respondent Selection Criteria - 21-45 years of age - Smokes at least 3½ packs per week - Minority representation - Demonstrates unaided recall of 7 or more brands - Smokes multiple brands or has switched - Demonstrates creativity, spontaneity, and speed P Interview Technique - Word associations with brand names - Smoker descriptions - Famous personality - cigarette associations - Alternate product - cigarettes associations All respondents chosen from Chicago metropolitan area Results - List of words and phrases - Brief brand family profile P FIELD INTERVIEWS Sample Design - 1000 Respondents - 8 cities Milwaukee St. Louis Memphis Albuquerque Seattle Philadelphia Tampa Tulsa - central mall locations Respondent Selection Criteria - demographic distribution male 21-34 27% male 35+ 23% female 21-34 24% female 35+ 26% - smoked 4 or more packs per week - must smoke one of 24 base brand families - must not work for market research firm, nor tobacco company P FIELD INTERVIEWS (CONTINUED) Questionnaire Design - 5 brands selected by respondent primary brand - most smoked pack primary brand - least liked pack least favorite brand previous switched brand or second preference most different from primary brand - 20 random attributes each questionnaire for each of the 5 brands independently computer generated to insure uniform random distribution of all attributes Total Sample Created - 100,000 total responses - across 24 families x 262 attributes - across 131 styles x 262 attributes average ratings per style = 3 average ratings per family = 16 P DATA REDUCTION PHASE I METHOD Calculation of Degree of Discrimination (measured by F ration) across Family (based on 24 aggregate brand families) Style (based on intra-family differences) Preference (Based on 5 brand selections for each respondent) Calculation of Attribute Variance From Mean by Brand SELECTION CRITERIA Selection of Those Attributes that Gave Highest Discrimination on All 3 Categories Independently Selection of Ratings of Brown and Williamson Brands P DATA REDUCTION PHASE I (CONTINUED) RESULTS Reduction of Data to 103 Attributes 34 Product 69 Smoker 8 B & W Brand Rating F Ratio Cutoffs P FINAL ATTRIBUTES P P RESULTS Total Attribute List - 262 (merged past studies and group and one-on-one list) Total Product Attributes = 48 Total Smoker Attributes = 214 Total "New" Attributes = 44 Total from merged lists 271 Total from personal interviews 130 Duplicates 86 P MAIN STUDY 1800 Microcomputer Interviews (in 20 sampling areas) - 900 with B & W smokers; 50 in each of 18 major styles - 900 with competitive smokers; 50 in each of 18 major families During interview, each smoker: - rates custom-designed set of 10 brand styles on each attribute - rank orders the 10 by preference - indicates ("claimed") importance for each attribute - answers classification questions Primary analytic techniques - perceptual mapping, based on attribute rating information - contour mapping to indicate density of demand Timing - Interviewing Jan. 5 - Feb. 4 P - Topline results Mar. 4 - Final Presentation Mar. 31 DATA REDUCTION PHASE 11 FINAL ATTRIBUTE SELECTION Procedure for final selection from list of 103 as follows: Factor analysis run on set of 103 - accounted for 80% of the variance - used 10 factors Highest "+" loading and "-" loading attributes selected on each factor Best product and best smoker selected from each - selection based on best discriminators highest correlations Final reductions done by removing very highly correlated pairs (+ and -) B & W brand rating attributes forced on final list P P P P P P P P P P P