Tobacco Documents Online

THE INFLUENCE OF BRA.N'D IDENTIFICATION AND IMAGERY ON SUBJECTIVE EV.4LUATION OF CIGARETTES REPORT NO. RD.1752-C RESTRICTED 18.7.1980 AUTHOR: R.P. Ferris GROUP LEADER: H. Oldman ISSUED BY: D.J. Wood DI STRIBUTION: Dr. L.C.F. Black~mn Dr. I.W. Hughes Dr. R.A. Sanford R.M. Gibb, Esq. R.S. Wade, Esq. R.G. Nicholls, Esq. Herr E. R/~ershaus Dr. F. Seehofer Dr. C.J.P. de Siqueira Mr. A.J. Kruszynskl Dr. D.G. Felcon Library ClibPDF - www.fastio.com Copy NO. I, 2, 3, 4, 5, 6 It II 7 " " 8, 9 .... I0 " " ii, ]2, 13 " " 14, 15 " " 16 " " 17 " " 18 " " 19 " " 20 " " 21, 22 COPY NO.~7 PkD rko rk~ -.,j

Group Research & Developmen~ Cent;e, British-American Tobacco Co. Ltd., SOUTHAMPTON. 18th July 1980. TH~ INFLUENCE OF BRAND ~DENTIFICATION AND IMAGERY ON $UEIECTIVE EVALUATION OF CIGARETTES (Report No. RD.1752-C &estricted) SUHMARYAND CONCLUSIONS In a future where Increasingly the product may have to sell itself through the pack, a fuller understandinE of the way in which perception of such packs affects perception of their contents is desirable. In this report a technique is demonstrated which allows controlled measure=ent of the ex=ent and quality of influence which brand identification and pack image have on the individual's impression of a cigarette's smokinE characteristics° It is shown that, for some products, such imagery powerfully and measurably modifies the perceived smoking characteristics of the cigarettes associated with i:. Further research is intended to determine both underlying bases of pack image influence (e.g. colour, pattern, etc.) and levels of responsiveness within ~he consumer population to the influence of imagery.. Application of the technique to the screeninE of pack design alternatives for products in current development may now be considered. ClibPDF - www.fastio.com P'O F,,D P',D ,..,,Q

-2- /-- INTRODUCTION The way in which a product is perceived and ultimately received in the marketplace relies on two basic processes: (a) EVALUATION: the individual's ability to obtain measures of that product's physical or sensory at~rlbu~es (cos~, size, welgh~, caste, smell, etc.)° (b) IMPRESSION: the individual's interpretation of the "image" of the produc~ (the term "image" usually being defined and discussed in the marketing literature as an abstrac~ concept incorporating the influences of brand name, pack design, paso promotion and reputation). The tobacco industry has developed several means of appraisimg and quantifying the evaluative component of bo~h in-house and consumer product perceptions. These means are typified respectively by panel testing procedures of ~he WOODROSE and DELTA type, and by a range of consumer product tests. It is accepted practice Eo carry out produc~ evaluations in "de- idemtlfied" format; that is, using blank whlce packs and clgarecces on which brand identlficacion markings have been obliterated or "masked" by a variety of ~echniques (usually overprinting or taplng~ which differ in complexity according to the type and distribution of identifiers present on the cigarette. Such precautions are deemed necessary in order Ehau the potential effect of brand image factors should not in£1uence the results of the smoking evaluation. ClibPDF - www.fastio.com F~D r-,D r-o ---.,4 p,D

-3- Brand de-!dentlficatiou procedures are generally costly and time consuming, co~ronting the product tester with three serious problems. First, for any given set of cigarettes under test, since it is desirable that the masking format is common to all, the amount and complexity of masking used is dlc~ated by the worst case (i.e. the cigarette with mosu identification marking). Second, the physical appearance .of cigarettes which have been masked is altered and it is possible that this sort of manipulation affects the tester's view of the product. Third, and most importantly, it is likely that current masking practice (particularly tape which can obliterate perforations and adversely affect paper porosity) alters the sensory smoking characteristics of cigarettes. It should therefore be understood that the precautions which the industry takes in order to control the potentially confounding influences of brand identification/imagery from the sampling situation, themselves generate variabilities which are uncontrolled. It is the purpose of this report to demonstrate that the variability which brand identificatlon/imagery contributes to the sensory evaluation situation may in fact be quantifiable, and the extent of this variable influence sensibly related to the sensory parameters on which a given product may be evaluated. In short, the report will show how hltherco poorly controlled or ignored influences bearing on produc~ perception (imagery variables) mlgh~ usefully be incorporated into the integrated empirical approach to product development, providing a further stage of refinement in the development of a productfs market aim. ....a. 0 r~D r~ ~0 ClibPDF - www.fastio.com

BACKGROUND The research literaEure dealing wlch Ehe iuf!uence of brand imagery on produce perception is dominated by hypotheses generated from psychology theory, parrlcularly those dealing w~th the concept of "self" (I). ~u terms of this theory, an individual's behavlour is largely a fuuction of his self image, such that he will tend to behave in such a manner as to maintain and enhance his view of himself. As a resulE, consumer preferences should tend to be determined, not so much by the product's objective properties, but by the relevance of these properties for Ehe individual's aECempt to maintain and enhance self image. This should mean uhaE products which complemeu~ the individual's self image should be preferred to Ehose which are seen to fail to complement it. ~ence individuals should prefer those brands in a given product category which have images slm/lar to their self image rather than those which have dissimilar images. (This prediction would be equally supported by cognitive dissonance theory which suggests that when an individual becomes aware tha~ his expectations differ from his perception of reality, dissonance occurs and Ehe individual will seek co reduce his dissonance by aligning his percepEions to his current choice or by seeking a new alternative.) These theoretical predictions have been tested, notably by M arEineau (2), Westfall (3), and Evans (&), who each confirm that, faced wi~h brand alternatives in a given produce category~ consumers tend to prefer the one whose image is most similar to their self image. Evans~ for example, found that he could discriminate owners of Ford and Chevrolet cars to a statls~ically significant degree with Ehe use of personallty variables, ClibPDF - www.fastio.com

-5- demographic variables, or both in combination. Similarly, Reid and Buchanan (5), and Woodside (6), for cac foods and beer brands respec=ively, find that consumers attribute different personality characteristics to purchasers of different brands. Interestingly, cigarettes are a favoured subject for the study of brand imagery, largely because the user is free to purchase in response to individual needs without serious complications due to different prices, availability, or trade-offs among joint users or uses. Fry (7), for example, finds brand preference for cigarettes on the Canadianmarket to be related to their perceived image (determined by non- metric scaling to fall into two groups, one of which was seen to be milder, more feminine and more elegant than the other)° Again, the influence of the imagery was importantly related to a number of personality measures including social confidence, masculinity/femininity and need for change, these being further moderated by socioeconomic variables. In the same vein, Vitz and Johnston (8), investigating the relationship between masculinity of smokers as measured by personality inventory and rated masculinity of brand images, found for both male and female groups a low but statistically significant positive correlation between Eheir masculinity and the rated masculinity of their regular brand cigarette. Friedman and Dipple (9) have tested this relationship empirically in a study designed to determine the influence of brand name on consumer product evaluation. ~n this case all subjects smoked identical cigarettes. However, one group of subjects knew the cigarettes by the name of "Frontiersman", a masculine name, while the other group knew the cigarette as "April", a feminine name. Male and female subjects were asked to rate O F~D F~ ~O ClibPDF - www.fastio.com

-6- the cigarette on seven sensory attribute measures and the results showed chac women gave a more positive evaluation to the cigarette purportedly named '*April", while men Save a more positive evaluation to the identical cigarette when the7 thought ic was named "Frontiersman". in addition, women reacted more s~ronsly than did men to brand name influence. Orpen and Low (i0) have united the type of research conducted Into car and cigarette imasery by Evans (4) and Vitz and Johnston (8) respectively, in a single experiment which sugsests that the effects of brand imagery, and in particular, the way in which the consumer's psychological disposition relates to that imagery, may be Eeneralised across differing product categories. In this case, correlations were computed between subject's self ratinS of masculi~/ty, an inventory measure of masculini=y and the rated masculinity of inm~e of the subject's most liked and disliked cigarettes and motor cars. The correlation between the self racine and the inventory measure of masculinity (r = 0.75) was significant at the 0.01 level, and as predicted the correlations between both measures and the rated masculinity of their most liked clgarette were also significant at the 0.01 level. The same relationship held for mos~ favoured motor cars. In addition, the mean masculinity rating of the most liked motor cars was significantly higher (p < 0.01) than tha~ of the most disliked motor cars, and the same relationship held for ~he cigarettes. This indicates tha~ the cars and cigarettes with more masculine images were preferred to those wlth more fem/nine ImaEes despite the fact that the sample cqusisted of au equal uumber of males and females (this fiudinE relates well to the fact that sales of Marlboro, a cigarette with a distinctively masculine image, are supported by a large female contingent). "-4 C~ ClibPDF - v.dv,,/vv.fas~[io.col]]

-7- Within GR&DC t-ao reporued studies (L1, I2) have exam:L~ed 4£scri=in,~tions made by ten subjects of £1fty cigarette packs from the Malaysian market, following a pilot study of ~he discriminations of one individual on Guatemalan packs. The data in this case was analysed by multi-dimensional scaling from whlch a t~o-dimensional model of discrimination was suggested, each axis being tentatively labelled on the basis of subjective and objective characteristics. Axis I appears to refer to "Io= colour contrast - high colour contrast" and "old fashioned - modern" and axis II might be labelled "small pack size - large pack size", "cheap - expensive" and "low class - high class". With the introduction of repertory grid research to GP.&DC, imagery Investigation was extended to three areas: (a) BRAND IMAGE: the mental associations stimulated solely by the pack appearance. (b) INFERRED SMOKING CE4RACTERISTICS: the type of smoking experience which might be anticipated (in the absence of sampling) from a cigarette contained in such a pack. (c) SMOKER ATTRIBUTION: the anticipated personality/character ~ype of ~he sort of person who migh~ be expected to regularly smoke clgare~tes contained in such a pack. This report (13) demonstrated that a pack not only generates powerful independent images, bu~ also provides important and predictable cues or suggestions about the ~ype of smoke which may be expected from a cigarette contalned is such a pack, and even the type of person who might typically smoke such cigarettes. The power of brand image in inferring smoking characteristlcs is clearly demonstrated in the case of Marlboro, a brand ~h a s~rong prominent image which almost invariably suggests, even to ClibPDF - ww~,v.1astio.conl

-8- those who actually smoke the brand, char the cigarette is fully middle tar in strength, hence belying the fact that the product is lo~-mlddle tar. From this position ic is a short step to the understanding that brand image factors will significantly modify the individual's perception of the actual smoking attributes of his cigarette, supporting tentative research by D~qare (14) indicaEing tha~ differences in objective parameter ratings of "mildness", "distinct taste", etc. depended upon whether consumers were judginE on the basis of their impression of the cigarette (a) unsmoked but with brand image prominent, (b) as smoked and accompanied by brand image, or (c) as smoked in the absence of any cues relating to brand. The possibility of measuring the influence which brand image has in modifying the individual's perception of the "objective" sensory attributes of his clgaret~e is of great importance, particularly if the influence can be quantified with sufficient precision to allow prediction of ~he extent and type of modification from partlcular aspects of the brand image. For example, i~ is already anecdotal knowledge that bright red on a pack connotes a snronger produc~, but it would be considerably more useful it if were possible to enpirically relate such a thing as area of red on the pack to the amoun~ by which the consumer's evaluation of that cigarette's "impact", say, is increased. If the influence of brand image could be rendered quantifiable and predictably related to modifications of the perceived sensory attributes of clgaret~es, we migh~ profitably add "image engineering" to the series of processes by which a cigarette ks designed and developed as a total product. Imagery mlgh~ then be "de-mys~ifled" and made to work An the ClibPDF - www.fasti0.com 0 .,"0 r,o b,o

-9- serv~.ce of complementing or eve= aus~aencing product design c.~aracteriscics. This report demonstrates a possible means of achieving this aim. RESEARCH PROCEDURE The object of this study is co quantify the extent co which brand image variables influence individuals' assessments of the sensory properties of cigarettes° The natural baseline or con=rol against which comparisons may be made is therefore cigarettes with brand identification markings masked and packs absent. Against this baseline, cigarettes may be compared ac two stages of imagery exposure: (a) With brand identification markings only visible, or UNMASKED. (b) Contained in ~he pack from which the cigarette would normally be oaken, or in PACK CONTLKT. The means by which comparisons are made of ~he sensory properties of cigarettes evaluated under these conditions is critical, and must satisfy two criteria° First, it must deal with a useful list of sensory aturibuces, such that any changes in the evaluation of cigarettes across the three experimental conditions may be quantified and rela~ed ~o specified and commonly understood elements of the smoking experience (e°go taste, impact, draw effort). Second, ~he method used mus~ be known to exhibi~ high inter-repllcation trial reliability, such tha~ any deterioration of this in~er-trial comparison (assuminS that the same cigarettes are compared each time) may be confidently attributed ~o the influence of experimen~Rl man£pulatlons of the imagery variable, rather chat to trial-co-trial variability of ~he assessmen~ method itself. ClibPDF - www.fastio.com J ,',o I',,D -'-.4 m "-4 ',,O