Tobacco Documents Online

Visit to Advanced Tobacco Products, San Antonio, Texas~ 19th Februar7 ~985 Compiled b7 : A. L. Heard Present: N. Jacobsen E.G. Vfmond J.P. Ray I. Hill J.D. Simonsen C. Warren J. Unlack J.A.B. Kellagher R.P. Ferris A.L. Heard Chairman, C.E.O., A.T.P. Chairman of Marketing Commltcee President V.P. Technology V.P. Administration T.T.L., Canada BATCo Ltd. BATUKE, R&D BATCo Ltd. Back~roun~ In November 1985, C. Warren and A.L. Heard visited ATF to explore whether the potential existed for some form of collaborative arrangements between their respective companies and ATP, who had recently launched FAVOR, a smokeless cigarette. At that time, ATP foresaw no problems in covering distrlbutlon through USA but were interested In finding partners who could develop non-USA markets. They were well disposed to the BAT Group as a potential partner. Products were-supplied to both ITL and BAT foI consumer test work. BAT and ITL agreed that initially BAT, through the facilities available to R.P. Ferrls of the BATUKE ReD Laboratory, would conduct a detailed but small scale evaluation of regular FAVOE with a brief assessment of some product developments made available during the visit co ATP. Additionally, B&W would be asked to monitor the performance of FAVOR in the Texan cities where it was being launched. Purpose and Scope of Meetin,~ In the period between the November visit and this visit, ATP had suggested to ITL that they had further developments on FAVOR including some which related to conventional smoke generatlng products. At preseat, in the secrecy agreements signed by ITL and BATCo, our interests are confined to smokeless products and, so far, have been strictly limited to cigarette-like products (FAVOR). ~a view of BAT research interests and background in smoke generating products, including altering the nicotine/tar and flavour contents, it was stated at the outset of this meetlu~ by A.L. Heard that we did not wish to include on the agenda any smoke-related information. The purpose of the meeting was thu= ~o be confined to FAVOR and to the marketing/product development issues. mmlm~ O O~ <:D C li b PB'F-~- ~,'.G ~,',; ~:,,:~:f a's t'i'o-, c'o'm

- 2 - In view of the value of the detailed findings obtained in the UK study, ATP were asked to sign a secrecy agreement. They agreed (the signed copy is nov with Legal Department, Millbank). Finally, in order that, if necessary, information supplied to us could be shared vith B&W, an alternative form of a secrecy agreement, supplied by B&W, was signed by ATP and returned to Legal Department, Bag. Conclusions There appear to be several important issues governing our future involvement with ATP: 1. Do they have a concept which meets a real consumer demand? On the basis of all the consumer studies done by ATP or BAT, there appears little doubt that the consumer is seeking a product of this kind. 2. Does the ATP product meet this concept (or will it by reasonable developmental evolution)? We must concede that, although there may be a small segment of smokers who find the current product satisfactory, the majority of consumers appear reluctant to use FAVOR Regular. Whether this product can be modified to retain the consumer is a key issue. We are confident that the ATP development team, particularly if backed by BAT Group resources, could make considerable strides in improving 'smoking mechanics' and flavour. The real point at issue is whether ~he change in smoking style demanded of the consumer by FAVOR relative to conventional cigarettes will be acceptable in view of its lower per puff nicotine and very low overall mass delivered per puff. 3, Only extended practical experience will answer this question. Certainly, the more the basic mechanics are improved the greater the likelihood of the consumer to make the trade-off (which, of course, he may have to do many times each day if he uses FAVOR as a partial supplement). Is the ATP product protected by sufficiently robust patents that others (including ourselves) cannot easily copy the product? The patent position is not clear and needs very close investigation by expert counsel. 0 0 C IJ b PD'P-~- ~i:~t,?'. fiTsTi'oTc'o'm

-- 3 -- 4 • Alternatively, is ATP know-how such that they have a valuable lead-tlme for us? (We cannot answer that question in terms of our competlclou.) The know-how that ATP have is certainly worth 2 years lead time for BAT even if we gave such a project a high priority in R&D, itself an improbable situation if viewed realistically. We believe that subject to the continued employment of the key Personnel (especially J.P. Ray and I.D. Hill) the Company may be a most interesting proposition because of the pot~ntlal extension of the technology Into ordinary cigarettes, where use of the plug carrying nlcocine/flavourants in a filter can improve quality of conventional smoke. 5. ghat is the worth of ATP? On current performance in the test markets (using largely ATP indications) the company must be seen as high risk and worth far, far less than the $300 m projected by ATP. We are basically looking at an R&D company wlch a good concept, good know-how, useful production facilities but no proveu market - indeed, if anything, it has a disillusioned market. Other factors which may adversely affect the viability or profitability of Favor in the US market are the possibility of FDA regulation (even to the extent of prohibiting the product), of becoming subject to Federal Excise and Sta~e taxes as a tobacco product, and of the imposition of Advertising restrictions. Any such development would also then have implications for the product in other markets. If we wait until market share rises and approaches the 1% ATP expect by 1987, the buy-out price could be around $300 m. If, on the other hand, market share remains at 0.4% or, indeed, falls, or if the product returns predicted by B&W materialise, the company is likely to be in serious financial problems and could be purchased at a low price ($5 - I0 m?). &ecommendatious The sltuaCion argues for the following possible courses of action: Technical/Markettn# Strate$7 1. Check ATP patent situation. If not robust, we should consider withdrawing since others will 'move in'. 2. Check the toxicological information, taking expert advice on the safety of FAVOR. 3. ITL should proceed wlth their plan to run an extended study on 400-600 consumers using the FAVOR variant best suited to the Canadian market. <:D ~O C I i-b-P[9F'-~- t,? ~,,'J ~,,'Jtla's.t'i'o'r¢'o'm

- 4 - ~o . 6Q BAT should extend the A0-person monitor on potential FAVOR usage iniClaced in January 1986, Co test the ~Ime dependence on views elicited from thac pilot study. B&W should complete the 1000-smoker Texas-vide telephone study scheduled for completion end-March. This-should provide corroboration of the ATP study nov being analysed by Marketlng Department, Hillbank. Evaluation of current marker position in the field by BAT Marketlng management. Business Strategy 1. In the short term ITL and BAT should maintain dialogue rich ATP. In the light of knowledge gained from the above initiatives, BAT/ITL might consider in exchange for: (i) input of their own technology (i£) assistance An distribution from B&~ (ili) S one-off payment to ATP a) Exclusive rights to all smoke-related applications of ATP's current future technology and patents. b) An agreement that ATP ~ill restrict the expansion of FAVOR during the next 12 months to permit further Joint development of the product co eliminate or minlmise the negatives revealed so far by research. c) BAT/ITL would be given an option on the purchase of shares in the company producing FAVOR on a performance related formula: the option to. remain valid for 18 months. d) Any payments or know-how contributions from BAT vould be capicallsed if the option co purchase a share and participate in the equity of ATF is taken up. O G C7~ C l i I~P'D'F - vv vv vv. f ~--~t, o. c o Ill

- 5 - Karketln6 and Sales Situation No Harketlng presentation was made by the ATP Board to give us any real information about strategy, advertising, current sales and distribution arrangements etc. We were informed only that FAVOR vas launched in the last week of September 1985 in Austin, Texas (which represents O.3Z of the United States market) using a Broker and Wholesalers and that distribution yes extended to Hetro Dallas in November and rolled out to Houston in January. They intend to continue a roll-out to cover the whole of the US in ten steps completing step one with extension to the remainder of Texas, California, Arizona and Nevada in the near future. The objective stated by Ed Vtmond (vho was hilled as Director and Chairman of the Marketing Committee) was ~o achieve a major share of the US Cigarette Market. It was only subsequently that it emerged that this major share objective vas in fact 1Z. Hr. Vimond who vas President and CEO of RJR Xnternattonal from X980-1982 was not particularly impressive. Subsequent to the visit J.A.B. Kellagher read two New York stock exchange reports relating to ATP, one of which dated 30th August 1985 included In the 1let of Officers and Directors, Hr. Jim Rarkness under the title of VP Harketing. There is no mention of him however in another report dated 9th September 1985 and he was neither present nor mentioned at our meeting, wo it may be assumed that he has left the Company. ATP has recently completed a Favor Attitude and Usage study based on a telephone interview survey carried out in the Austin initial market after the first thirteen weeks. No management analysis of these results was presented except for the information that the product had achieved 91Z awareness amongst smokers, that 55X had tried Favor, 33Z had actually bought Favor themselves and 4Z are currently using the product. The share of market calculated by them on the basis of the research Information results in a figure of 0.9Z of the Austin cigarette market for the whole 13 week period and 0.4Z current share. Share figures, incidentally, are all calculated on a per pack, not per stick, basis. The research revealed that current users are 65Z female and tend to be upper income, white and aged 35+. The full Survey report has, however, been broughc back to Hlllbank and a more detailed analysis will be prepared. ~r. Vtmond claimed chat ATP had gained a good breadth of distribution and that only two firms had turned down the product buC admitted that ATP eere lacking in depth o~ distribution due to the absence of any field-force. This, of course, is one of the reasons that they are anxious to form an association with a large F~CC company and preferably a tobacco company. C li t~'P-D-F - ~la-"~i~i

- 6 - ~r. Vimond was informed that the information we had received from B&R suggested that although substantial quantities of Favor had been dls~ributed to the Trade much of this stock would, in due course, be returned. Hr. Vlmond denied that any stock had been returned. It is worth ootlng, however, that B&W obtained this information from their Broker in Austin which is the same ss that used by ATP. Hr. Vimond predicted that market share would now steadily increase, claiming in effect that 0.&X share is the bottom of the curve. When questioned ou this subject, however, there appeared to be no'statistical or other objective Justification for this expectation. Production process A visit was arranged Co see the production facilities and laboratories. The factory is clean, bright and tidy as are the adjoining quality control/research facilities. FAVOR comprises an extruded plastic (85 mm) tube containing a small mioroporous plug (9 mm) vlth a brovu tipping paper at the mouth end ~remoce from the plug). The microporous plug contains adsorbed high purity free nicotine and selected volatile flavourants. Impure nicotine (extracted from tobacco) is imported as a 95Z (impure) extract from India and is distilled (in a plant located outside of city limits for t6xicological reasons) to better than 99.9X purity. This is diluted eith ethanol and flavourants added. A batch of plugs checked for pressure drop is floated ou this solution and by applying a vacuum, the air is removed enabling the nicotine solution to penetrate. Solvent is removed by rotary vacuum drying leaving pure nicotine/flavourants adsorbed on the pore surfaces. According to Dr. Z. Hill, the dimensions of the plug/pores places an upper safe limit on nicotine loading, a feature ehich will need careful probing if collaboration is pursued. The plugs are transported in sealed churns to the assembly factory situated at the rear of the ATF office premises. This occupies 25,000 (?) sq. ft. Currently, ATP have two making lines from blank cubes to wrapped cartons, ac present laid out in the typical U-conformation of a cigarette making assembly. (This rill be replaced by an in-line conformation when these two machines are transferred co the adjoining building where tvo other assembly lines are being built.) Blank tubes of double FAVOR length are fed by hopper Co a grooved block which transports them approximately 1.5"metres by which time plugs are rammed into either end of each Cube. Tn the subsequent operation, tipping is applied co the centre of this dual rod and the rod is cut. The product (6 to a pack) is sealed ou a heat seal chocolate wrapping type machine and packed in a hinge-lid box. G ClibPDF - www.fastio.com

- 7 - Production, which is currently packer speed llmlced, is between 80 and 130 packs/minute. ATP claim chat on four shift operation the four production lines soon co be available will be capable of supplying the 1Z market share they foresee for FAVOR within the next year. Patent Position ATP claim that their issued US patent (No. 4,284,089) covers securely the basic principle of nicotine volatisacion implicit in YAVOR. They also cited the EEC patent applications (already known to us) and stated that other patents on details of construction etc. are in progress. A.L. Heard expressed his concern for the robustness of the US patent following a first survey by BAT Patents Department. He expressed an intention, fully agreed by ATP, to explore in depth the soundness of the ATP patent position since this is a keystone of any collaboration. Toxicolo~ical DaCe relatin~ co FAVOR ATP provided both BAT and TTL with a copy of a 30-page document eutitled "Toxicology of Nicotine Base - an assessment for ATP" by Dr. Daniel M. Fords a consultant toxicologist. This covers the acute and chronic toxicity of nicotine and goes on Co consider implications for FAVOK. Abuse, risk to children and workers are considered. The conclusions suggest no real concern but the document has now been passed to T.G. Mitchell of BATUKZ KaD for expert comment. ATP were made aware of the kind of requirements that the BAT Additives Cuidance Committee would seek in approving use of ~AVOR or similar items. It was made clear that BAT would, if it pursued further with ATP, ask for all confidential details of ingredients, materials etc. to be disclosed to a nominated person. UK Consumer Stud~ on YAVOK The results of a small-scale consumer study carried out in UK using Regular FAVOK were shared with ATP (under a Secrecy Agreement). The'trial was based on 40 subjects selected by delivery, sex and concern (expressed as ability to cope for • periods without cigarettes). The study was divided into three stages: Cr~• 0 C I J bpIDF - ~J. f ~TWti 0-. c o ill

- 8 - 1. introduction co concept and produ:c ~a a ~-hour period of clKarecce deprivation buc free access co FAVOK a &-day period of free choice - conventional clsareCCes or FAVOR ac will The findings are summarlsed in the attached uoce by R.P. ¥erris. ATP vere excreme&y interested in our findings and were clearly i~pressed by the skill shown in both constructing and implementing such an informative study ac short notice. They were quick co see the benefits that access co the BAT Croup techniques and knowledge of cigarette design and human smoking/preference could brine Co them in the developmenC of FAV0~. Development Potential of FAVOR Two features distinguish FAVOR from ocher nicotine-delivering alternatives co the conventional cigarette (e.g. sums, tablets, dry and wet snuffs): a) the delivery device is physically and visually very similar to conventional cigarettes b) the mode of delivery, albeit vapour and noc aerosol based, relies on inhalation in a personally controlled series of puffs. These are extremely important features in chat Chef offer familiarity to conventional cisaretCe smokers. However, FAVOR has important known constraints. Firstly, nicotine delivery per puff has been maxinised at 10 us, only 10Z of the typical nicotine level/puff from a eouvenclonal mld-car cigarette. The potential of one stick to yield up to tO0 such puffs leads co a behaviouraI chanse in the direction of more rapid puffins by the user. Hore importantly, the Coral matter delivered per puff of FAVOR is merely ~Z of Chat from a conventional mid-tar cigarette. This will call for considerable ingenuity in product modification co restore traditional sensory cues such as mouthful and reward for draw effort. Dr. Hill will seek to achieve such improvements by selection of approved volatile flavourancs. o C~ Cr~ U~ ~mmm~ Clit~PDF - www.fastio.com

- 9 - Potentlal gAY Asslst~nce in FAVOE Development There are three ways in which ATP development of FAVOR could be accelerated if BAT resources were made available: C1) Sensor7 Development ATP currently use ad ho¢ in-house evaluations for sensory assessment of the product. Lack of a panel approach using a set of standard attributes means that ATP have not crated their progress with a development history of comparative sensory profiling. BAT research has established a basic sensory pro£iling technique which can be used to indicate sensory shortcomings and provide guidance on levels o£ sensory magnitude which would improve product acceptability. Notably TAVOE ~CULAR shows problems on taste level (too low) irritation (coo high) and mechanics (poor perceived draw due to the fact that ATP wrongly define draw ef£ort in terms of measurable draw resistance instead of against the consumer criterion of reward for draw effort). BAT sensory pro£iling of ATP's more recent ex~erimental prototypes shows chat these problems are surmountable ~o some large extent. Use of a sensory testing approach would significantly £ocus and speed up the current empirical approach and should concentrate on determininS the interdependence og mechanics and taste masuicude perceptions (the two seem to interact such chat increasing taste magnitude will produce improved perceived mechanics even against a fixed measurable draw resistance baseline). The method should also allow an understanding of the relative contribution or salience of particular sensory attributes co overall product acceptability. Ic will be important to compare such a "contribution heirachy" for FAVO~ with that more typical of conventional cigarettes. Looking at the same problem from the consumer perspective, it will be important to understand relative sensory requirement hairachies of low and full £1avour smokers if KEGULAK and LIGHTS versions of FAVOR are Co be successfully di£ferentially marketed to such consumers (at present, data on FAVOR REGULAR shows no differentiation of sensory response by delivery level of smoker, buc a clear dif£erenclacion by acceptability). Cli) Behavioural-Heasuremeut Diary monitoring of FAVOK use patterns indicates that the product is used in a very dif£erenC manner to conventional cigarettes (often in use continually for periods of over one hour) and chat its use alongside cigarettes is associated with a reduction in cigarette consumption. Sensory testing also indicates that 'getting the mechanics right' (i.e. achieving an appropriate effort/reward relationship) is key to product acceptability. A central issue concerning C7~ C li tiP D-F - ~1 a-~'tT6T'_~oWfi

Clii) - 10 - eventual success/failure of FAVOR will be whether users can learn to trade getting nicoClne in large doses from a small number of puffs for the FAVOr alternatlve (small per puff dose, large number of puffs). The observation that cigarette consumption declines when FAVOR is used supports the view that FAVOr Is satisfylng a pharmacological demand. Therefore the prime use of behavloural measurement would be to ensure that the sensory cues driving smoking behavlour are optlmised. Smoker effort can be measured dlrectly as puffing pressures aud'volumes using smoking behaviour equipment. It should be technically feasible to convert current equipment to monitor FAVOR smokers and derive the following important information: a) A concrete measure of smoker effort being applied to an7 FAVOR variant, which can be used as a criterion for arrlving at satisfactory 'mechanics', and may also be correlated to sensory perceptions of draw effort. b) Characcerisation of physical smoking style learned in using FAVOR relative to conventional cigarettes. c) Examination of how critical the role of satisfactory mechanics will be in FAVOr acceptance. Will the most acceptable FAVOr simply be the variant with the best effort-reward relatlonshlp (i.e. closest to conventional axpectatloos) or will the quallt~ of reward £n taste and irritation terms support trade offs against the effort criterion? Product *Cosmetics' UK qualitative work has defined a number of product appearance factors which aeem to contribute significantly to product acceptance/rejection. Most notable of these are tactile feel of the rod, requirements for more perceived solidity at the mouth end (perhaps by use of a baffle), stick use/depletion indication, improved (preferably single stick) inner wrap, improved pack design (in terms of function and appearance). While ATP are larsely aware of these aspects, the use of a panel of *Innovation sensitive' consumers, such as are available in the UK for. qualltatlve evaluation would he a considerable asset. .....a. O r~D O', O ClibPDF - www.fastio.com

- 11- ATP Business Plan ATP have noc flnallsed a business plan. Nevertheless, Dr. Jacobsen van prepared co oucllne the way he saw FAVOR earnings developing and the options for involvement vlch potential partners. Earnings: 1986 - profitable 1987 - $25 m net earnings pre-tax 1988 - $50 m " " " 1989 - $100 m " " " This is based on 25Z profits on sales assuming no 'tobacco' tax is imposed. However, ATP were keen to stress that the current product is certain ~o reduce in cost as new ideas and process improvements are made. The high gross margin (>70Z) for this product (which would be still >60Z if tax were imposed) means that providing the market share reaches a level of general availability, profits should be very high. ATP believe FAVOR could be a 'white hat' product (implying that it has an attraction Co medical groups, government and even anti-smoking factions,) and could be strongly supported. Zn the patent area ATP believe they are strong but recognise that their tr~e technical asset is in 'know-how' which they estimate equates to a 2-year lead rime. (We do not disagree.) In view of the positive outlook, ATP see for their product (once optlmised), they estimate company worth at $300 m based on the 2-year post tax profit ($15 m) combined with a 20 multiple. In terms of participation by other companies, Dr. Jacobsen saw three possibilities: i. Buy out 2. ~uy in at a percentage 3. Licensing arrangements or Joint Ventures. He seemed to favour option I. with the second option being 2. but 3. was not particularly favourably viewed. However, a combination of 2. and 3 was not discouuted. One possibility that ATP foresee for spreading activities internationally with a partner is to operate a 'Coke syrup' philosophy, l.e a key element of the process could be manufactured centrally in San Antonio and shipped to YAVO~ manufacturing locations (it is easy Co envisage the ntcotine/flavour plugs being so handled). Although not made clear, there was a suggestion Chat ATP are encountering distribution problems in USA and would welcome collaboration with a domestic US cigarette company. ALH/DET 7th March 1956 Distribution: Messrs. E.A.A. Bruell J-L Mercier C. Warren J.A.B. Kellagher R.P. Ferr£s O I J b P~E~.F---v.,z.vv.wd@s+i.o-:c o1~:1

~ote to J.A.~. Kellagher, Esq. FAVO~ Research P~porc - United SCares You asked for coffiments on the data relating Co the post-launch research study " of FAVOR smokeless cigarettes in a rest-market area of the United States. 519 smokers were questioned by telephone, a sample which yielded 20 current users of FAVOR and in order co boost this nuaber a further 30 users were contacted. Before ccmnenCinE on the findings there are one or two points which I should make about the val/dity of the data. 11 Our ability to say very much ~rlCh confidence about the outcome of any research study must always depend on its scale and where this particular type of project is concerned IC is only rarely thac we find thac sufficiently large samples have been used to permit a reliable comparison of Important sub-groups - for example buyers versus non-buyers of the p~oducc in terms of their comparative make-up by age and regular brand. Thus, such comments as ~ ,m able to make on differing patterns of behaviour or acclcude as-between the breakdowns by which the data is analysed mUSts for the most part, be tentative rather than firm: there are many tendencies and indications apparent in the data but few conclusions to which sign/ficance in terms of the conventionally acceptable level of statistical reliability can be attached. ~e Current users of a test-product are inevitably of particular interest in such a study but severely constrained by total sample-size. In this particular case the initial study yielded only 20 current users of FAVOR and even when boosted they totalled only 50. Such samples permit us to say very little with any confidence and in this instance, the situation is worsened by the possibility that the boosre~-sample has distorted the picture: the additional users were found by non-random means and comparison with the ma/n s"mple data suggests that they are excessively skewed towards women and older smokers. Without a 1-arKer user-sample, found by more representative methods (e.g. by inserting questions on an Omnibus or sim~larly large study) it is impossible to say vhat is the true age, brand and sex profile of those who are still using the product. Finally, in order to maximise the information on which analysis was based these additional users were added into the sample wherever possible and althouEh the effect is more diluted, this Coo may have had some adverse effect on such data as the total buyers' normal cigarette-type. 3, It is difficult to assess the representativeness of the basic smoker-sample: the profile by segment under-represents smokers of regular cigarettes and over-represents lights smokers when set against Maxwell nation-wide sales figures, the only data available to me. However, this could he due to regional differences. CD ~4D G t.n °. C li b P D F - vvw-w~l a.st.i,(~c~(~.r-l-1 ....

~o -2- Awareness Smoker-awareness of the new product is remarkably high, at 91%. It is higher among men (93Z) than women (88Z); it rends Co be slightly higher among the youngest age-group (the 18-24's) and lower among the over 50'8. And smokers whose daily-consumption of conventional clgarecces is classed as 'moderate' appear also to be rather more aware of the product than ocher smokers. Trial Over half of all smokers quesCloued had Cried the product, and men to a greater extent than women. The pattern of trial by age and by normal clgareCte usage is similar to that for awareness. Purchase One in three of all smokers had got as far as buying the produnC. Men and women appear Co have done 8o Co an equal extent, as did all age-groups bu.t the over-50's, of whom only one in four made a purchase. 'Moderate' consumption smokers were signlflcancly more likely to buy than ochers. Of Chose who had ever bought the product, however, only 12Z (4Z of all smokers) were still using £t aC the C£me of the research and-although they were drawn equally from the sexes and the three cigarette-cons,-.pt$on groups, the age-groups show a more marked fall-off in buying among the under-35's than older smokers, There are indications (but no more than this) Chat users may be " sllghcly'skewed towards the upper educaclonal groups and towards ~.ce~ rather than Blacks. The brand-share profile of chose smokers who were aware of FAVOR, Cried lc or purchased it approximates to that of the smokar-semple as a whole but Marlboro smokers appear to he slightly under-cepreeented when it comes to pnrchasln8 the product. Source of purchase There were no significant differences in the place of purchase as between the first occasion ou which FAVOR was bought and the most recent occasion. Nearly half of all buying Cook place in a Supermarket (46Z), and the next most important outlet was the Convenience S~ore (37Z). Drugstores accounted for 7Z of purchases, and r~scounc Stores 3X. The incidence of buying in Supermarkets and Drug Stores was markedly higher among women and older smokers, while men and the under-35's were more likely to make their purchases in a Convenience Store. ~xD C3~ Crx C li b P D F - w.w.w-l.a.s¢i,®-.e®,m

-3- The pattern of source of purchase as between first-films buyinE and the most recent occasion vas the same. ~uanetcy purchased Whereas normal cigarette purchases vere predominantly (67Z) made in cartons, purchases of FAVOR vere almost invariably (96Z) in indLvidual packs: 84Z of purchases ou the last buying occasion cook the form of a single pack, and a further 10Z, two packs. Female and older buyers, however, tended to show a higher incidence of two-pack buying. Froduct-type purchased The majority (84Z) of smokers of "regular" cigarettes bought the regular version of FAVOR, and 75~ of menthol cigarette smokers boushC the menthol version. Those who normally smoked lights, however, were much less likely co buy the lights version: only 37Z did so, the majority (58Z) buying regular FAVOR. The nec result of this was ehac over half of all buyers chose the regular version, a quarter Che menthol, and the'remainder the lights. Former users 88Z of those smokers who had st esme t4me bought FAVOR had discontinued using it. Of these nearly half had made use of the product for only one day, and tthree-quarters for less than one week. Regular smokers of a menthol brand and those ~ao boushC the menthol verslou of the FAVOR product appear to have abandoned FAVOR =ore quickly Chart other smokers: 87Z of buyers who normally smoke menthol dropped the use of FAVOR ~chin a week of trial. Nevertheless, a small core of smokers (around 10-15Z of FAVOR purchasers), ![drawn to a similar extent ~rom the regular, lights and menthol segments of the market, claim Co have persisted with FAVOR for as lone as three months or more. ~easons for quiCCtn8 42X of Chose who vere no longer us£nS FAVOR crtcicised its taste and there vat little difference in the level of Chic criticism as between buyers of the regular, iishCs or menthol versions. Those who normally smoked lights cigarettes, hovever, shoved a Cendeucy Co be less critical In this respect. The second most frequently mentioned criticism (34Z) was of a lack of satisfaction, a complaint which appeared to be more frequent among those who had bought the menthol version of FAVOR. 29Z of former users cricictsed the product for excessive irritation, a ~omplainc which tended co be higher among women, older smokers and those who normally smoked lights cigarettes. C li b PDF - ~l~R~'t'l o.c o m 0 ',...rl ",,I

.j m -4- Asked the likelihood of their buying the product at some c~ne in the future, those former users who had last bought the lights version of FAVOR and who were in the 25-34 age group Cook a rather more positive attitude than other respondents. • Current users As my opening remarks indicated, the small size of the coral user sample (51) and uncertainty as to £Cs reliabilicy restrict firm to,went on this particular sub-sample, and especially the user-profile. However, if we ass~e Chat it fairly reflects current buyers of FAVOR, the following points may be noted:- Ii 14% of current buyers had made use of the product on the day they were contacted, and about 50% between one and seven days ago. Only 14% of the sample used more than one FAVOR cigarette per day. 2o 8% of current users claimed to have bought a carton at their last purchase buc nearly three-quarters had bought only one pack. 6 1~ used FAVOE only in situations in vh/ch they were free to smoke conventional cigarettes; 40% used FAVOR where conventional smoking was forbidden, and the remainder in both smoking end non-smoking situations. 4. It is particularly interesting Co note ChaC a high proportion of current FAVOR users said that they were using the product in order to help cham cut dovu their e~ok.'Lng of regular cigarettes (25%) or as an aid Co quirting sacking all together (22%). The next most frequently mentioned reason was as an alternative co regular cLgaretCes in a non-smoking situation (35%); and, third most often mentioned, as a means of avoiding offending others (29%). t" ~%." of current users believed ic co be safer than regular ~igarectes. 5. Asked ~hat they liked about the product ~ 44% of current users praised ice Satisfying qualicles. Posiclve c~eut on Caste aspects, however, li were considerably lower, at 16%. The majority of favourable responses related to the absence of "smoke (58%), and 22% singled out the fact Chat FAVOR was noC offensive Co ochers. No ocher favourable comments were made by more than 6% of users • 0 G~ ClibPDF - www.fastio.conl --

-5- 7e Users' volunteered dislikes were ma£nly centred on the pc.duct's taste characteristics (35Z) and a lack of satisfaction (31Z). OoJ.y 14% missed the smoke from convenct0nal cigarettes. Crct7 one-third of current users agreed chac Chef would definitely buy the pc.duct in future, although a further 45Z said thac they would probably do so. Main conclusions lm Awareness, trial and initial purchase In the rest-marker area were undoubceck~y high. B Hovever, early rejection of the produce after initial trial was also high, largely for Caste and irritation reasons and a lack of satisfaction - criticisms made even b7 current users. The menthol version appears Co be less successful in produce terms than the lights and regutar versions. . The absence of smoke does not appear to be a major obstacle to acceptance. Information on the reasons for purchase is confined to tie small s=mple of current users. Nevertheless this indicates a strikingly high claimed level of the use of FAVOR as an aid to cutting down or quitting smoking, as well as a solur~ton co the problem of nou-mok£ns situations. . Those smokers clain~nS Co be us~ FAVOE at the time of the research totalled'some 4Z of all smokers; however, less than one in seven current users were found to have used the produce on the day of interview, and oul~ one in three said that they would deftnicely continue to buy the produce. Thus, although the concept has plainly g~nerated a high degree of interest and trial the attrition-race is also high and it would be interesting co re-cent-act current users ac a later date co find ouc at whac level purchase finally settles. On its present showing the product, as currently marketed, does not appear to represeuc a serious threat Co cigarette-sales. 7thPLarch 1986 C.J. Pendrr cJP/~0c o LIl C IJ b P DF~--~,7~,'~W-f astioTc-orfi ....

ADVANCED TOBACCO PRODUCTS, INC. February 19, Z986 lCeeCing: 9:00 AM Par~£cipan~s Alan Heard Cliff Waz-ceu Jim Yniacke Robb ¥erris - BAT - Imperial Tobacco Ltd. - Imperial Tobacco Ltd. -BAT Alan F,.ellagher - BAT ATPI Norman L. Jacobson, Chairman and Ch'(e£ Ex.ecueive Officer ~./e-mmzl~a~ J. Philip Ray, President e"~'m~,~Q,/~ee,KC/4,,,Cmb,o~ Edmond C. Vimond, D£.reccor, Chairman o£ MarkeCCnK Committee ~ ~ /~t Ira Hill, Vice Pzes£deut, Technolosy - t~/¢,e, mv-¢~.,~e,,.m,m,~ James D. S~noneeu, Vice President, Administration ~ ~,~/~ zese~mes/waNh~r 0 r'O 0", 0 0", 0 C l i bPDF - ~ cq--Stl o. c o ii1 " "

FAVOR U.K. RESEARCH FOLLOW UP: FURTHER CONSIDEP~TIOIIS Previous U.K. research on the FAVOR product has indicated that, in the short term, response to the concept Is hlghly favour~ble whllst response to product ts somewhat less favourable. Price elasticity of demand for FAVOR REGULAR is htghly Inelastic, but becomes elastic tn relation to other product variants sampled. It now becomes necessary to examine longer run trends fn attitudes and usage of FAVnR. Three al ternatlve research strategies are posslble, all of which assume that a pre-product screening exerctse will have been performed on ATP's range of product variants in order to determine the optimal product for placement: Ii EXTEHD THE CURRERT POSITIOiI i.e. 6 week placement wlth existing 40 respondents using dlary monltoring, weekly attltudes/usage questionnaire, butt collectlon and weekly product placement. STRENGTH: low cost. WEAKtIESS: professlonalt sed and biassed panel. e COST: c. £6,000 VALIDATE nH A DIFFERS)IT PAHEL i.e. Inltlal concept aualitatlve sessions followed by 6 week placeuent wtth new sample of naive consumers {c. ~0) again using diary monttortnq, weekly attitudes/usage questionnaire, butt collection and wc~ekly product ptacenent. STRE)fGTH: validation of earlier work, avoidance of 'loyal responding'. WEAKt!ESS: cost of pre-recruitment. 1 COST: c. £9,000 SITI'JLATED TEST tIARKET MOnFLLItlG: i.e. to use a new sample and apply a classic ASSESSOR evaluation to promote trial, re-trtal and subsequent modell I ng of long run market share. STRENGTH: WEAKNESS: will provide a prediction of product performance in the marketplace in ter~s of share. within reasonable costtng,saemle stze would be small (c. 60) therefore accuracy of prediction would have high error variance. Loss of richness in usage/attitudinal information would occur. COST: C. £11,000 C~ ",C) C~ (~ CP~ C li b P D F °--'VC/'C'~W':fra's't'i'oTc-o~ll

-2- Choice of option is heavily dependent on whether the objective ts to learn more in depth about attitudes and usage trends with time, or to atilt predictions of market share potential on the basis of some reallstlc purchase opportunity situation. The former objective is addressed by options I and 2, of which option 2 is the most desirable since it combines validation with a move to a less biassed sample base. Ootlon 3, while attractive in commercial decision making terms, really requires e much larger sample slze to be done well and Is, in any case, somewhat pre- empted by Al~ providing, In effect, a ready made test market (assuming we can monitor it satisfactorily). PRICING EFFECTS Earlier discussions lead to the view that product placement would be by purchase of both FAVOR and own brand conventional cigarettes at an equated, lower than Market, rate. Whlle provldlng a welcon~ element of 'hard reality' Into the testing sltuatlon, the Introductlon of a purchaslnq varlable carrles several problems: I. Sample size is already small and would reduce to unrealistic levels, even by the most optimistic prediction, If a purchasing criterion were introduced. 1 Consumers are very uncomfortable buylnq supplies of clgarettes on a weekly basis. . To achieve an equated price slt.atlon means either conventional cigarettes will be priced unreasonably low, or FAVOR may be orlced too hloh (relative to perceived costing data from the earlier study). In either case. 'normal' consumption patterns will be significantly distorted. 4t Relaxation of the requirement to equate on price begs the nuestton "what should be the price for FAVOR?'. This question is complex and outside of the scooe of a small scale exploratory study since price elasticity of response at various pricing pitches would have to be modelled. RECORV£NOATION Since the scale of thls U.K. research wlll not support a credible evaluatlon of long run market potential of FAVOR, and since this objective is currently better addressed by researching the llve market and by the proposed Canadian study, tt ts recommended to droo the purchase variable from U.K. research stnce at best tt wtll reduce available data and contHbute significant bias. Concentration on the objective of tn-depth diagnostic research on usage and attitudes would be most suftably achteved through Option 2 (with perhaps a low cost parallel placement to the orlgtnal sample). R.P.FERRI$ RPFIBCH 13.3.86 0 P~ C~ P~ C l i b P-D.F - ~ v.w.w~.l.a-s.t.i,o-, e®.m

Mr. R.P. Ferrls/JJ 27th February, 1986 DEVELOPMENT POTENTIAL OF THE FAVOR PRODUCT Two features destlnsuish FAVOR from other nicotine delivering alternatives to the conventional cigarette (e.g. gums, tablets, dry and wet snuffs): a) the delivery device Is physically and visually very similar to conventional cigarettes. b) the mode of delivery, albeit-vapour and not aerosol based, relies on inhalation in a personally controlled series of puffs. The presence of these features is highly significant, since the former all.as the user to re-assemble the familiar gestural routines associated with smoking around the new behav£our of raping, and the latter ensures the high efficiency pulsed dose lung uptake which is central to the development of a behavlourally reinforced product use pattern. In developin8 the potential of the FAVOR product, one known constraint Is thac the nicotine delivery potential per puff hss been optlmlsed to a ceiling of c I0 micrograms per puff, therefore future development wlll be about optimlslng stimulus values of the sensory and physical context which the product places around a nicotine element which is essentially providing lower per puff nicotine levels across an extended puff number potential (90 - 100 puffs per stick). It is significant that FAVOR development thus far has been considerably faster than new product development cycles typical of corporate R&D. This is due co the small scale entrepreneurial nature at ATP linked with unique high level capability in nicotine delivery development and flavour formulation. A corporate R&D approach to support of ATF should be avoided in favour of a fast, praKumCi¢ and product oriented input, ideally 'hands on' alongside Ira Hill on whom most tertiary development of the basic product wlll now begin to rely. BAT could prow/de product development support at 3 levels: Sensory Development ATP currently use ad hoc In-house evaluations for sensory assessment of the product. Lack of a panel approach uslnE a set of standard attributes means that ATT have not traced their progress wlth a development h/story of comparative sensory profiling. RAT research has established a basic sensory profiling technique which can be used to indicate sensory shortcomings and provide guidance on levels of sensory maEuitudewhiohwould contribute to improvements of product acceptability. Notably FAVOR REGULAR shows problems on taste level (too low) irritation (too hl8h) and mechanics (poor perceived draw due to the fact that ATP wrongly define draw effort in terms of measurable draw resistance instead of against the consumer criterion of draw effort in relation to achievement of a given sensoric reward). P~D CD C li b P D F - vvw_w,.ta,sI, i.o,.,c- o.ra.1.

-2- BAT sensory profiling of &TP's more recent experimental prototypes shows that these problems are surmountable to a large extent. Use of a sensory testing approach vould significantly focus and speed up the current empirical approach and should concentrate on determining the interdependence of mechanics and taste magnitude perceptions (the ewe seem to interact such that increasing Caste magnitude will produce improved perceived mechanics even against a flxed measurable draw resistance baseline). The method should also allow an understanding of the relative contribution or salience of particular sensory attributes to overall produce acceptability. It will be important to compare such a "contribution he~rarchy" for FAVORer lib that more typical of conventional cigarettes. Looking at the sme problem from the consumer perspective, it ~ll be important to understand relative sensory requirement heirarchies of low and full flavour smokers if ~ and LICRTS versions of FAVOR are Co be succesfully differentially marketed to such consumers (at present, data on FAVOR ~CULAR shows no differentiation of sensory response by delivery level of smoker, but a clear differentiation by acceptaSility). A highly time/cost effective model of consumer preference testing will also be key to screening variants follo~rfng the rather more diagnostic role that sensory panels may provide. BAT's PR~FAN model using small samples wtth statistics which preserve independence of respondent scoring behaviour would be particularly appropriate. Behavioural Measurement Diary monitoring of FAVOR use patterns indicates thac the product is used in a very different manner to conventional cigarettes (often tn use continually for periods of over one hour) and that Its use alongside cigarettes is associated with • reduction in cigarette consumption. Sensory testing also indicates that 'getting the mechanics right' (i.e. achieving an appropriate effort/reward relationship) is key to product acceptability. A central issue concerning eventual success/failure of FAVOR~r~ll be whether users can learn to trade getting nicotine in large doses from a small numSer of puffs for the FAVOR alternative (small per puff does, large number of puffs). The observation that cigarette consumption declines when FAVOR is used supports the view that FAVOR i seattefying a pharmacological demand therefore the prime use of behavtoural measurement would be to ensure that the sensortc cues driving smoking behaviour (notably chose concerned ~th perceived effort) are 6pttmised. Smoker effort can be measured directly as puffing pressures and volumes using smoking behaviour equipment. St should be technically feasible to convert current equipment to monitor FAVOR smokers and derive the follo~ng important information: i) A concrete measure of smoker effort being applied to any FAVOR variant, which can be used as a ~riterion for arriving at satisfactory 'mechanics', and may also be correlated to sensory perceptions of drav effort. ii) CharactertsaClon of physical smoking style learned in using FAVOR relative to conventional cigarettes, and evaluation of smoker effort modelled by delivery level of smoker. mmmab G Cr~ ClidP-D-F - v.~/~/,,/\,vi,'a{[io.col]l --

-3- ill) Exan~natlon of hc~crltlcnl the role of satisfactory mechanlcs will be In FAVOR acceptance. Will the most acceptable FAVOR simply be the variant ~rlth the best effort-reward relationship (i.e. closest to conventional expections) or will the quali~of reward in taste and irritation terms support trade offe against the effort cr~terion? Product 'Cosmetics' U.K. qualitati~e work has defined a number of product appearance factors which seem to contribute significantly to product acceptance/rejection. Most notable of these are tactile feel of th~rod, requirement for more perceived solidity at the mouth end (perhaps b7 use of a baffle), stick use/depletion indication, improved (preferrably single stick) inner wrap, improved pack design (In terms'of function and appearance). While ATP are Largely aware of these aspects, the use of a panel of 'innovation sensitivet consumers, such as are available in the UK for qualitative evaluation is strongly indicated. Conclusion Sensible support for ATr should be designed to run in step ~rlth the current pragmatic empirical approach, avoiding the temptation to 'go fundamental' and aiming to ~atch up FAVOR performance against what we know of the 8eneory/behavioural performance of conventional cigarettes. R.P. FERRIS O C7~ Cr~ ClibPDF - www.fastio.com

This to to confirm that my personalpar:tclpatlon in ~mrter Research actl~rltles is on an entirely voluntary basis and of ~, ovn free rill under the following terms: le 6 ~e That my relatlonship to Charter Research representatives be governed b7 ethical ~uldellnes currently isld down by the Rrltlsb Psychological Soctet7 pertaining to the pro£esslo~al conduct of PsTchologlsts- That m7 relatlonshlp to Charter Research representatives be governed by the Xarket Research SocletT code of conduct. That, in the case of any exerclsee l=lvin~ produce evaluation, products e~sloated should be of marketable standard. I have read, understand and accept the above conditions regarding "wy voluntary relationship to Charter Research Associates. Signed • • • ....... • Dated . • • • • ... • • • • .....=b C~ ClibPDF - www.fastio.com

---