Abstract
We run a benefit segmentation of 2017 insurance consumers in order to analyze the structure and heterogeneity of the German term life insurance market. The consumers’ preference information has been obtained through a choice-based conjoint (CBC) experiment and a subsequent hierarchical Bayes (HB) estimation routine. Drawing on their part-worth utility profiles, we first construct a diverse cluster ensemble, comprising a total of 1624 hierarchical and k-means solutions based on different linkage criterions and sensibly drawn starting points. Then, final group memberships are determined by means of consensus clustering. Our empirical results indicate that the market divides into three segments characterized by substantially different consumer types with distinct demands and needs. While the first group is clearly driven by the premium, the opposite holds true for the brand-loyal group. Additionally, the market is completed by a third segment with in-between preference structures. Hence, both brand insurers and companies with a lower reputation face consumer groups that almost perfectly fit their provider profiles. More specifically, by offering segment-oriented products, an efficient resource allocation is fostered and the basis for long-term business relationships is laid. This is becoming increasingly important, because ongoing regulatory efforts, low interest rates, and market entrances from InsuranceTech start-ups and tech giants aiming to utilize the market’s enormous hidden potential are changing the competitive environment significantly. A consequent alignment of important strategic decisions related to product innovations, pricing, and distribution channels to our identified consumer segments enables incumbents to maintain a stable and sustainable market share and profitability.
Similar content being viewed by others
Notes
As pointed out by Miles et al. (1978), adapting a decades-long and proven business model to a changing market environment is associated with substantial difficulties. This applies even more to the insurance sector with its strong regulations and high entry barriers for new market players, which renders most insurers primarily tradition-minded and rather reluctant regarding adaptation processes (Grossmann et al. 2004; KPMG 2014).
Most studies employ the methodological family of conjoint analyses (Green and Krieger 1991) in order to determine individual benefits.
A recent exception is the work of Farzanfar and Delafrooz (2016), who apply benefit clustering to derive customer lifetime values for various insurance lines. Benefit segmentations focusing on financial services in general are also scarce, although a few articles exist (Kinnaird et al. 1984; Harvey 1990; Cermak et al. 1994; Chang and Chen 1995; Machauer and Morgner 2001).
This is particularly the case if the company has information on individual preferences and is able to transform them into a meaningful profile of its representative customer (Weinstein 2002).
An overview of commercial conjoint applications focusing on Germany, Austria, and Switzerland is provided by Sattler and Hartmann (2008).
In their article, the authors laid CBC’s theoretical foundation by combining major aspects of discrete choice theory with the concept of conjoint analysis.
Despite its strengths, CBC is also associated with a few shortcomings. First, such an experiment is artificial by nature, implying that a hypothetical bias might exist (Cummings et al. 1995; Ding et al. 2005; Ding 2007). Second, each decision requires a high level of attention, which may decrease if the number of repeated choice tasks becomes too extensive.
Theoretical details of the CBC approach and the HB estimation routine can be found in Braun et al. (2016).
The respondents were entitled to participate in the study after stating that they either make decisions about insurance policies by themselves or are at least involved in their conclusion.
The health status was reflected by a respondent’s smoking behavior (nonsmoker vs. smoker).
We interpolate the utility values between attribute levels to display smooth lines. This state-of-the-art depiction of preference data is particularly useful in the case of continuous attributes, such as the insurance premium and term assured, respectively.
This was enforced during the estimation process by the so-called tying after estimation method (Johnson 2000).
Punj and Stewart (1983) provide a comprehensive overview on applications of cluster analyses in marketing research and highlight that most studies are targeted at market segmentations. Given its strengths, cluster analysis has been widely applied in various disciplines over the past decades other than marketing, such as archaeology (Sutton and Reinhard 1995), astronomy (Faúndez-Abans et al. 1996), bioinformatics and genetics (Selinski and Ickstadt 2008), psychiatry (Farmer et al. 1983), as well as weather classification (Littmann 2000).
Drawing on unstandardized raw utilities during the clustering would lead to distorted results since the absolute utility values from more consistent respondents compared to less consistent respondents are generally larger in size.
“The diameter of a cluster is just the largest dissimilarity between two of its objects.” (Kaufman and Rousseeuw 1990).
The inter-cluster distances are weighted according to the inverse of the number of persons in each cluster (Everitt et al. 2011).
Single linkage clustering (nearest-neighbor) has not been applied since it produces elongated instead of compact cluster solutions (Orme and Johnson 2008).
MacNaughton-Smith et al. (1964) proposed divisive hierarchical clustering as an improvement compared to agglomerative hierarchical methods since the risk of transferring “wrong” mergers from early to subsequent stages is mitigated.
In case this criterion applies to two persons, one of them is chosen randomly (Kaufman and Rousseeuw 1990).
Note that if all \(\xi _{j}\) are negative, the splinter group remains a singleton and the second divisive step starts.
A more detailed description of this divisive algorithm called DIANA can be found in Kaufman and Rousseeuw (1990).
All of the following procedures have been recommended and outlined by experts from Sawtooth Software who found good performance with regard to cluster recovery as well as solution reproducibility (Orme and Johnson 2008).
As highlighted by Retzer et al. (2009), these methods have been shown to lead to efficient solutions.
Orme and Johnson (2008) argue that comprehensive ensembles with many partitions and group solutions lead to good consensus solutions. Furthermore, these are more stable if the ensemble reflects a high diversity.
Please note: “Same cluster” refers to the classification of a pair within one partition only. More specifically, two respondents (a pair) are both in \(c_1\) in \(P_1\) and both in \(c_2\) in \(P_2\), for instance, where \(c_1\) does not have to be identical to \(c_2\).
The numbers \(\phi \), \(\chi \), \(\psi \), and \(\omega \) are most easily obtained by means of a \(2 \times 2\) contingency table of \(P_{1}\) and \(P_{2}\).
The necessary steps for assembling the indicator matrix to be employed as input of CC in the approach suggested by Orme and Johnson (2008) are contained in “Appendix 1”.
With 30 replications in the first CC step, a \(30 \times 30\) reproducibility matrix for each \(\kappa \) is obtained.
Relative attribute importances are ratio data and add up to 100% for each respondent.
Please note that the graphical interpretation of dendrograms should be taken with caution. Since our ensemble comprises solutions from 2 to 30 groups of each method, however, the likelihood of drawing wrong conclusions is minimized.
The F-values are obtained from an analysis of variance (ANOVA) on the individual-level part-worth utilities and indicate on which attributes the groups in the respective solution differ most. Since we compare solutions, however, solution-specific aggregated F-values have to be considered. The latter generally decrease with an increasing number of groups, which also holds true for the adjusted reproducibilities.
It is not common to remove outliers before clustering since the number of groups and the respective characteristics are unknown a priori. Hence, there is the risk of eliminating respondents that form a common group.
This step is needed since the individual-level part-worth profiles are subject to different arbitrary scale origins within each attribute (Orme 2010).
Please note that the respondents classified as outliers remain in the market. From an insurer’s point of view, however, it is preferable to focus on the segment kernel when developing and implementing product and pricing strategies.
Please note that we have not removed the outliers from our analysis since they are located within two standard deviations from the respective cluster averages.
Accordingly, these attributes are associated with the highest F-values in the ANOVA.
The setting has been adopted from Orme and Johnson (2008) with only slight variations.
The heatmap in Fig. 4b shows the diversity matrix of the reduced ensemble E with 112 partitions.
An exception is the reduced ensemble F.
To facilitate a comparison, we draw on the segment memberships obtained from the full ensemble with 1000 persons for the following more detailed segment analysis.
Please note that running a cluster analysis on a subset of respondents should be handled with caution. Even though it has been drawn randomly, with equal probability for each respondent to be included, it might be possible to inadvertently ignore an entire group. This, in turn, would then result in a completely different preference distribution.
References
Aaker, D. A. (2013). Strategic marketing management (10th ed.). New York, NY: Wiley.
Akaah, I. P. (1988). Cluster analysis versus Q-type factor analysis as a disaggregation method in hybrid conjoint modeling: An empirical investigation. Journal of the Academy of Marketing Science, 16(2), 11–18.
Andreasen, A. R. (1966). Geographic mobility and market segmentation. Journal of Marketing Research, 3(4), 341–348.
Auty, S. (1992). Consumer choice and segmentation in the restaurant industry. Service Industries Journal, 12(3), 324–339.
Bahn, K. D., & Granzin, K. L. (1985). Benefit segmentation in the restaurant industry. Journal of the Academy of Marketing Science, 13(3), 226–247.
Bass, F. M., Tigert, D. J., & Lonsdale, R. T. (1968). Market segmentation: Group versus individual behavior. Journal of Marketing Research, 5(3), 264–270.
Braun, A., Schmeiser, H., & Schreiber, F. (2016). On consumer preferences and the willingness to pay for term life insurance. European Journal of Operational Research, 253(3), 761–776.
Brouhns, N., Guillén, M., Denuit, M., & Pinquet, J. (2003). Bonus–Malus scales in segmented tariffs with stochastic migration between segments. Journal of Risk and Insurance, 70(4), 577–599.
Brown, J. D. (1992). Benefit segmentation of the fitness market. Health Marketing Quarterly, 9(3–4), 19–28.
CapGemini. (2016). Top 10 trends in insurance in 2016.
Carrillat, F., Jaramillo, F., & Locander, W. (2004). Market-driving organizations: A framework. Academy of Marketing Science Review, 5(1), 1–14.
Cattin, P., & Wittink, D. R. (1982). Commercial use of conjoint analysis: A survey. Journal of Marketing, 46(3), 44–53.
Cermak, D. S. P., File, K. M., & Prince, R. A. (1994). A benefit segmentation of the major donor market. Journal of Business Research, 29(2), 121–130.
Chang, T.-Z., & Chen, S.-J. (1995). Benefit segmentation: A useful tool for financial investment services. Journal of Professional Services Marketing, 12(2), 69–80.
Cormack, R. M. (1971). A review of classification. Journal of the Royal Statistical Society. Series A (General), 134(3), 321–367.
Cummings, R. G., Harrison, G. W., & Rutström, E. E. (1995). Homegrown values and hypothetical surveys: Is the dichotomous choice approach incentive-compatible? American Economic Review, 85(1), 260–266.
Currim, I. S. (1981). Using segmentation approaches for better prediction and understanding from consumer mode choice models. Journal of Marketing Research, 18(3), 301–309.
Dahlby, B. (1983). Adverse selection and statistical discrimination: An analysis of canadian automobile insurance. Journal of Public Economics, 20(1), 121–130.
De Kluyver, C. A., & Whitlark, D. B. (1986). Benefit segmentation for industrial products. Industrial Marketing Management, 15(4), 273–286.
DeSarbo, W. S., & DeSarbo, C. F. (2007). A generalized normative segmentation methodology employing conjoint analysis. In A. Gustafsson, A. Herrmann & F. Huber (Eds.), Conjoint measurement—Methods and applications (4th ed, pp. 321–345). Springer, Berlin.
DeSarbo, W. S., Oliver, R. L., & Rangaswamy, A. (1989). A simulated annealing methodology for clusterwise linear regression. Psychometrika, 54(4), 707–736.
DeSarbo, W. S., Ramaswamy, V., & Cohen, S. H. (1995). Market segmentation with choice-based conjoint analysis. Marketing Letters, 6(2), 137–147.
DeSarbo, W. S., Wedel, M., Vriens, M., & Ramaswamy, V. (1992). Latent class metric conjoint analysis. Marketing Letters, 3(3), 273–288.
Ding, M. (2007). An incentive-aligned mechanism for conjoint analysis. Journal of Marketing Research, 44(2), 214–223.
Ding, M., Grewal, R., & Liechty, J. (2005). Incentive-aligned conjoint analysis. Journal of Marketing Research, 42(1), 67–82.
Doyle, P. (1995). Marketing in the new millennium. European Journal of Marketing, 29(13), 23–41.
Epetimehin, F. M. (2011). Market segmentation: A tool for improving customer satisfaction and retention in insurance service delivery. Journal of Emerging Trends in Economics and Management Sciences, 2(1), 62–67.
Ester, M., Kriegel, H.-P., Sander, J., & Xu, X. (1996). A density-based algorithm for discovering clusters in large spatial databases with noise. In Proceedings of 2nd international conference on Knowledge Discovery and Data Mining (KDD-96) (pp. 226–231). Menio Park, CA: The AAAI Press.
Everitt, B. S., Landau, S., Leese, M., & Stahl, D. (2011). Cluster analysis (5th ed.). Chichester, UK: Wiley.
Farmer, A. E., McGuffin, P., & Spitznagel, E. L. (1983). Heterogeneity in schizophrenia: A cluster-analytic approach. Psychiatry Research, 8(1), 1–12.
Farzanfar, E., & Delafrooz, N. (2016). Determining the customer lifetime value based on the benefit clustering in the insurance industry. Indian Journal of Science and Technology, 9(1), 1–8.
Faúndez-Abans, M., Ormeño, M. I., & de Oliveira-Abans, M. (1996). Classification of planetary nebulae by cluster analysis and artificial neural networks. Astronomy and Astrophysics Supplement Series, 116(2), 395–402.
Finn, D. W., & Lamb, C. W, Jr. (1986). Hospital benefit segmentation. Journal of Health Care Marketing, 6(4), 26–33.
Frank, R. E. (1967). Correlates of buying behavior for grocery products. Journal of Marketing, 31(4), 48–53.
Frank, R. E., Massy, W. F., & Wind, Y. (1972). Market segmentation (Prentice-Hall international series in management). Englewood Cliffs, NJ: Prentice-Hall.
Gordon, A. D. (1987). A review of hierarchical classification. Journal of the Royal Statistical Society. Series A (General), 150(2), 119–137.
Gower, J. C., & Legendre, P. (1986). Metric and Euclidean properties of dissimilarity coefficients. Journal of Classification, 3(1), 5–48.
Green, P. E. (1977). A new approach to market segmentation. Business Horizons, 20(1), 61–73.
Green, P. E., & DeSarbo, W. S. (1979). Componential segmentation in the analysis of consumer trade-offs. Journal of Marketing, 43(4), 83–91.
Green, P. E., & Krieger, A. M. (1991). Segmenting markets with conjoint analysis. Journal of Marketing, 55(4), 20–31.
Green, P. E., Krieger, A. M., & Wind, Y. J. (2001). Thirty years of conjoint analysis: Reflections and prospects. Interfaces, 31(3), 56–73.
Green, P. E., & Rao, V. R. (1971). Conjoint measurement for quantifying judgmental data. Journal of Marketing Research, 8(3), 355–363.
Green, P. E., & Srinivasan, V. (1978). Conjoint analysis in consumer research: Issues and outlook. Journal of Consumer Research, 5(2), 103–123.
Green, P. E., & Srinivasan, V. (1990). Conjoint analysis in marketing: New developments with implications for research and practice. Journal of Marketing, 54(4), 3–19.
Grossmann, M., McCarthy, R. V., & Aronson, J. E. (2004). E-commerce adoption in the insurance industry. Issues in Information Systems, 5(2), 467–473.
Hagerty, M. R. (1985). Improving the predictive power of conjoint analysis: The use of factor analysis and cluster analysis. Journal of Marketing Research, 22(2), 168–184.
Haley, R. I. (1968). Benefit segmentation: A decision-oriented research tool. Journal of Marketing, 32(3), 30–35.
Haley, R. I. (1985). Developing effective communications strategy: A benefit segmentation approach. Wiley series on marketing management (1st ed.). New York, NY: Wiley.
Harvey, J. W. (1990). Benefit segmentation for fund raisers. Journal of the Academy of Marketing Science, 18(1), 77–86.
Honkanen, P., Olsen, S. O., & Myrland, Ø. (2004). Preference-based segmentation: A study of meal preferences among norwegian teenagers. Journal of Consumer Behaviour, 3(3), 235–250.
Howe, N., & Strauss, W. (2000). Millennials rising: The next great generation. New York, NY: Vintage Books.
Huber, J. (1997). What we have learned from 20 years of conjoint research: When to use self-explicated, graded pairs, full profiles, or choice experiments. Sawtooth Software Research Paper.
Hubert, L., & Arabie, P. (1985). Comparing partitions. Journal of Classification, 2(1), 193–218.
Jain, A. K. (2010). Data clustering: 50 years beyond \(k\)-means. Pattern Recognition Letters, 31(8), 651–666.
Jang, S. C., Morrison, A. M., & O’Leary, J. T. (2002). Benefit segmentation of Japanese pleasure travelers to the USA and Canada: Selecting target markets based on the profitability and risk of individual market segments. Tourism Management, 23(4), 367–378.
John, J., & Miaoulis, G. (1992). A model for understanding benefit segmentation in preventive health care. Health Care Management Review, 17(2), 21–32.
Johnson, L. W., Ringham, L., & Jurd, K. (1991). Behavioural segmentation in the australian wine market using conjoint choice analysis. International Marketing Review, 8(4), 26–31.
Johnson, R. (2000). Monotonicity constraints in choice-based conjoint with hierarchical Bayes. Sawtooth Software Research Paper.
Kamakura, W. A. (1988). A least squares procedure for benefit segmentation with conjoint experiments. Journal of Marketing Research, 25(2), 157–167.
Kaufman, L., & Rousseeuw, P. J. (1990). Finding groups in data: An introduction to cluster analysis (1st ed.). Hoboken, NJ: Wiley.
Kinnaird, D., Shaughnessy, K., Struman, K., & Swinyard, W. (1984). Market segmentation of retail bank services: A model for management. Journal of Retail Banking, 6(4), 52–63.
KPMG (2014). Transforming insurance: Securing competitive advantage.
Lance, G. N., & Williams, W. T. (1967). A general theory of classificatory sorting strategies: 1. Hierarchical systems. The Computer Journal, 9(4), 373–380.
Littmann, T. (2000). An empirical classification of weather types in the mediterranean basin and their interrelation with rainfall. Theoretical and Applied Climatology, 66(3–4), 161–171.
Loker, L. E., & Perdue, R. R. (1992). A benefit-based segmentation of a nonresident summer travel market. Journal of Travel Research, 31(1), 30–35.
Louviere, J. J., & Woodworth, G. (1983). Design and analysis of simulated consumer choice or allocation experiments: An approach based on aggregate data. Journal of Marketing Research, 20(4), 350–367.
Luce, R. D., & Tukey, J. W. (1964). Simultaneous conjoint measurement: A new type of fundamental measurement. Journal of Mathematical Psychology, 1(1), 1–27.
Machauer, A., & Morgner, S. (2001). Segmentation of bank customers by expected benefits and attitudes. International Journal of Bank Marketing, 19(1), 6–18.
MacNaughton-Smith, P., Williams, W. T., Dale, M. B., & Mockett, L. G. (1964). Dissimilarity analysis: A new technique of hierarchical sub-division. Nature, 202(1), 1034–1035.
MacQueen, J. (1967). Some methods for classification and analysis of multivariate observations. In L. LeCam & J. Neymen (Eds.), Proceedings of the fifth Berkeley symposium on mathematical statistics and probability (Vol. 1, pp. 281–297). Berkeley, CA: University of California Press.
Mao, J., & Jain, A. K. (1996). A self-organizing network for hyper-ellipsoidal clustering (HEC). IEEE Transactions on Neural Networks, 7(1), 16–29.
Mariorty, R. T., & Reibstein, D. J. (1986). Benefit segmentation in industrial markets. Journal of Business Research, 14(6), 463–486.
McQuitty, L. L. (1966). Similarity analysis by reciprocal pairs for discrete and continuous data. Educational and Psychological Measurement, 26(4), 825–831.
Miles, R. E., Snow, C. C., Meyer, A. D., & Coleman, H. J, Jr. (1978). Organizational strategy, structure, and process. The Academy of Management Review, 3(3), 546–562.
Milligan, G. W. (1980). An examination of the effect of six types of error perturbation on fifteen clustering algorithms. Psychometrika, 45(3), 325–342.
Milligan, G. W., & Cooper, M. C. (1987). Methodology review: Clustering methods. Applied Psychological Measurement, 11(4), 329–354.
Myers, J. H. (1976). Benefit structure analysis: A new tool for product planning. Journal of Marketing, 40(4), 23–32.
Ogawa, K. (1987). An approach to simultaneous estimation and segmentation in conjoint analysis. Marketing Science, 6(1), 66–81.
Orme, B. K. (2010). Getting started with conjoint analysis: Strategies for product design and pricing research (2nd ed.). Madison, WI: Research Publishers LLC.
Orme, B. K., & Johnson, R. (2008). Improving \(k\)-means cluster analysis: Ensemble analysis instead of highest reproducibility replicates. Sawtooth Software Research Paper.
Punj, G., & Stewart, D. W. (1983). Cluster analysis in marketing research: Review and suggestions for application. Journal of Marketing Research, 20(2), 134–148.
PWC. (2016). Opportunities await: How InsurTech is reshaping insurance.
Rand, W. M. (1971). Objective criteria for the evaluation of clustering methods. Journal of the American Statistical Association, 66(336), 846–850.
Retzer, J., Alberg, S., & Yuan, J. (2009). Achieving consensus in cluster ensemble analysis. Sawtooth Software Research Paper.
Retzer, J., & Shan, M. (2007). Cluster ensemble analysis and graphical depiction of cluster partitions. In Proceedings of the 2007 Sawtooth software conference, Sequim WA.
Sattler, H., & Hartmann, A. (2008). Commercial use of conjoint analysis. In M. Höck & K.-I. Voigt (Eds.), Operations management in theorie und Praxis (1st ed., pp. 103–119). Wiesbaden: Gabler Edition Wissenschaft.
Sawtooth Software. (2013). CCEA v3. Software for Convergent Cluster and Ensemble Analysis Manual.
Selinski, S., & Ickstadt, K. (2008). Cluster analysis of genetic and epidemiological data in molecular epidemiology. Journal of Toxicology and Environmental Health, 71(11–12), 835–844.
Seog, S. H. (2009). Insurance markets with differential information. Journal of Risk and Insurance, 76(2), 279–294.
Shoemaker, S. (1994). Segmenting the U.S. travel market according to benefits realized. Journal of Travel Research, 32(3), 8–21.
Smith, W. R. (1956). Product differentiation and market segmentation as alternative marketing strategies. Journal of Marketing, 21(1), 3–8.
Sokal, R. R., & Michener, C. D. (1958). A statistical method for evaluating systematic relationships. University of Kansas Science Bulletin, 38(2), 1409–1438.
Sørensen, T. (1948). A method of establishing groups of equal amplitude in plant sociology based on similarity of species content and its application to analyses of the vegetation on Danish commons. Biologiske Skrifter, 5(4), 1–34.
Speed, R., & Smith, G. (1992). Retail financial services segmentation. Service Industries Journal, 12(3), 368–383.
Steenkamp, J.-B. E. M., & Wedel, M. (1993). Fuzzy clusterwise regression in benefit segmentation: Application and investigation into its validity. Journal of Business Research, 26(3), 237–249.
Steinley, D. (2004). Properties of the Hubert–Arabie adjusted rand index. Psychological Methods, 9(3), 386–396.
Strehl, A., & Ghosh, J. (2002). Cluster ensembles—A knowledge reuse framework for combining multiple partitions. The Journal of Machine Learning Research, 3(3), 583–617.
Sutton, M. Q., & Reinhard, K. J. (1995). Cluster analysis of the coprolites from Antelope House: Implications for Anasazi diet and cuisine. Journal of Archaeological Science, 22(6), 741–750.
Swiss Re. (2013). Life insurance: Focusing on the consumer. Swiss Re Sigma 6/2013.
Swiss Re. (2014). Term life insurance in Germany: The consumers’ perspective—A need for preferences-orientated product design? Swiss Re Economic Research and Consulting Report.
ter Hofstede, F., Steenkamp, J.-B. E. M., & Wedel, M. (1999). International market segmentation based on consumer-product relations. Journal of Marketing Research, 36(1), 1–17.
Thomas, A., & Pickering, G. (2003). Behavioural segmentation: A New Zealand wine market application. Journal of Wine Research, 14(2–3), 127–138.
Twedt, D. W. (1964). Some practical applications of the ‘Heavy Half’ theory. In Advertising research foundation 10th annual conference, New York, NY.
Vriens, M., Wedel, M., & Wilms, T. (1996). Metric conjoint segmentation methods: A Monte Carlo comparison. Journal of Marketing Research, 33(1), 73–85.
Ward, J. H. (1963). Hierarchical grouping to optimize an objective function. Journal of the American Statistical Association, 58(301), 236–244.
Wedel, M., & Kamakura, W. (2000). Market segmentation: Conceptual and methodological foundations (2nd ed.). Norwell, MA: Kluwer Academic Publishers.
Wedel, M., & Kistemaker, C. (1989). Consumer benefit segmentation using clusterwise linear regression. International Journal of Research in Marketing, 6(1), 45–59.
Wedel, M., & Steenkamp, J.-B. E. M. (1991). A clusterwise regression method for simultaneous fuzzy market structuring and benefit segmentation. Journal of Marketing Research, 28(4), 385–396.
Weinstein, A. (2002). Customer retention: A usage segmentation and customer value approach. Journal of Targeting, Measurement and Analysis for Marketing, 10(3), 259–268.
Wind, Y. (1978). Issues and advances in segmentation research. Journal of Marketing Research, 15(3), 317–337.
Wittink, D. R., & Cattin, P. (1989). Commercial use of conjoint analysis: An update. Journal of Marketing, 53(3), 91–96.
Wittink, D. R., Vriens, M., & Burhenne, W. (1994). Commercial use of conjoint analysis in Europe: Results and critical reflections. International Journal of Research in Marketing, 11(1), 41–52.
Woodside, A. G., & Jacobs, L. W. (1985). Step two in benefit segmentation: Learning the benefits realized by major travel markets. Journal of Travel Research, 24(1), 7–13.
Yankelovich, D. (1964). New criteria for market segmentation. Harvard Business Review, 42(2), 83–90.
Yankelovich, D., & Meer, D. (2006). Rediscovering market segmentation. Harvard Business Review, 84(2), 122–131.
Author information
Authors and Affiliations
Corresponding author
Appendices
Appendix 1: CBC experiment and consensus clustering
1.1 Levels for the attribute insurance premium
1.2 Illustration of consensus method
The approach applied in this paper goes back to the work of Strehl and Ghosh (2002). Later, it has then been adapted by Orme and Johnson (2008), who suggest to follow the subsequent steps: First, one needs to derive multiple solutions from various clustering algorithms and techniques in order to generate a diverse cluster ensemble. In the second step, for each ensemble partition \(P_i\) (\(i=1,\ldots ,M\)) and individual j (\(j=1,\ldots ,n\)), the assigned group labels (e.g. 1, 2 for a two-group solution) are transformed into dummy coding such that every group represents a single column in the dummy matrix. Generally, “1” indicates “Group member”, whereas “0” stands for “No group member”. Then, the M dummy matrices are used to construct the so-called indicator matrix. In contrast to Strehl and Ghosh (2002), who employ a so-called graph partitioning approach at this stage to achieve correspondence between the labels of the individual solutions, Orme and Johnson (2008) refrain from this step. Instead, they suggest to take the columns of the indicator matrix as new clustering variables and directly subject them to the consensus analysis, which can be regarded as clustering on cluster solutions.
This short example illustrates how the input matrix for the consensus stage is generated.Footnote 42 Let us assume that we have a total of five respondents j (\(n = 5\)) and four ensemble partitions \(P_i\) (\(i=1,\ldots ,4\)) as shown in Table 7.
It can be seen that partitions \(P_{2}\) and \(P_{4}\) are two-group solutions with inverse labeling, while \(P_{1}\) is a three-group solution, and \(P_{3}\) a four-group solution. The dummy coding of partition \(P_{1}\) results in the following three-column matrix as shown in Table 8.
By converting all M ensemble partitions into the form of Table 8, the so-called indicator matrix is obtained (see Table 9), which is then subjected to the final consensus clustering.
Appendix 2: Sensitivity analyses
According to Retzer and Shan (2007), two critical aspects in cluster analysis exist: deciding on the appropriate number of clusters and differentiating between their quality. In order to demonstrate that the results presented in this paper are robust and meaningful, we now run several sensitivity analyses on our selected number of segments and our selected ensemble. Additionally, we further show that the results remain stable if the sample is randomly reduced to 1000 persons.
1.1 Number of market segments
We construct six reduced ensembles based on different selections of clustering techniques and numbers of groups, which are then subjected to the consensus approach. The reduced ensemble A consists of 784 partitions with 2- to 15-group solutions derived by our eight clustering strategies on all seven clustering bases (\(7\,\times \,8\,\times \,14 = 784\)), while ensemble B has 504 partitions with 2- to 10-group solutions (\(7\,\times \,8\,\times \,9 = 504\)). Ensemble C, in turn, consists of 609 pure k-means strategies (\(7 \times 3 \times 29 = 609\)) that cover 2- to 30-group solutions. Its counterpart ensemble D, in contrast, is derived by the five hierarchical methods (\(7 \times 5 \times 29 = 1015\)). Finally, ensembles E (\(1 \times 8 \times 14 = 112\)) and F (\(1 \times 8 \times 9 = 72\)) exclusively resort to the combination of all six attributes.Footnote 43 As becomes apparent from the last row, diversity generally increases with the number of partitions.
The adjusted reproducibility rates in Table 10 provide evidence that the three-cluster solution is first-best for at least two reasons. First, although reproducibility generally falls with an increasing number of segments, the reverse holds true for three segments, independent of the selected ensemble. Second, it reaches the highest absolute values at almost equal to 100.0%. As pointed out by Orme and Johnson (2008), such high reproducibility rates clearly indicate that a reasonable structure has been found. Additionally, it can be seen that the second-best solution is a division into two segments, with an average adjusted reproducibility of 91.8%.
In order to further illustrate that three segments should be chosen, Table 11 shows its cross-tabulation with the two-group solution. As can be seen, the latter consists of two almost equally sized segments with 776 and 734 respondents. More specifically, it assigns all price-sensitive consumers to its first cluster (upper left cell), while all brand-loyal respondents are allocated to the second segment (cell at bottom center). However, since the algorithm was forced to form two groups, the waverers are distributed to both clusters. That is, 347 to \(C_1\) and 211 to \(C_2\). As a consequence, the preference distributions in the groups are diluted and less clear-cut than in our three-segment solution. In the first cluster, for instance, the average relative importance of the premium amounts to 59.5%, which equals a decline of 8.3 percentage points (p.p.). The importance of the attribute brand in the brand-loyal group, in turn, decreases by 2.8 p.p. to approximately 24.3% on average. At the same time, it increases by 5.5 p.p. to 23.9% for the insurance premium. Thus, the two-group solution ignores an important market segment that is substantially different from the price-sensitive and the brand-loyal groups.
1.2 Selection of the ensemble
In the next step, we analyze how individual group membership assignments depend on the ensemble composition. To do so, we take all reduced ensembles from Table 10 as clustering bases and examine their three-segment solutions by calculating the corresponding average relative attribute importances. Table 12 shows all results.Footnote 44
The upper part of the table contains both the importances for the full sample and our final ensemble with 1624 partitions from Table 5. At first glance, it becomes evident that the price-sensitive, the brand-loyal, and the waverer groups are obtained under all ensembles. Moreover, the absolute relative attribute importances are almost constant for all three segments and similar to those of the full ensemble. For the price-sensitive and waverer segments, the importance ranking of the six attributes remains unchanged, while some minor deviations can be observed for the brand-loyal consumers. Regarding the segment sizes, the price-sensitives always account for the lowest market share, amounting to approximately 20.0% to 22.0%.Footnote 45 The brand-loyal market share, on the other hand, varies between 26.0% and 33.0%, while the waverers are associated with market shares between 21.0% and 29.0%. As can be seen from the last column, approximately 25.0% of the respondents are classified as outliers under all ensembles. Particularly when comparing the reduced ensembles E and F with our full ensemble, however, some differences in both the relative attribute importances and market shares emerge. This can be explained by the fact that the former are based on the combination of all six attributes only, while the single attributes have not been considered.Footnote 46 To sum up, this sensitivity analysis is a further indication of the robustness of our approach.
1.3 Number of respondents
Finally, we construct a subsample of 1000 respondents at random and reconstruct selected ensembles from Table 10 (Full, A, C, D), for which we then calculate adjusted reproducibilities, relative attribute importances, as well as market shares. Similar to our previous analyses, Fig. 7 shows that the adjusted reproducibilities for the three-cluster solutions exhibit an atypical trend. That is, for all ensembles, it increases compared to the respective two-group solution. In contrast to the full sample with 2017 persons, however, our full ensemble achieves a slightly lower reproducibility figure (96.3%), while the reduced ensemble A with a total of 784 partitions is associated with the maximum of 99.4%.Footnote 47
Table 13 shows that we were able to reproduce the preference patterns from the full sample.Footnote 48 As can be seen from the first row, both the absolute relative attribute importances and attribute rankings are almost identical (see Table 5). While the premium is by far the most dominant product characteristic, the remaining five attributes are of minor importance. Regarding the segments, the first group (\(\mathrm {C_1}\)) accounts for an equal market share (22.0% vs. 21.3%) and exhibits a similar price sensitivity. Within the second cluster (\(\mathrm {C_2}\)), the average relative attribute importance for the brand (26.9%) corresponds to the full sample analysis, which also holds true for its market share. Moreover, the importance distribution of the product attributes remains unchanged, with the insurance premium being ranked third. The members of the waverer group, in turn, have an average preference structure that lies between the price-sensitive and brand-loyal customers. A comparison of rows one and four shows that these respondents almost resemble the one-group solution. Finally, regarding the aggregated market share, we also identified 25.0% of the respondents as outliers.
Rights and permissions
About this article
Cite this article
Schreiber, F. Identification of customer groups in the German term life market: a benefit segmentation. Ann Oper Res 254, 365–399 (2017). https://doi.org/10.1007/s10479-017-2446-y
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10479-017-2446-y