Multichannel segmentation in the after-sales stage in the insurance industry

Abstract

Purpose The purpose of this paper is to present a multichannel segmentation approach to identifying customer segments based on actual customer channel usage in the post-purchase phase in the health insurance industry. Design/methodology/approach A multinomial regression model and count regression models were estimated to describe the profiles of customer segments and the frequency of channel usage based on generations and sociodemographic variables. Findings This study identified generational differences in channel usage. Single female customers from the Pre-Boomer or Baby Boomer generation and customers living in states with lower incomes are more likely to use call centres. Website users tend to live in regions with higher per capita income. Multichannel users are, on average, more frequent users of both the website and call centres. In terms of sociodemographics, they display a more heterogeneous profile. Research limitations/implications The proposed segmentation needs to be enriched with additional variables such as customers’ health status or channel usage motivations. Practical implications Customers, who are male, married and from Generations Y and X, are more likely to use the website. Their propensity to switch to a digital channel could be investigated further to develop targeted migration strategies. Multichannel users are, on average, more frequent users of all channels. To avoid increased channel costs, segments should be analysed in terms of their size and profit potential to help allocate marketing investment more efficiently. Originality/value As opposed to existing research, the proposed segmentation approach is based on transactional data of channel usage from a real company, combined with analyses using generations and sociodemographic variables.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 14 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.