The research register for this journal is available at http://www.mcbup.com/research_registers The current issue and full text archive of this journal is available at http://www.emerald-library.com/ft European Journal of Marketing 35,9/10 1058 Received June 2000 Revised January 2001 European Journal of Marketing, Vol. 35 No. 9/10, 2001, pp. 1058-1073. © MCB University Press, 0309-0566 Strategic political segmentation A new approach for a new era of political marketing Gareth Smith and Andy Hirst Loughborough University Business School, Loughborough, Leicestershire, UK Keywords Politics, Marketing, Market segmentation, Strategy Abstract This article begins by analysing developments in political segmentation over the last decade. Using an appropriate database and statistical approach, segments of the British electorate are identified. Conservative and Liberal Democrat segments are then analysed and issues affecting their likely electoral performance discussed. The Labour segments split into distinctive "old" and "new" Labour camps. As attitudes differ widely across these segments, the two most different segments are targeted for further analysis. The issues which most discriminate between these two Labour segments are highlighted and some suggestions forwarded on how policies might be positioned for these disparate segments. The article concludes by considering the stability of political segments over time. It also discusses the Emits of strategic segmentation in politics and identifies further research opportunities. Introduction: developments in political marketing over the last decade Over the last decade there has been a marked increase in research into marketing's relevance and application to politics, both of a theoretical (Wring, 1997; Butler and Collins, 1999) and applied nature (Hayes and McAllister, 1996; O'Cass, 1996). In 1990 it was possible to conclude that, using the idea of evolutionary stages in the development of marketing (Keith, 1960), political marketing had moved from sophisticated selling (using advertising/election broadcasts, etc. to "push" party political ideas) to a "nascent marketing era" (Smith and Saunders, 1990, p. 296). This marketing era in politics manifests itself in such activities as image building, issue tracking, the targeting of voters, timing of elections and aiding in policy formulation (Kavanagh, 1995). The plethora of new research covering these issues makes the epithet "nascent" look increasingly redundant. Ten years on, there is evidence that we are now moving to another stage of marketing's development in politics, namely the strategic marketing era (Newman, 1994; Butler and Collins, 1996). This latter stage manifests itself in the use of marketing not merely as a short-term tactical device, primarily used for information gathering in the run-up to elections, but as a longer-term and permanent activity to ensure continued governance (Nimmo, 1999). For the strategic era to be fully embraced in politics will require a concomitant development of strategic segmentation. Porter's (1980) generic, market-based strategies for achieving competitive advantage support this assertion. Both "focus" and "differentiation" require strategic segmentation. Focus requires the targeting of segments and development of appropriate marketing mix strategies. Differentiation involves "positioning" the offering of the organisation in the minds of the consumer (Ries and Trout, 1986). Typically, positioning is viewed as part of the strategic segmentation process (Kotler, 1997; Hooley and Saunders, 1993; Myers, 1996). Accordingly, it seems reasonable to conclude that those political parties wishing to embrace the new "strategic marketing era* and gain a long-term competitive advantage over their opponents, will need to develop their segmentation in a more strategic way. Recent developments in political segmentation Since 1990 the development of segmentation has seen a shift in emphasis from the traditional (geographic and demographic) methods of segmentation towards an increased use of psychographic/attitudinal bases to segment political markets (see Table I). The Labour party appear to have started the trend for psychographic segmentation in Britain in the early 1970s when it identified segments such as "Jack" (new style Labour), "old Fred" (old style Labour), "floating left" and so on (Rosenbaum, 1997). Since then its use has increased in line with computational power and appropriate statistical techniques. In 1992 the Tory party employed Richard Wirthlin, who, having proved influential in Reagan's successive presidential elections, developed a psychographic/attitudinal segmentation approach purpose built for application to political marketing (Rosenbaum, 1997). However, the fact that a more sophisticated range of segmentation bases has been used over the last decade does not, of itself, support the notion of more strategic segmentation. A strategic approach to segmentation emphasises a long-term customer focus and requires the integration of three discrete activities, namely segmentation, targeting and positioning (Kotler, 1997). Although there has been theoretical acknowledgement of segmentation as a strategic activity in politics (Newman and Sheth, 1985; Baines, 1999), the applied research (Table I) shows little evidence of this approach. Political segmentation is thus following the practice of most organisations in using segmentation simply as a means of breaking down markets to aid short-term managerial decision making (Piercy and Morgan, 1993). To clarify what constitutes strategic segmentation, an approach, amended for use in the political domain, is supplied next. Applying strategic segmentation to British politics The strategic segmentation process in Figure 1 begins by discovering the segments that exist using appropriate criteria. Clearly this is critical as segments may be constructed using any base, however irrelevant (e.g. using attitude to DIY to assess voting behaviour/party loyalty, for example). Second, decisions need to be made over which segments are attractive enough to be European Journal of Marketing 35,9/10 1060 Segmentation method Political maketing up to 1990 Methodology Example Political marketing post 1990 Methodology Example Geographic/ geo-demographic A priori Demographic A priori Behavioural A priori Attitudinal/ psychographic Limited post hoc Post hoc Table I. Development of segmentation in politics since 1990 North-South Divide (Heath, Jowell and Curtice, 1985) ACORN3 Geodemographics (Yorke and Meehan, 1986) Social class (Butler and Stokes, 1971) A priori US Presidential Election (Shelley and Archer, 1992) A priori UK regional analysis (Curtice, 1996) A priori Gender rates across Europe (Scott, Braun and Alwin, 1998) A priori Middle-class politics in Britain (Heath and Savage, 1994) Economic theory of A priori democracy US politics UK politics (Smith and Saunders, 1990) Late deciders in UK politics (Hayes and McAllister, 1996) New Labour vs Old (Bottomley and Curtice, 1999) Unknown Psychographic segmentation of UK politics (Wirthlin quoted in Rosenbaum, 1997) Sophisticated The "American a priori dream" research (Newman, 1999) A priori Attitudes and politics under Thatcher (Heath and Park, 1997) A priori Political partisanship and attitudes to the EU (Evans, 1998) Note: a ACORN is an acronym for A Classification Of Residential Neighbourhoods. It links consumer behaviour to demographic and geographic classification targeted (i.e. segments that are very disparate/small, etc. might be ignored or given less attention). Third, a better understanding of the targeted segments is required. From this better understanding, parties and their policies may be positioned more effectively; some issues having greater resonance with some segments than others. Then the appropriate marketing mix activity may be formulated in terms of the actual policy/issues to highlight or downplay for a given segment, the way to communicate with them and so on. The segmentation, targeting, and positioning (STP) approach (Figure 1) promises better targeted policies towards identified and attractive segments. However, the political market is different from the private sector business market (Lock and Harris, 1996). Positioning in politics, for example, is not a value free activity. Edmund Burke (1774) told the electors of Bristol "Your representative owes you, not his industry only but his judgement; and he betrays instead of serving you if he sacrifices it to your opinion". His point is a strong, modern warning against any "full-blown" STP approach that elevates public opinion over the views of elected representatives - with all that this would entail on issues such as repatriation, asylum seekers, the death penalty and so on. The question of how far the STP process ought to be applied in a liberal democracy is not clear-cut therefore and will be returned to later. Research methodology As indicated in Table II, "a priori" and '"post hoc" approaches are the two basic methods available for segmentation research (Green et al, 1988; Myers, 1996). A priori segmentation occurs when the researcher decides how the segments will be formed prior to conducting the research (e.g. using voters with high, medium or low levels of party loyalty as the pre-determined segments whose views will be researched). Post hoc segmentation involves no pre-judgement of the segment bases at the outset. Instead the segmentation is achieved using a statistical technique which places respondents into groups with others who have similar views/responses to questions asked. According to seminal research in the area, segmentation for "understanding the market" should involve many bases for segmentation (Wind, 1978). Given that the position faced at the outset of this research was to understand the segments that exist in the British political market, all of the possible generic bases of segmenting (attitudinal, behavioural, geographic and demographic) were used. Also, as there was no real theoretical guidance for deciding STAGE 1 STAGE 2 STAGE 3 'Segmenting' 'Targeting' 'Positioning' the Political Market Political Segments Political Segments 1. Identify bases for segmenting electorate. 3. Develop measures of political segment attractiveness. 5. Develop positioning for each political segment targeted. -w 2. Develop profiles of resulting political segments. 4. Select the target political segment(s). 6. Develop appropriate marketing mix for each target political segment. European Journal of Marketing 35,9/10 1062 Features Table II. The British social attitudes database, 1997 Annual longitudinal survey (since 1984) 1,355 respondents 46 attitudinal questions covering: Economic issues/taxation Social services TheEU Single currency Transport/car usage Environmental protection The green belt Job satisfaction Work ethic Left-right perceptions3 Authoritarian-litertarian perceptions3 Importance of social class Two specific voting/behavioural questions: Likely voting intentions Strength of party identification (partisans, sympathisers or residual identifiers) Eight classificatory (geographic and demographic) questions: Sex Age Income Occupation Social class Education Region Marital status Note: 3 Left-right and authoritarian-libertarian are derived variables from other attitudinal questions within the survey. See Appendix for details appropriate segmentation bases at the outset, the logical approach was to adopt a post hoc methodology involving no prior judgement from the researchers. Once segments are better understood, specific bases may be chosen a priori to investigate issues that arise (Wind 1978). Thus, post hoc followed by a priori approach is used herein. First, a wide listing of segmentation bases are used in a post hoc way to break the market down into its constituent segments. A priori segmentation is then used to investigate those specific issues signalled by the previous post hoc segmentation. This latter stage equates to a party manager subjectively choosing to investigate further an "attractive" segment with a view to developing a positioning strategy for it. This research process follows, therefore, the STP process identified in Figure 1. Sampling frame The latest available data from the British Social Attitudes survey (just prior to the 1997 election) was used to segment the political market. It was chosen because of the quality of its data that seeks to "monitor and explain changes in the social and political climate" (Jowell et al, 1999). The survey combines enduring social issues relevant to politics (such as the economy and social services) with currently salient issues (such as the environment, transport policy, etc.). It also contains information on the partisan allegiances of respondents plus relevant demographics and social classificatory data. Table II provides more detail of the database's content. Data reduction and segment evaluation As the database is so large, the first step was to use principal component analysis to reduce it to more manageable proportions. Due to missing data, the final sample size had 500 usable cases. Principal component analysis using Varimax rotation was conducted with a solution determined by eigenvalues greater than 1. The final rotated solution explained 63.1 per cent of the variance in the data. The individual factors were examined to establish internal consistency using Cronbach's coefficient alpha (Devellis, 1991). Factors that exhibited both strong internal consistency (i.e. alphas between 0.65 and 0.95) and substantive meaning were summated to produce a single variable, and variables that did not, were kept as a single item to prevent any significant loss of meaning. This left 16 variables that were then used to cluster respondents. Examination of clusters led to a clear seven-cluster solution. The attitude data plus an appraisal of party identification and other demographic and non-parametric data was then used to describe the nature of the clusters. The results are presented in Table III. After pre-examining the univariate and bivariate statistics, multiple discriminant analysis (MDA) was performed using all 16 variables to examine two important Labour segments (combined number of cases = 120). A stepwise approach was used because theory provided little guidance for model specification (O'Gorman and Woolson, 1991). Typically MDA is performed with a hold out sample to measure the stability of the coefficients produced. The low variable to case ratio meant that an alternative approach was needed. The stability of the coefficients was therefore determined by examining the results of progressively smaller samples (Kohli, 1989). The stepwise procedure found five variables to be significant at the 0.001 per cent level and, as can be seen from the classification, the variables chosen through the stepwise MDA correctly classified 95.6 per cent of cases, which is much higher than would be expected by chance. When the analysis was conducted on smaller samples the results remained consistent. Therefore we can be confident that: the coefficients are stable; and the variables provide good predictive validity. The results of the analysis are presented in Table IV. Analysis of results Stage one. Identifying the segments - the magnificent seven Table III provides a synopsis of the findings resulting from the factor and cluster analysis. Segment II III IV V VI VII Name Share of populace (%) Party share of segments Conservative (%) Labour (%) Liberal Democrat (%) Party identification3 Attitudes Right vs left3 Libertarian vs authoritarian3 NHS (run well) Social class (more important) Performance of GB economy Lower taxes (at expense of social services) Increase tax to increase services Positive/negative about EU Pound vs Euro Feel good factors More prosperity Job satisfaction "Old Labour" Poor 19 25 69 6 Labour partisans Left wing No Yes Poorly No Keep pound "Feel Gooders" 5 47 35 18 Tory Labour Mainstreamers Nationalists 27 60 30 10 Conservative Conservative partisans partisans and sympathisers Yes Very well Yes No Well No No 12 26 58 15 Labour partisans Left wing No Yes No No Negative Keep pound Keep pound Keep pound Yes Yes "Underwhelmed" Loyals 14% 41 54 4 Labour and Conservative partisans and sympathisers No Champagne Socialists 10 72 20 Labour partisans "Light Green Lefties" 15 23 59 18 Labour partisans and sympathisers and residual identifiers Very left wing Very left wing Libertarian No No No Yes Very well Definitely not Definitely not Yes Keep pound Keep pound Yes Positive Keep pound Yes (continued) Segment II III rv v VI VII Feel good about employer Work ethic (good?) Environmental issues Environmental protection Green belt (need to protect) Car usage (controlling) Transport policies (well sorted) Demographics Age Sex Income Education Socio-economic group Region Yes Yes Yes - More needed More needed Agree Agree strongly Agree More females Many poor (<£10k) "0" levels Junior white collar, skilled blue collar and unskilled Northern > 45 Many poor (<£10k) Graduates or "0" levels Junior white collar and skilled blue collar < 44 Spread - Low and high earners (<10k and >£40k) No Agree Disagree Agree 24-44 More females Middle (£10-20k) Mix of professional, managerial and blue collar "0" levels Junior white collar Agree Yes Yes Much more Much more needed needed Agree strongly Agree strongly Agree Disagree More males Middle (£10-20k) Professional/ managerial plus supervisory/ unskilled South East Better off (more >£20k) Graduates Professional/ managerial Southern 35-54 More males High (many >£20k) Graduates Professional/ managerial South East Notes: a The definition of these variables is supplied in the Appendix. - Indicates where no obvious difference within the segment is present a 0 01 w CD Crq o n W European Journal of Marketing 35,9/10 Control car usage +ve or -ve about EU Performance of economy Green belt protection Lower taxes/worse social Old Labour Discriminant variables Coefficient (mean) 0.652 0.384 0.372 0.356 3.22 3.23 3.66 2.1 New Labour (mean) 1.85 2.53 3.04 1.35 01d(%) New(%) Old Labour 95.6 4.4 New Labour 4.3 95.7 95.6% correctly classified Predictive group membership 1066 -0.751 3.7 4.67 services Table IV. Results of multiple discriminant analysis Summary statistics: Eigenvalue = 2.365 Wilks's lambda = 0.73 Chi-square = 159.556 Df = 5 sig 0.000 The seven clusters/segments that resulted reflect the overall popularity of the Labour party in the immediate run-up to the election. They are the clear leaders in four segments (I, IV, VI and VII). By way of contrast, the Conservatives are the main group in two (III and II). As with a similar analysis ten years ago, (Smith and Saunders, 1990), Liberal Democrat problems are reflected in their not "owning" any segments. The Conservative position is an interesting one. Segment III is by far their biggest segment. It also unites a wide range of Tory voters in terms of their level of allegiance to the party. Both partisans and sympathisers are well represented therein (see Appendix for definitions). Also, there is a notably wide spread of income, education, occupation and region. Despite these differences, the Tory voters in this segment are remarkably similar in not feeling very strongly about any of the social and economic issues covered. It is for this reason that they have been dubbed "Tory Mainstreamers" - though "middle of the readers" would fit equally well. Interestingly, this segment shows similarities with another segment, the "Underwhelmed Loyalists" (Segment V). It too has a strong presence of partisan Conservatives and they similarly do not feel strongly about the issues of the day. As Segment V also has a high number of Labour loyalists it is possible that both segments are relatively apolitical despite their partisanship. They may well reflect voting behaviour that is habitual and historical, as typified by social class/family based voting (Butler and Stokes, 1971). It is also tempting to interpret the Tory Mainstreamers' moderation as a sign of resignation about the general election to come but this would be pushing the data too far. It may be that there are other factors driving behaviour in these segments that have not been identified from the data collected. The other Conservative segment (Segment II), whilst much smaller, is dominated by Conservative partisans. As signified by the title allocated to them, these "feel gooders" are happy about lots of things - their job, the increased prosperity of the country, the way the economy is performing. They are even happy with the way the NHS is run. Their salaries are not great yet and their age suggests they are starting rather than finishing their careers. Most of the Liberal Democrats' support (63 per cent) is in segments dominated by Labour. Also, given the party's long-term commitment to libertarianism and environmentalism, it must be of concern that they do not perform better in the Labour Segments VI and VII, where these issues are seen as particularly important. In fact, their support is relatively uniformly spread across all the segments, despite the marked differences between the segments. It is interesting, for example, that the biggest single group of Liberal Democrat voters is in the largest Conservative segment, the Tory Mainstreamers (i.e. segment III). It may be that the party means many things to many people or their supporters are voting against the other parties rather than for Liberal Democratic policies. Reasons for this spread need to be better understood before any positioning strategy is formulated. The position of the Labour party is much more positive, as might be expected given their imminent electoral victory. However, from a party manager/marketer's perspective there are a number of issues that the analysis highlights. First, it is notable that although there are four strong Labour segments, they are split into two broad camps. Segments I and IV are similar in that they exhibit "Old Labour" characteristics. They believe social class is still an important issue, they are not very interested in the environmental issues, do not have much by way of a feel good factor and, in segment IV, are anti the EU - little sign of Blair's New Labour here (see Bottomley and Curtice (1999) for more on the "Third Way"). They are markedly different from Segments VI and VII who are the most "left wing* of all the segments. The "Light Green Lefties" segment is so called because of their attitude to environmental issues but also because they are always less radical than Segment VII. The latter segment has the most party loyalists and is the most left wing. It is also the wealthiest segment overall and by far the best educated - hence the title "Champagne Socialists". Strategic political segmentation 1067 Stage two. Targeting segments: Labour's dilemma. At this point each party faces different problems on which segments to target. Labour is chosen as our focus because they have four potential target segments to aim at. Moreover there is a clear polarisation of these segments. Consequently, Labour provides more scope to test the application of the STP approach than, for example, the Conservatives with their dominant, main segment. Segmentation theory suggests that "good" segments are those that possess criteria such as measurability, substantiality, accessibility, differentiability and actionability (Kotler, 1997). As the segments are already known through the clustering process and are relatively sizeable (with the possible exception of Segment II), the first two criteria are satisfied. With the raft of direct marketing techniques available to political marketers and the ability to target messages in party political/election broadcasts, the criteria of accessibility also seems European Journal of Marketing 35,9/10 1068 satisfied - especially considering the range of classificatory data that is also available from the data to target the segments. The critical issues are those of differentiability and actionability. Are the segments different enough on important issues to merit being treated as discrete segments and will the views they hold allow for action via a coherent manifesto/platform. The issue of joining the Euro does, for example, clearly differentiate voters. It is much less clear if the "wait and see" versus the "not within the timescale of the next Parliament" strategies of the two main parties is an acceptable compromise for the electorate or not. The preceding cluster analysis put into segments people with similar attitudes, but could not give sufficient guidance on what issues are the most important discriminators between segments, i.e. their relative differentiability and actionability remains unclear. The two Labour segments that look the most different at a superficial level of analysis are Segments I and VII (a fact confirmed in general terms by the partitioning process in the cluster analysis - those at the extreme ends being the most different in their expressed attitudes). They are the two segments most likely to pose problems for the party managers and, as such, they have been chosen for further analysis. The choice is a managerial/a priori one reflecting the likely information needed by party managers seeking to develop a coherent positioning strategy across apparently disparate Labour segments. Stage three. Positioning "Old Labour Poor" (OLP) versus "Champagne Socialists" (CSs) The results of the discriminant analysis are produced in Table IV. Table IV identifies those issues that most differentiate the two Labour segments targeted. The biggest differentiator is their attitude to lowering taxes at the expense of other welfare issues (e.g. pensions). Not surprisingly it is the CSs who are most strongly against this with the OLP slightly in favour. Given the relative disparity of wealth between segments, this is hardly surprising. The OLP may well feel that they need lower taxes more than the CSs. The next most important discriminator is their differing attitudes to the car. The CSs are very positive about controlling car usage while the OLP are slightly against such a move. Again it may be that the well-off CSs living in the South East are economically able to use all forms of transport whilst experiencing the negative effect the car has on urban life. This may be contrasted with the more Northern OLP group where the car is more positively viewed and seen as less of a threat to the environment. As shown in Table III, the research indicates more OLP coming from the North than CSs (chi-square 13.06, sig. 0.001) whilst several segments, including Segment VII (Champagne Socialists), have a southern bias (chi-square 14.92, sig. 0.001). This suggested North-South divide links together the remaining three discriminating issues. Segments I and VII are both in favour of protecting the green belt but, not surprisingly, the CSs are most in favour. Once again, their educational platform may be playing a role here as well as the relatively different pressure on land between the North and the South East. The OLP are most negative about how the economy is performing which, given their relative wealth and the relative disparity in economic performance between the two areas, might be a factor. The final discriminating issue concerns the EU. The OLP are more in line with the other segments by being generally opposed to the EU. The CSs, perhaps because of their geographical proximity and/or intellectual conversion to the benefits of closer co-operation, are the most positive of all the segments. As such, they are closest to the "New Labour" agenda on this issue. It is tempting to overplay the New versus Old Labour argument when contrasting the two segments. This would be dangerous however. Despite the differences just discussed, on most issues the two segments do not differ significantly (see Table III). This means that a common message on these issues may be used for both groups. With regard to the two environmental issues where they differ, a strong message on environmental protection that both groups are in favour of (as shown in the cluster analysis) might be effective if it did not highlight the potentially divisive issues of car usage and the green belt. Failing this, a more targeted approach using direct marketing might be a viable way of talking to the two separately. With regard to their differences about how well the economy is doing, the positioning decision is straightforward. As the opposition party in the run-up to an election, a line critical of Conservative economic management would suit both segments. With regard to the EU issue, this was seen to be splitting the Conservative Party more than the Labour party at the time. The obvious positioning strategy for the two segments would therefore point out the opposition's divisions whilst keeping Labour's own options as open as possible. Finally, the lowering taxes versus improving other social services (such as pensions) would be an issue that the Labour party would wish to keep separate, reflecting as it does the old "tax and spend" Labour image. In hindsight, one of Labour's great successes in the 1997 election campaign was the fact that they managed to convince the electorate that they would manage the economy prudently, not put up taxes whilst prioritising the most needy in society - a major positioning achievement. Conclusions This research has segmented the British political market into seven distinctive groupings. The segments are understandable and offer an opportunity for each party to communicate their policies more effectively to the electorate. It is also possible to compare the segments from this research with those of a decade ago. A study using an earlier British Social Attitudes database and the same segmentation approach produced a six-cluster solution (Smith and Saunders, 1990). The fact that clear and understandable segments have been identified in both studies lends weight to their validity. Also it is possible to compare and contrast them to assess segment stability over time. Some differences are readily observable, as one might expect given the changes to the political landscape. Developments include a re-packaged and attractive opposition (New European Journal of Marketing 35,9/10 1070 Labour), a divided government (the Conservatives over the EU), greater cynicism towards politics generally, all set against a backdrop of reduced party loyalty (Crewe, 1993). So, for example there are two radical left wing segments (VI and VII) in this research compared to one a decade ago. However, there are signs of segment stability as well with the "Old Labour Poor" (Segment I) looking very similar to the "Poor Outsiders" segment of the previous research. Similarly, the "Tory Mainstreamers" of this study have similarities with the so-called "Hardcore Conservative Traditionalists" of a decade ago. The comparison therefore suggests change but with enduring segments. Certain generic influences such as traditional conservatism and poverty will produce segment stability over time. Other segments that are influenced by recent social and political factors (e.g. a new left-of-centre offering) may well change from election to election. The article has also considered the targeting process as applied to politics by focusing on two disparate Labour segments. There are, however, numerous other ways that might have been chosen for targeting segments. Obviously, Conservative and Liberal Democrat party managers would be more interested in targeting those segments where their support is strongest. Targeting "competitor" segments would also be possible and a potentially useful approach. For example, the Conservatives might be particularly interested in those Labour segments where anti-EU sentiment is strongest, with a view to converting voters. It has already been noted that the Liberal Democrats might be interested in those Labour segments (VI and VII) which show radical, libertarian values. In addition, targeting segments by their relative partisanship would be of interest to party managers of all persuasions. For example, targeting those segments with high levels of other parties' "residual identifiers" (i.e. weakly partisan voters) and finding out the issues that are most important to them might well produce an electoral advantage. Each of these targeting approaches presents, therefore, a new research avenue for marketing academics and practitioners. The successful use of the STP process to guide this research clearly shows that it is possible to follow a strategic approach to market segmentation in politics. Two segments were chosen deliberately because of their differences and it was still possible to develop a positioning strategy that would be viable for both, whilst not being contradictory. However, the danger remains that a "full blown" STP approach would push parties to be "all things to all people" resulting in them looking short-term and opportunistic in the eyes of the electorate. Accordingly, boundaries need to be set when using the STP approach within politics. The enduring ideology of the three main parties (socialism, conservatism and liberalism) offers the most likely safeguard against any excess. The fragmentation of messages across segments that positioning might encourage would be seen as lacking ideological consistency and credibility and thus be rejected by voters (Butler and Collins, 1994). Finally, it is worth noting that, although the research reported here focuses on party politics in Britain, the methodology can be applied in a wider political context. It will be relevant to any democratic process where two or more parties or directly elected individuals (as in the case of presidential/London Mayoral/ local elections) are competing for the support of an identifiable group of electors. We started this article by introducing the concept of a strategic approach to segmentation. Having identified the lack of any real sign of such a process either in practice or from a theoretical perspective we have suggested a methodology to proceed and provided a worked example for a major political party. Now, as ten years ago, it is clear that the pursuit of a competitive advantage is going to force an ever-increasingly sophisticated marketing approach in politics. Political parties will become better at positioning their parties towards identified targeted segments with attractive policies. In 2010 we may well be confirming the presence of the strategic era of marketing underpinned by strategic segmentation and asking the question, "Quo Vadis political marketing next?". References Baines, P.R. (1999), "Voter segmentation and candidate positioning", in Newman, B.I. (Ed.), Handbook of Political Marketing, Sage, Thousand Oaks, CA, pp. 73-86. Bottomley, C. and Curtice, J. (1999), "Is there a third way?", in Jowell, R. et al. (Ed.), British Social Attitudes. Who Shares New Labour Values?, Ashgate, Aldershot. Burke, E. (1774), "Speech to the electors of Bristol, November 3rd", in Hill, B.W. (1975), Edmund Burke on Government, Politics and Society, Harvester Press, Brighton, p. 157. Butler, P. and Collins, N. (1994), "Political marketing: structure and process", European Journal of Marketing, Vol. 28, pp. 19-34. Butler, P. and Collins, N. (1996), "Strategic analysis in political markets", European Journal of Marketing, Vol. 30 No. 10/11, pp. 32-44. Butler, P. and Collins, N. (1999), "A conceptual framework for political marketing", in Newman, B.I. (Ed.), Handbook off Political Marketing, Sage, Thousand Oaks, CA, pp. 55-72. Butler, D. and Stokes, D. (1971), Political Change in Britain, Macmillan, Basingstoke. Crewe, I. (1993), "Voting and the electorate", in Dunleavy, P. (Ed.), Developments in British Politics, Macmillan, Basingstoke. Curtice, J. (1996), "One nation again", in Jowell et al. (Eds), British Social Attitudes: The 13th Report, University Press, Cambridge, pp. 132-49. Devellis, R.F. (1991), Scale Development: Theory and Application, Sage Publications, Newbury Park, CA. Evans, G. (1998), "How Britain views the EU", in Jowell, R. et al. (Eds), British Social Attitudes, Ashgate, Aldershot, pp. 173-89. Green, P.E., Tull, D.S. and Albaum, G. (1988), Research for Marketing Decisions, 5th ed., Prentice-Hall, Englewood Cliffs, NJ. Hayes, B.C. and McAllister, I. (1996), "Marketing politics to voters: late deciders in the 1992 British election", European Journal of Marketing, Vol. 30 No. 10/11, pp. 135-46. Heath, A. and Park, A. (1997), "Thatcher's children", in Jowell et al. (Eds), British Social Attitudes: The 14th Report, Athenaeum Press, Gateshead, pp. 1-22. Heath, A. and Savage, M. (1994), "Middle class politics", in Jowell et al. (Eds), British Social A ttitudes: The 11th Report, University Press, Cambridge, pp. 132-49. Heath, A., Jowell, R. and Curtice, J. (1985), How Britain Votes, Pergamon Press, Oxford. European Journal of Marketing 35,9/10 1072 Hooley, GJ. and Saunders, J. (1993), Competitive Positioning. The Key to Market Success, Prentice-Hall, New York, NY. Jowell, R., Curtice, J., Park, A. and Thomson, K. (1999), British Social Attitudes: The 16th Report. Who Shares New Labour Values?, Ashgate, Aldershot. Kavanagh, D. (1995), Election Campaigning: The New Marketing of Politics, Blackwell, Oxford. Keith, R.J. (1960), "The marketing revolution", Journal of Marketing, January, pp. 35-8. Kohli, A. (1989), "Effects of supervisory behaviour: the role of individual differences among salespeople",/owrwfl/ ofMarketing, Vol. 53, pp. 40-50. Kotler, P. (1997), Marketing Management: Analysis, Planning, Implementation and Control, 9th ed., Prentice Hall International, Englewood Cliffs, NJ. Lock, A. and Harris, P. (1996), "Political marketing - vive la difference", European Journal of Marketing, Vol. 30 No. 10/11, pp. 21-31. Myers, J.H. (1996), Segmentation and Positioning for Strategic Marketing Decisions, American Marketing Association, Chicago, IL. Newman, B.I. (1994), The Marketing of the President: Political Marketing as Campaign Strategy, Sage, Thousand Oaks, CA. Newman, B.I. (1999), "A predictive model of voter behaviour: the repositioning of Bill Clinton", in Newman, B.I. (Ed.), Handbook of Political Marketing, Sage, Thousand Oaks, CA, pp. 259-82. Newman, B.I. and Sheth, J.N. (1985), "A model of primary voter behaviour", Journal of Consumer Research, Vol. 12, pp. 178-87. Nimmo, D. (1999), "The permanent campaign: marketing as a governing tool", in Newman, B.I. (Ed.), Handbook of Political Marketing, Sage, Thousand Oaks, CA, pp. 73-86. O'Cass, A. (1996), "Political marketing and the marketing concept", European Journal of Marketing, Vol. 30 No. 10/11, pp. 45-61. O'Gorman, T.W. and Woolson, R.F. (1991), "Variable selection to discriminate between groups: stepwise logistic regression or stepwise discriminant analysis?", The American Statistician, Vol. 3, pp. 187-93. Piercy, N.F. and Morgan, N.A. (1993), "Strategic and operational market segmentation", Journal of Strategic Marketing, Vol. 1, pp. 123-40. Porter, M. (1980), Competitive Strategy: Techniques for Analysing Industries and Competitors, The Free Press, New York, NY. Ries, A. and Trout, J. (1986), Positioning: The Battle for Your Mind, McGraw-Hill, New York, NY. Rosenbaum, M. (1997), From Soapbox to Soundbite: Party Political Campaigning Since 1945, Macmillan Press, London. Scott, J., Braun, M. and Alwin, D. (1998), "Partner, parent, workers: family and gender role", in Jowell, R. et al. (Eds), British Social Attitudes, Ashgate, Aldershot, pp. 19-37. Shelley, F.M. and Archer, J.C. (1994), "Some geographical aspects of the American presidential election of 1992", Political Geography, Vol. 13 No. 2, pp. 137-59. Smith, G. and Saunders, J. (1990), "The application of marketing to British politics", Journal of Marketing Management, Vol. 5 No. 3, pp. 295-306. Wind, Y. (1978), "Issues and advances in segmentation research'1'Journal of'Marketing Research, Vol. 15, August, pp. 317-37. Wring, D. (1997), "Reconciling marketing with political science: theories of political marketing", Journal of Marketing Management, Vol. 13, pp. 651-64. Yorke, D. and Meehan, SA. (1986), "ACORN in the political marketplace", European Journal of Marketing, Vol. 20 No. 8, pp. 73-86. Appendix. Definition of composite and derived variables 1. Party identification Respondents classify themselves as: Strategic political segmentation • supporters of that party (partisans); • closer to it than to others (sympathisers); or • more likely to support it in the event of a general election (residual identifiers). 2. Derived variables The following questions were used to develop these scales: (1) Left-right scale (Cronbach Alpha 0.82) • Government should redistribute income from the better off to those who are less well off. • Big business benefits owners at the expense of workers. • Ordinary working people do not get their fair share of the nation's wealth. • There is one law for the rich and one for the poor. • Management will always try to get the better of employees if it gets the chance. (2) Libertarian-authoritarian scale (Cronbach Alpha 0.74) • Young people today don't have enough respect for traditional British values. • People who break the law should be given stiffer sentences. • For some crimes, the death penalty is the most appropriate sentence. • Schools should teach children to obey authority. • The law should always be obeyed, even if a particular law is wrong. • Censorship of films and magazines is necessary to uphold moral standards.