.I~ P~~I~IOIS i, Against a clssr liina~ in the iiclb fn rlspsct or OPtrD1 el the point of ssIs lnvironmsnt, svaluation of osrohsndising ostsrial in visual prominooos a 'atlnd out' tara is underdovslopsd within Villa. ms in-hoUss rsacoeh rtlds us to sons bltllt IdC 1Wd in ths svaluation 61 av pet Qcsfa and msrchandising dssign projscto, 2. lo Ipai~oll4 rcs6rehld guibClincs a~af to direct b ~a61 lacs as positioning stratsgy with P,O,S. Cotrils, 3. Pc guidslinsa 6dst on d~rlrsatiP preh~dog polict bJ stJlsllooaUonlatte o~ eltsr·l sdvsrtisingl I, ~lonltion ~~ the speibe inpsct ar pack/brand iaslerY Ca ~esked asnaorf propcrtiaa or dpa~ttas is gc·FalJ undsrdsvslopsd. ICrIOI POfNTS 1, It is sdvissd that l~mlotio·l taohistoscops scrssddg stags ha ~~nsd as a routios inclusion in ths usual sspuancs d aw paok de~depcPt, a modilicltfcn to cdatipg asckn, Bid step volld provide idoatioP or ctiloi,Pg of dssfp Illnd~Pt lad Ihodd ccDpleslCnt i~~mation c~ deli8n'loathstlcs, SllilarlY, projectio~ ~o~ltolocpe ~Pd eys las ·qifodpg CquipnCnt m tl IsCd to svaluats Ilriclencsa or Plrch~ndiainl ~ltcr~~, Idditioalll, it Ip bl of valus to rsvinr ths potsottd or ooaputsr bassd Ilotlarch·~aing dtaip laras oa adlabla log such allta hr bssn dadapd at dl i OC), 21 It Is rsconnsndsd that cbnsidsration be givsn to providing guidslinsJ b salca Ircl personnel on product CosItioalag strategy in CPtrisl, tbs ge gzs ~adtainl systca an be used to dsvllop ~Qlfaal~ ~e crr6nt OP i De slstsa is portable cad should ideally bs ussd litMI tbs oulturs ia whish the rllDlts Ira to bs applfsd. L ~I· prllol91n ald~ ~W~~a.l~ 1··,, gyl~ ddocr ogtrtiaiPg ia attrlolipt sys gscc sad sPb~clrt raalU r(CopdtiOP CC ahr svsilablc, I r(POrl on the efe III IOIitOrill study which iJalltod those principles vill shortly be ndlohls to Pills, It i, sdvised that Itch pidllial IB k usslul in d ianPadPI outdoor Igtrfldpg lita pac~~l policy, ClbPDF-~I~, ~~w~~n · --- I Zr · I, it is recommended that measurement of the impact or brand image on basic product senso~rT perceptions be evaluated using the brand image measuremer;t model which can be applied in·hou~ using DaTA, or preferably with consumers using HILE in replicated blind and branded trials, This type of measurement gives an indication or the short term impact of brand imagery operating at paints of brand choice and switching, In the WILE model, some se~Pentatiaa of the respondents is possible by relative imagery responsfvenea. - · O IV· CC Clil;PDF - !::!r!r:!:f3stlc.;:nlil SPECIFIC ~eBOJE~:T~.DrSCUSSInKS O h' Cc O Clil;PDF - !::!!::a::!.f3stlc.;:nlil i, Discussions held with Snack Food Division were necessarily limited, however, it is plainly apparent on brief exposure that the division 13 strikingly under in9ested In ronsumerrPsearch · to the point of vulnerability in the case of in·house panel testing and canJ~mer product testing, considerable loss of opportunity must also be represented by the lack or up to date imagery research, Priority considerations should be given to equipping the division with a satisfactory sensory testing and consumer product testing support, perhaps au~ented by a new image surtey and point or sale prominence research, in that order, The lack of recognition or the usefulness of the diagnostic (trial) element of the assessor technique is particularly surprising, The value or this approach for screening new product concepts made snack foods one of the major pioneers of its early development in the O,S,A, a, project Ship (joint space mapping of the physicc-chemtcal and subjective properties of leading Ilustralian market brands) was reviewed, 'IZle conclusions to be drawn at the current stage of completed consumer pilot work are: a) That more vocabulary elicitation work is required prior to the main exericise, b) That the sensory maps shou product discrimf~tion (albeit not strictly manufacturer based) largely against a global strength and a burn rate criterion, and arising from characteristicallY low levels of genuine discrimination ability, c) That results of the pilot stage, methodological fansideratfens, and the logistics of Wills product testing, indicate that an independant Jamplinglattribute mofadic approach such as that conducted in the U,S,A, would be inappropriate, it is concluded that a sequential moradic approach Would be mot appropriate in the context of Ship Australia, 3, It was ageed that as a starting point to researches on product evaluation incorporating smoking behaviour measures, eR i DC would O\ undertake the loanlsupply of a suitable monitoring device, and support its use by analysing data collected In the Australian market, Initial studies would concentrate on taking a commercial product and Clit; PDF -!::!!::!!::!.f3 StlC.;lnlil LI evaluating its (oonsumer) aoldng behaviour performance relative to si~ilu porlormanoe measures derived trom design variants or this co~erEiai NlntroC 4, Discussion held with the Australian Tobacco Institute yielded a fruitfll erebang ol views, and it was agaed to consider seleotive exposure d [I,R, I~BRdIkBIES and COOBTESY CIMPAICN ~~ormation, contingent entirely on Wills and central BIAlr, policy. Clit; PB F~T,r, i.! ~f3StlC.i: nill YLU*UIR J"1L6JnWIoLIYLTU 10 j Ir / 3D oiu(wrm~l .,, i P1 I "" I~v~oCd~~ I~ayg W~GHtaf Snm(i~6 DBCLw~cl,XA6 ~TR~Lrm) OP~LJIRHt~ /DPVH SLIII(~LOI I~P(niul Sw1~3 C h' Cc h' W C:lil;PDF - !::l!::l·~::cf3stlc.;:nlil B,A,T (U,K, b Export) limited, Research & Development Centre, SOUMAMP'TON I RPF/JP/46E fHE RELATIONSHIP OF BEHAVIOURAl SMOKING SfY1E PO CONSUMER SEGMENTATIOB AID ADAPTIVE RESPONSE TO SWITCHING REPORT NOI RD,2056_R~ SUMMARY This report considers the hypothesis that smokers develop a learned pattern of behaviaural responses associated with own brand smokingl this pattern of physical puffing actions~on the cf garette, the 'rmoklng style', Has both evidenced and found to he consistently associated with own brand smoking In the consumer switching study reported herein, furthermore, in terms of the abrupt switching involved in this study, it was demonstrated that, In the fmnediate stages of trialling a new brand, smokers will ' seek to maintain the behavioural style associated with the previous brand, The extent to which the new product falls short of delivering an expected pharmacological/sensory reward in relation to this pre-established 'pattern of effort' criterion is a likely prediction of extent of product rejection, The study also examines the scope far segmenting consumers on the basis of theft smoking styles with own brand, Four substantial segments based on different smoking styles are revealed; these segments are independent of tar band membership and show a relationship to sex of smoker, It is also shown that it is possible to segment consumers in terms of the particular style of their adaptive response when snitching to a product of a differing tar band to their usual own brand, The substance of this h' -i· CC ~~t. n , r ,rl Y ,n,l e,,.,'l-~hf~rwn ma~al k copied of Ihovn in unluthclisd pmca, C: I i I; PDF -!::!!::!!::!.f3 StlC.i: Dill segmentation a~pearr to be a distinction between smokers who attempt to 'Ennpensate' for reward loss by effort increase and those who opt to maintain their smoking style (ff not actually reduce effort) Indicating a product rejection response, The introduction of smoker style measurement into conventional segmentation routines is now feasible given recent advancer in portable smoking behaviour monitoring devices, The information given in 'ityle' regloentatfon may be pf strategic value In giving product developers a concrete behavicural criterion against which to evaluate prototype cigarette designs for their respective capabilities to meet appropriate effort·reHard relationships for the consumer, Style segmentation of the switching component in mature markets may be used with the aim of 'stabillslng' a dlsproportlonately high share of available switchers by producing product designs most sympathetic to the demand characteristics Inherent In the smoking style of the target switchers, ·ii· ~U-L ..I 1 ,,.I T~~~ ,,,, PUII not br~epitd or !hovn in unluLi~ ptnonr. Ln C: I i I; PDF -!::!!::!!::!.f3 StlC.i: 0111 ·29· TABLE 13 MI001E TAR SWITCHING C1USfERS x STABILITY Cluster 1 Cluster 2 Stable 3 6 Unstable 5 4 CONC/USI:ON this study set out to examine the evidence for consumers showing a learned behavloural style as a consistent response to nun brand smoking. The conclusion from this research, with the caveat of a weight of smoking effect inherent in the experinental design, indicates preliminary evidence in favour of the existence of a stable behavioural smoking style, Additionally, the results suggest that, when comparing smokers with long and short tenure to an 'own brand', the balance of evidence indicates that adaptation of smoking behaviour in response to switching products if a learned phenomenon and occurs only after smoking the new product for a period of time, At least in terms of the abrupt switching involved in this study It is possible to conclude that, in the immediate stages of smoking a new brand, smokers will seek to maintain the behavioural style associated with their previous brand. The extent to which the new product falls short of delivering an expected pharmacological Isensory reward in relation to this pre·established effort criterion is a likely prediction of extent of product rejection, Clil;PDF - !::!!::l!::!.f3stlc.;:niil . .... ..~~~..,,~e 1~~~1 ?I~cr?r;i~ulh6n~rd Dcnonr. ;hls study also set auf to exami ne the scope ton the assumption If consistent hehaviourat styles being measurable) fsr.estab- I Ishf ng a segmentatf on methodology based on consumer smoki ng ;tyler, It should be stressed that this study Is strictly at Illot scale, A corollary of this Is that any segnentatlon Iffect discerned at this level would represent substantive 'indings at the scale more usual for segmentation studies. :t is contludtd that using own bond aakiog Lehaviollr as a 1 ;egmentatfan data base in this study produced evidence of at east four discrete segments baled on different smoking styler, ]ithln the limited sample site It may be concluded that sloking ;tyle segment membership is statistically related to sex but ~ot to stability or tar band membership, larger scale research ,auld be requf red to determine the likely relationship between ;tyle segnent membership and speci ff c own brand usage. :inally, it may be concluded that it is also possible to segment :anslaners in tenos of the particular style of their adaptive esponre to switching to a prodrt of a different tar Band to ,hefr usual own brand, I .he substance of this segmentation appears to be a distinction I/ ~etween smokers who attempt to 'colapensate' for reward loss by !ffort increase and those who opt to maintain their usual lehavioural smoking style (if not actually reduce effort), ndlcating a product rejection response, he immediate conclusion of this report, given the sample sizes ·mplayed, must be a recommendation for a more substantial fallow· Ip study, However, it Is clear from the i imited data herein that ·egmentation by behaviounl snaking style is a viable concept, 'he introduction of smoker behaviaural style measurement into onventional segmentation routines is now made possible by .ecent advances in partable smoking behaviour monitoring devices, he i nf amati on given i n 'style segmentat i an' may be of st rategic (: I:p[~F~III:~i:~:nll:;"nl L~mad I~L~I~I~L!:1Dl'd'" IhuvnD unolhumnlp~n*n~ BRmsH AMERICAN ToeAcco CopyR~quatFarm RepoestNumltcU Organization: Dept, of Justice- Legal Health, Canada i: Request Details I~quatDaB: la'S:f4 )1111·11·11··111(01··1)(11101(111·1(~((11111)(1)11111101~)~11111)1))1)(1111(110 ~ 2~s I ...,.Box Number: ,......fa.i.?i~~....,......... File Number: ,..,..,i........,,.,..., Page Range: First Page I Llast Page olzl~li~l4 I8l";ilYI· I ITa lz16141rlclid ~n.i~41~1~9 h- I· I /10/i121olo10 Requoted By: print 8rmoi.~h~l., ~..~~tFN..,,,,.,. ,,,,,,,,,,,, ,,,,,~~~~ Details below will be filled in by Depository Staff Only n Copy Details cara~.~ LL )11111)~11((1~11(111()1((1(~(1~(1111111)1)11()1)1(1111 1 9 MAY 1994 111 )(111111(1111111111111(1 111111111(()(1Il)t)ll~l ~.Time: S_.Zlt.e Date: h c~wn.td4:.....,...,.....?.;...1~!! ' 1111111·111 111)111·(1111 )1)1)1((11 l)ate: 111)(11)1(111111111(11111111~)11 111 Delivery Details Checked By: 111)((11)111)11(01((11(11111)~ 11(·)1(1)1(111111111)1111111)1111 Clit; PDF -!::!!::!!::!.f3 StlC.i: 0111 CONFIDENT~AI 1985 B.A,T. RESEARCH CONFERENCE "CONSUIIER RESEARCH TECIIQ~ES" DISCUSSION PAP"YR R.P, FE,PRfS October 06, 1985 ao . Clil:POF- i:·':'i:· i~jtlc.;:e ii ------------ ~B~B BESIYQ QaIqms -- This paper will consider the following areas in sequence: I, Criteria ior consumer research, a. The current Blr scenario, and a critique. 3, gt~ developments and future directions in cdnsumer research, 4, The case ior research on consumer methods, CBf~BId POB CO~S~B ~8Xd~B Criteria for consumer research Ire easilp derived and generally agreed, It the most collnon level the criterion ~ould be: To lore fully charleterise the consu3lar in order to more Iccaralely guide marketing actions. the specific level, the hel criteria ~opld be: i. To measure consumer reaction to blind and I or branded product offerings in such a way ns to understand sensory response and determine product I brand preference or choice, (PaDCCT ~ Png~BB ~STIIG), a, To differentiate ~ithin the consumer population in such I tpaY Po to derive clearly defined sob-groups to~nrd which a more refined and targeted marketing strategy may be directed, (SGGIIEIUT~IOB ~QIa~S), 3, To test-market prober I brand in the field in such way 98 to predict potential mPr~et shore. (TgSTU~TS WD T~P~ OPWB ~STIBC], I, To understand shifts in consumer motivations and attitudes in the marketplace lhleh may impinge on product usage patterns dr Predict new aaes (bRIT~T~BI SUBVEYS), it is I central thess of this paper that ~hors~o criteria for consaPer research are generally agreed, criteria of eterllsnee C ore insuificieotlg considered, cltpr~7~ - To consider criteria of excellence, we would have to add the :ollo~ving concepts to our discussions of consumer research: i, SENSITIVITP : the extent to which any method demonstrates sufficient precision and accuracy in measuring the phenomena it purports to measure, 2, RELEV~VCE the extent to which the method measures that which is relevant t~ measure rather than that which is readily measurable; and the extent to which findings are subsequently operatianalisable, 3. PREDICTIVE POWER the eaaabilitg of the method to providk accurate predictions of consumer behaviour in the marketplace, 4, TI~E/COST EFFECTIVENESS the extent to which the method enables the client to achieve his objectives fully within the least time and cost, It is the central theme af this paper that manp of the research techniques currently in use around the BAT Group would benefit from a stringent examination against these criteria, THE CURRE~~JT BE SCENARIO ~ND d CRITIQUE The mast general observation to make about BAT consumer research, comparable for example to that of our main competitors, is chat it is marked by an adherence to traditional methods, This situation is underpinned by three factors: i, Functional / departmental sub-divisions within Operating Companies between product; strategists and brand strategists, which inhibits Ihe use of newer, more flexible methods which rely on freer, twolway interfaces between functions, 2. Conservative attitudes toward the adoption o~ new methods, exacerbated by lack of central in-Group evaluations I advices on the performance characteristics of new research " methods which have been increasingly available over the O last two decades, h\ 3, Reluctance on behalf of market research agencies to adopt and supply new methods, especially where BaT is nor pressuring such agencies to do so while providing guidance and support; and most especially where such new methods are time I cost-~eclive, It should be clearly understood that consumer research authorities would agree almost all major innovations in research methods have occured within the last 20 years, a period during ehich BAT has been 04 par aith competition in awareness and conceptual understanding of what is going on, but notably behind competition in adopting such nea practices. Revieving dominant research practices around the Group, it is not surprising to find a greeter level of sophistication within the C.1.C. Operating Companies. This broadly manifests in 6 areas: i. Ilore direct and extended. access to consumers,invalving a more immediate interface between the consumer and in-house personnel, iYotable examples of this are the [T.If. Charter Research facility, the U,S. Dupdnt Circle facility, and the German Postal panel approach, The strengths of such approaches have been demonstrated as a) providing an environment in which to research ways of doing research (esp, U,1(.] b] providing highly time i cost effective means of collecting consumer data (esp U.S. and Germany) and c) synergising the relationship between product development and brand marketing (esp, Germany), 2. I]se of more sophisticated segmentation strategies which do not simply rely on that which is easiest to measure (i,e, sociodemographics, price I brand usage, delivery I brand usage), but which attempt to measure those factors which are considered salient (price sensitivity, life style, smoking and health I social concern attitudes, imagery responsiveness jmast notably U.S, and Canada), 3, Commitment to large scale surveys which may have no immediate brand marketing goal, but Dursne the broader objective of understanding the develbpment of attitudes I motivations and behaviours in the marketplace, with the O objective of better directing future marketing actions (notably Canada and U,K, [libn and driesi i, 0~ 4, d move to Product testing with an emphasis on assemblicg Of SimU~at: -ng the total offer (esp, IJ,S, and Buslralia)~ ~ ,,,,,,,, from monadic or pair i~ise conceptions of praduet / brand comparirons to product acd marter ~Efi~ tall C,A,C, Companies), 6. Developing experiences with new-generation market research methods. (gatably simulated test market techniques in U,S,, Australia and Germany; and product cornmunication techniques in D,S,), Examining the broader perspective of the BAT Group, we see a consumer research scene still dominated largely by the quality of service and range of techniques offered by local market research agencies, Correspondingly, a heavy dependence is placed on almost exclusively contracting research out to such agencies (~irh the exception of in-house sensory panels), In general, product testing is by sequential monadic or simulataneous placement, restricted to pair comparisons of products, Methodological debate engages at the level of a) sequential ;oanadic v, simurtaneous placement b) pre v, past positioning of the preference question, and c) direct v, com~arative attribute scaling. Debate tends not to extend to abeSher or not such an approach is inadequate per se, despite the fact that traditional preference testing of this sort is highly time I cost inefficient, very sensitive to consumer sampling effects, and incapable of setting the results in context of the brand space structure of the market. Segmentation techniques are largely by sociodemographics or price I delivery categories in the market, The tendency is to use segmentation. retrospectively i.e, to attempt description of particular brand / delivery smokers, A more important criterion for segmentation research should be one of prediction of likely product or brand acceptance I reject'ion by particular consumers , The sequence from product testing to brand testing still tends to be conducted in a series of discontinuous stages with several discontinuities in the attribute sets at each stage and a disturbingly high level of attribute irrelevance and redundancy. For practical purposes, an irrelevant attribute is O one that can be neither interpreted nor operationa!ised, aad~a redundant attribute is one which measures substantively the same thing as another attribute. C.d,C. Company methodological studies have amply demonstrated the existence of irrelevance and redundancy in product test questionnaires, and have elaborated the kind of techniques which can be used to bring about improvement (factor analytic methods), In terms of general critique, several major issues can be identified, rectification of which have two prime impacts on the business; i, ~n reducing the probability of failure of any brand in the marketplace, and identifying previously unacknowledged consumer needs I product opportunities, 2, fn encouraging use of more time I cost effective methods in order to increase the research buy of a given market I product research budget, These keg issues of critique are: i. back of integrated systems of attribute vocabulary, linking in-house product testing through to brand testing in the market: This situation, exacerbated by traditibnal professional and functional sub~divisions in companies leads to: a) inability to relate results of ialhouse testing to external consumer test results. b) inability of the marketing function to provide a brief to product development which is e~oressed in attribute terms which are operationalisable and measurable, 2. Use of segmentation models which are designed as guides to ---------- brand targeting bur, give very few clues about product targeting (beyoad broad delivery, construction and blend guidelines). In this respect, the point has been raised from the Canadian e Smoking Behaviour Conference that priority action should be r\J addressed to determining whether it is oossible to routinely C~ build into conventional segaen~a~ion apbroaches a hard ~O: behavioural measure of smoking style. Such measurement ma4' be used to refine ~~e design brief of products io order to ~D sympathetica:ly model to Farticular consumer Sehavioural Clil;PsF - !::!!::!!T.t3~lc.;:nlil demands, The guiding principle Would be that for any given market segment currently being used to direct brand marketing strategy, it should be desirable to characterise the smoking style of that segment in order to help product strategy to best complement the brand strategy, 3, Inadequate modelling of the consumer produce, choice process. Conventional techn~sues tend to treat all product attributes as independent and equally weighted. Consumer perception has neither of these qualities, Product test and preference models which take account of attribute interdeoendence and importance weighting are no~a becoming available jnotable examples being conjoint and trade-off preference analyses), Additionally, in the U,K, such models of preference are being explored in combination with segmentation styles of analysis ( supply of such techniques nill shortly be for~hcoming)~ The strength of using conjoint methods embedded in a segmentation scheme is that it becomes possible to segment consumer groups an the basis of contrasting product benefit requirements~ 4, Lack oipro-ac~ive, or mapping ~echnioues. losl~caatenrional ;ae~hods deal with measurement of current scenarios, such methods cannot deal with "~Rbatlifs", Techniques are available which do attempt to model alternative scenarios, in particular, the extent to which mapping techniques mag be used for both product and brand mapping nill bring about improvements in targering and, especially aar~ec gap analysis, Equally potential and under-used is the diagnostic feature non offered by many simulated.:est market methods. B simulated test market method (e,g~ Assessor) is most usually know as a test market simulation method which has the capacity to predict long run market share of new product launches without the necessity to go to full scale market test. Ihe diagnostic sub-routines of these models are less widely acknowledged and used, Diagnostics allow hypothetical product concepts to be evaluated for potential market success, and have sbau~n considerable Value in other produc~ areas for identifying genuinely innovative propositions (e,g, the snack food industry is heavily invested in diagnostic STY's), Clit; PDF -!::!!::!!::!.f3 StlC.i: Dill lack at insistence on high quality ~quali~ative research, ~rct~part a: the B1~T Group, qualitative research remains an euphemism for badly conducted 'chats' between people, and group 'gossips', Since creative qualitative work is a keystone to new product development; three areas of improvement are called for: a) Pressure for higher standards of agency research supply, b) Close involvement of 3rdducl oriented personnel in observing such groups, c) Implemenra;ion of new, more structured and extended qualita dye techniques which apparently provide disproport~onatelg higher creativity of results Isee especially the use of EPSP technique by BdT Suisse), In summary, and with the exception of leading companies, BaT consumer research tends to be: i, agency supply dominated, 2, Expensive and slow (note the cost of consumer research has been escalating dramatically over the lasr, decade throughout the world), 3, Out-dated, 4, ~ot particularly predictive, ,YEW DIRECTIONS .INI1 FI~URE DEVEIOP~E~TS IN CONSU~EBl BESEaffCII General future treads in consumer research are discernible in the literature and can be summarised as follows: i, d trend toward smaller scale and total proposition testing, and use of more time I cost effective methods, 2, 4lodels of testing which integrate the product development with the brand development process more closely, 3. Increasing sarhisticatioa of sample segmentation in O\· product tas~ing, C I i I;'P B F~I n:f j~f'l'C'li:l~l ii .. 4, Increasing sophistication of statistical models to deal with attribute contribution / salience in the choice situation, 5, Segmentation for prediction as well as characterisation, through correlation of behaviaural to subjective / attitudinal ~asures, 6, Increasing emphasis on unders~ar,diag switching patterns (which are key to understanding opportunistic marketing where precise product targeting is a central concern), TIIE CASE FOR RESEARCH ON CONS~ER SESE~RCA It follows from the discussion so far that considerable opportunity exists for development: of aBT product testing expertise into some or" the new generation techniques, For exa~nple: Time-cost effectiveness may be improved by adoption of small scale sample preferencetests and use of simulated test market approaches (which, it is recalled, have a capability of predicting short long run market share rPithau; the necessity of full test marketing), Creativitp of identifying future product ogportunities and 'what-if' scenarios will be enhanced by use of extended pualita;ive techniaues (e.g~ EPSY), the diagnostic STy technique, and con,oiir; approaches to benefit segmentation (currently in development in U,K,). Comprehensiveness of testing designed to evaluate the relative contributions of different classes of variables to overall product acceptance I rejection will be improved by the use of ; statistics joint space mapping techniques which use 'locking (e,g, canonical correlation) to relate variable sets (e,g, smoke chemistry, smoker behaviour, brand imagery, sensory evaluation), Comorehensiveness is also a feature of total offer test methods and S,T,U.'s, Another aspect of research coverage in the conte.u~ of continually increasing adoetlising restriction should C eaarinue to be eacoural·llear of ;ie use pi reehnioues tor h~ ~easurilg pack I me:cb~disil~ i advertising stand-out (see U.I resea.rches ii I:iS l~el, aid l..Z, r:dies !nc!udin(;he Pereeptor method), - Predictiveness, ~a~y new generation models of segmentation have featur~-~ strength in prediction, Yew approaches which are particularly promising are a) OPTIJ~ISED SEG~ST$TfON, which is an approach which accounts for the potential profitability of segments within the analysis and b) CO~BPONENTIII, SEG~E~TaTION which analyses on the basis of the interaction between consumer type va:!ab:es and product reaction type variables, These new generation approaches arel currently under evaluation within BaT ~I(& E R & D, RECO~EHDAT IONS ddvancing the BilT investment in consumer research should follow a number of routes: Conduct of a Qouolwide audit of consumer research practices, which should be directly evaluated against the earlier defined 'criteria of e~cellelce', 2, Support of an agreed locus for research on research methods, preferably consultative and mobile across BET functions, Operating Companies, and market research agencies, The role being to encourage appropriate near methods and assist in synergisin: company / agency relationships and R&D / Yarketing relliionshi~s, 3, Stimulation of new initiatives with research suppliers to move to newer time ; cost effective methods, 4, Promulgation of new generation methods, including those currently appearing in C.S.C. Co~pacies, into 9dT CO Companies where appropriate (i.e, consumer technology transfer), Debate on alternative functional I disciplinary structures which may im~roce product e:eG'elor,ment I brand mlr~eting synergy in OPe~sing Comcanies, Clit; PDF -!::!!::!!::!.f3 StlC.i: 0111 6, Delelopment 09 'custam' models and computer ana!psis programmes within and for the 38T Group, nrith the objec~ipe of a) reducing dependence on researct. agencies, b) reducing research costs, and c) gaining access to presently unavailable techniques, (For example, ISSESSOR is unavailable in Europe as such since the suite o~ programmes was initially modified far use in Switzerland by R,J,R,), a,P, FE891S Clit; PDF -!::!!::!!::!.f3 StlC.i: 0111