Tobacco Documents Online

io a) - 19- ALLEGED IRRITATION/BITTERNESS/DRYNESS OF SOUZA CRUZ PRODUCTS First step should be to check whether this is a true sensory observation, or a 'perceived' effect based on overreaction to consumer complaints/spontaneous co~,~ent data, and/or a brand image effect. The brand image measurement model can be used to show any consistent sensory biases relating to Souza Cruz products when presented in branded form in contrast to blind. It may be recalled, for example, that UK low delivery products when presented branded in the image measurement model'show a clearly consfs=enC bias to perceived incTeaslng irritation. b) This brand image measurement exercise should be conducted as part of an In-house DELTA market mapping exercise where the population of cigarettes used is representative of major brands from the leading manufacturers in the Brazilian market. Souza Cruz products should cluster out toward the irritation parameters if a true sensory problem is present. Note that extended cirritation vocabulary should be used in such a DELTA exercise (see next point). c) Upon smoking Souza Cruz products (notably Free v Galaxy and JPS v Marlboro) it was apparent that although irritation differences in terms of magnitude are not notable, there are some differences in terms of irritation quality. This is an important observation since conventional sensory/product tests only use magnitude scales and may therefore miss or obscure genuine irritation quality differences. 'Bitterness' for example may reflect a spontaneous'consumer attempt to express differences in irritation quality in a sitdation where no BATCo document for Legal Services : Health Canada 18 May 1999 O LJ, J (DO

- 20 - differences necessarily exist in terms of irritation amount or magnitude. It is therefore strongly advised that irritation quality scales are incorporated in Souza Cruz In-house testing and consumer product test approaches. The appropriate attributes to be considered (using graphic scales) would be PITCH and DURATION. d) In the short term, pair~ise comparisons of JPS/Marlboro/Pall Mall and Free/Galaxy could be made using OMEGA with the extended irritation vocabulary. In all cases of extended placement sequential monadlc exercises it is essential that an appropriate competitor referent product be included in the exercise (this point partlcularly refers to evaluations of 'Free' variants). e) In addition to irrltatlou quality observations during smoking of Souza Cruz products, some physical quality differences were noted. It is advised that the R&D OMEGA panel is used routinely to evaluate visual- tactile performance of Souza Cruz products compared with the opposition. f) In the event that these exercises ludlcatlng genuine sensory problems in the irritation area for Souza Cruz products, then adoption of the FTIR approach would be strongly indicated as the firs= approach to determining the objective (blend chemistry) basis for the subjective (irritation) problem. Precedents for this approach are now established in the U.S.A., Australia and Malaysia. .% o , ~UD BATCo document for Legal Services • Health Canada 18 May 1999

- 21 - . a) PROJECT TROPI~LIA Since a prime objectlve must be to determine flavour quality/level of appilcatlon i~teractlons, a number of samples at different dose levels per flavour require to be evaluated. This points clearly to the use of the in-house DELTA panel for preliminary screenlng. It is recommended that a flavour quality parameter be built into the DELTA panel procedure, and each flavour type be evaluated In separate exercises to include several levels of application and an untreated control. DELTA plots for each flavour type can be subsequently consolidated statistically into one overall plot: i , , ,,,, | , , PLOT 1 I PLOT 2 ! Flavour "A" at 5 levels of application~ plus I control. . Flavour "B" at 5 levels of application plus I control. COMPOSITE PLOT PLOT 3 Flavour "C" at 5 levels of application plus i control. b) Since sldestream irritancy/aroma amelioration is also claimed for these materials, a prime objective should be to flnd the trade-off point for each flavour between mainstream and sldestream qualities: Unobtrusive flavour in mainstream Negative or loss of sidestream benefits Obtrusive flavour In • mainstream ~:. •." Positive sldestream .b4nefi ~9. o . P~O :C~O --O CD BATCo document for Legal Services • Health Canada 18 May 1999

- 22 - It is recommended to convert two cubicles in the R&D panel room to sidestream evaluation cubicles, which can be used together in control v experimental dose/response studies of sidestream aroma and Irritancy. c) In the event of Tropicalla products subsequently being submitted for consumer product testing, it is advised that qualitative focus groups will be an important precursor to any product testing. These groups should comprise five types: i) Low Tar Smokers ll) ill) Males Full Flavour Smokers <Females Smoker/Non Smoker Groups o a) HOLLYWOOD BLEND RE-POSITIONING Clearly on such a major brand any blend changes would be advi%ed to be circumspect and guided as clearly as possible by an adequate picture of the blend 'landscape' of the market. The first step in mapping this landscape may be to treat the routine Brazilian Blend mapping work conducted in house with a principal components analysis. However, a shortcoming of this approach is that it will not include a preference vector and will, therefore, only have descriptive value. A parallel approach conducted in the marketplace and including preference questions may give the necessary map, and should indicate the preferred directlon, of Holl~o6d blend reposltlonlng. ! CD rxO BATCo document for Legal Services • Health Canada 18 May 1999

- 23 - Once the direction of re-positioning is determined, the same plot can be used as a background into which evaluations of prototype products may be incorporated to show freshly achieved positioni=gs in relation to the current Hollywood position and the desired new position. Since it may be desired to move to a new position in stepwlse changes, In-house duo-trio testing can be employed to ensure that any changes are controlled to near threshold levels of perceived significant difference. . a) PROJECT SHERIFF Since brand imagery response is generally a powerful component of the Marlboro following, it is advised to use brand image mapping as a background data base to inform the project. Ideally this should be achieved through using image attribute vocabulary elicited from consumers using repertory grid interview and consensus groups. It should be possible to generate three maps if the interviews are focussed appropriately: i) general imagery ii) expected smoker type ill) expected smoking experience. Maps can be produced by segment as well as for the overall respondent population, and by comparing these maps it should be possible to derive an understanding of relative image responsiveness in the market. Suggested segmentation would be initially by age~ sex~ soclo-economlc class and brand smoked. rxo BATCo document for Legal Services • Health Canada 18 May 1999

b> Motivational and ~motlonal response to b~and is also an important feature of the Marlboro phenomenon. In-depth motivational techniques can be employed to examine consumer response to Souza Cruz brands contrasted wi~h Marlboro. Most notable of these are Transactional Analysis and EPSY, documentation on these approaches has been supplled to Souza Cruz and iE is advised that EPSY is likely to be a more productive approach in This case. "'.. ". ..- 0 ,-~D BATCo document for Legal Services • Health Canada 18 May 1999

RPF/JP ~jc~/,i s~d v,, ~x ~. ,) ~.i .... STRICTLY CONFIDENTIAL v 16th July 19B5 VISIT REPORT: DATES: OBJECTIVES: ITL, CANADA gTH-IOTH JULY 1985 (i) To critically review draft questionnaire design for Project VIKING (2) To participate in a one-day synectics session designed to generate (a) product perception attributes for use in the questionnaire and (b) novel alternative product concepts for use in an additional concept testing phase of Project VIKING BACKGROUND # Project VIKING was initiated in early 1985 under the auspices of Wayne Knox, with broad objectives as Follows: i. To characterise the Canadian market with respect to smoking and health and social pressure attitudes and motivations. . To examine relative perceptions of product classes within the tobacco category (e.g. cigarettes, cigars, pipes, snuff, etc.). . To examine consumer perception of the tobacco product category relative to other product categories (e.g. food, alcoholic drink, personal accessories, etc.). To evaluate potential consumer response to radical alternative presentations of the tobacco product. Implied in this approach is an attempt to provide a prognosis of the future Canadian tobacco market. Philosophically, the approach differs from the GR&DC LIBRA/ARIES/NARINERS work in two respects: (1),. NARINERS is designed around the contention that, to fully understand the relationship between attitudes and behaviours in the marketplace (i.e. quitting and switching), it is necessary to monitor the same group of people across time. VIKING takes the alternative route of using a large population in a "single shot" model. (z) Whereas VIKING is essentially a prognostic indicator with an emphasis on seeking new product categories, NARINERS is aimed at developing more refined and predictive consumer segmentation models to improve share within a conventional cigarette marketing scenario. Prior to Wayne Knox joining Hardees, the agreed budget ~or VIKING was in the order of 300,O00"Canadian Dollars. Bob Bexon now assumes executive responsibility for the project and tee budget, As still retained pending a feasibility review. C~ r~ "-C) BATCo document for Legal Services • Health Canada 18 May 1999

-2- OUTCOME OF CURRENT MEETING i, QUESTIONNAIRE DESIGN The first draft questionnaire for VIKING has been produced by The Creative Research Group of Toronto, with input from a previous VIKING meeting (February 1985) and a series of consumer qualitative group sessions. This draft was reviewed (RPF) and the following observations have been made: (a) The questionnaire is too long (approximately 2-hour administration time) and alternative means of administration will have to be explored (including repeat visit and postal returns). (b) In current form, the questionnaire fails to cover several of the issues originally identified as important to measure. This problem can be resolved by building out to a questionnaire size unconstrained by timing issues and then reducing by eliminating redundant measures and non-discriminating or merely descriptive status items. (c) The inclusion of two lengthy personality inventories was questioned since one questionnaire (EPI) is designed to measure a factor already measured within the other questionnaire (16 P-F). It was advised to use only the 16 P-F and as a postal return measure. 2. SYNECTICS SESSION This synectics (brain storming) group comprised the following individuals: BOB BEXON DICK CROSBY ROB FERRIS PETER ILLYCH BILL SANDERS PAUL NESBITT (ITL, Montreal) (Retired President of Canadian Facts) (BAT Marketing Support) (ITL, Montreal) (ITL, Montreal) (Creative Research Group, Toronto) The first half of the one-day session was devoted to developing the product perception component of the VIKING questionnaire. This involved the generation of judgemental dimensions which may be used to achieve (a) a perceptual mapping of tobacco product classes relative to one another and (b) a similar mapping of tobacco products usage relative to uses of products from other classes (food, drink, etc.). Following this construct generation stage, the group was used to generate new a]ternative product concepts. The aim was to identify'six concepts which will be incorporated into a concept reaction test within VIKING. The test itself is of considerable interest sinceitrepresents a form of simulated test market (STM) methodology, an approach current]y being evaluated (RPF) for its potential utility to the BAT Group. r~ r~ l BATCo document for Legal Services : Health Canada 18 May 1999

-3- Briefly, the new product concepts generated were segregated into those showing most promise and those having less promise: MOST PROMISING io EXTRUDEDS: Th~s covers a collection of presentations ranging from snack food type shapes and packaging through chewing sticks to reel dispensers or toothpaste tube type dispensers. In each case the notion of individual control over how much of the material to break off and use is paramount. . DRY SNUFF: Perhaps the most clearly endorsed concept of all, this product would be presented bleached white and probably in an inert carrier, formulated to eliminate the sneeze reflex and marketed in high imagery/qua|ity containers. D MODULAR SMOKE DELIVERY SYSTEM: This concept would be a smoking device somewhat similar to a pipe into which the basic fuel could be mounted for combustion. Modular "clip-on" mouthpieces could be used to provide such smoker control options as "dial-a-tar" or "dial-a-flavour". . PIPE PLUGS: For the pipe smoker, this concept would be a pre-formed, pre-tamped, drop-in tobacco plug designed to fit the appropriate pipe bowl and sufficiently contained within the plug to allow subsequent lifting out of the spent unit. LESS PROMISING . AEROSOL DISPERSION DEVICES: The concept covers personal mobile inhalers in combination with nicotine dispersing air humidifier/ freshener units which could be used in controlled social situations. o IMPROVED SMALL CIGARS AND AROMATIC CIGARETTES: The use of blending/ flavouring to ameliorate harshness/dryness and enhance flavour/aroma of small cigars was seen as worthwhile. In addition, the use of pipe grades or pipe tobacco flavours (e.g. vanillin) to enhance conventional cigarette sidestream was endorsed. DISCUSSED AND REJECTED 7. NICOTINE ADDITIVES TO FOOD PRODUCTS . SHORT "FAST HIGH" CIGARETTES: I.e. cigarettes designed to provide optimum nicotine requirement in the minimum number of combusted puffs. In addition to these ideas, strong interest was shown in the identification of "safer" tobaccos, regions, design configurations and synthetic smoking materials. (RPF was requested to supply a retrospective review of the NSM experience in the UK.) " .... - "" 0 r~o r,o Ir~j ,.o r~o BATCo document for Legal Services • Health Canada 18 May 1999

-- 4 - OBSERVATIONS Without substantial revision of the VIKING questionnaire, this project is unlikely to proceed (and it would be ill-advised to do so). The agency involved was clearly so negatively impressed by the overwhelmingly anti- smoking feeling coming from the consumer qualitatives that they themselves are concerned about taking the project further. It is suggested (RPF) that, with adequate questionnaire design, the project is important and viable. The agency response to the consumer group qualitatives reflects lack of experience with this type of data - such groups are invariably extreme, representing, as they do, consumers' theories about what their attitudes should be rather than their true in-depth attitudes. The "re-surfacing" of a concern for generating new product concepts is interesting in the context of the R&D/Marketing conference experience. Some overlap of concepts is apparent across the two sessions; the interesting development here, however, is the commitment to moving on to evaluate the concepts using the VIKING consumer population. In terms of possible "knock-on" project work for GR&DC, the scope for input is very high, given past recent history of exploring these new avenues. While Southampton-based interest in these concepts must be highly circum- spect, given the confidential nature of VIKING, a potential feasibility response timed to follow VIKING consumer response to these concepts may be useful. This could be especially so for items I, 2 and 5 since these were Southampton contributions to this synectics session. R.P. FERRIS 0 Ir~_) r~ L~J BATCo document for Legal Services • Health Canada 18 May 1999