Student or educator? Sign up with your school email and get 6 months of Pro free.
CoChat

The DOI matches its Crossref record; 55 of the 57 references it cites were found in a registry or an index; none of the 19 sources its claims cite is checked for support yet.

Modeling CLV: A test of competing models in the insurance industry

Open access
Donkers, Bas, Verhoef, Peter C., de Jong, Martijn G. · 2007 · Quantitative Marketing and Economics

Abstract

Customer Lifetime Value (CLV) is one of the key metrics in marketing and is considered an important segmentation base. This paper studies the capabilities of a range of models to predict CLV in the insurance industry. The simplest models can be constructed at the customer relationship level, i.e. aggregated across all services. The more complex models focus on the individual services, paying explicit attention to cross buying, but also retention. The models build on a plethora of approaches used in the existing literature and include a status quo model, a Tobit II model, univariate and multivariate choice models, and duration models. For all models, CLV for each customer is computed for a four-year time horizon. We find that the simple models perform well. The more complex models are expected to better capture the richness of relationship development. Surprisingly, this does not lead to substantially better CLV predictions.

Overview

CoChat Summary

TL;DR

This paper compares a range of relationship-level and service-level models for predicting customer lifetime value (CLV) in the Dutch insurance industry. It finds that simple models perform well, but models focusing only on retention are insufficient because cross-buying must also be accounted for.

Key findings

  • For individual CLV levels, no single model dominates: the status quo model has the lowest MAE (19.4% of average CLV), while the multivariate Probit model has the lowest RMSE (40.2%); the Tobit II model has a high level hit rate (78.3%).
  • For ordering customers by CLV, the status quo, profit regression, aggregate retention, monetary-value retention, and Tobit II models all achieve the same hit rate (78.4%), so the simple status quo model may be favored for simplicity.
  • For valuing the total customer base, all models undervalue it; the profit regression model performs best (−1.2% deviation), followed by the independent choice model (−4.9%), while retention-only models are about 12% below actual value and service-level duration models perform worst (−21.5% and −22.3%).
  • About 12% of the company’s total CLV comes from services customers did not buy when predictions were made, showing that ignoring cross-buying substantially harms CLV prediction.

Methods

  • The application uses 6 years of behavioral data for 30,000 customers of a large Dutch direct writer; models are estimated on the first year for customers active on January 1, 1998, and CLV is predicted for customers active on January 1, 1999 over a 4-year horizon (2000–2003) with a 10% annual discount rate. Descriptive statistics report n=24,055 for customers active on 1/1/1998.
  • The paper estimates relationship-level models (status quo, profit regression, aggregate/segmented retention, Probit, bagging, duration, Tobit II) and service-level models (independent and multivariate Probit and duration models), using a factor structure for dependencies and Markov chains for multi-period CLV prediction.
  • Predictive performance is assessed with MAE and RMSE as percentages of average CLV, hit rates for CLV levels and ordering, and percentage deviation for total customer-base valuation.

Limitations

  • The paper states that results are based on a single case study in a relatively stable industry and that future research should replicate the study in different industries; it also notes that margins are not customer-specific because the company cannot accurately estimate individual-level risk.
  • The paper states that unobserved heterogeneity is not allowed in relationship-level models due to identification problems, and that service-level duration models use less information and perform poorly.
  • Evident from the text: CLV is predicted over a finite 4-year horizon rather than an infinite one, and the good performance of the status quo model may not generalize because in this application defection and cross-buying effects are similar in size, making the net effect small.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1007/s11129-006-9016-y

Cites

55 of 55 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Verification

Claims the paper makes

0 Valid · 1 Not valid · 18 Not checked · 89 Own results and interpretations

63 more claims

Create a free account to see all 108 claims and every source verdict.

See all claims — free account
  • It may state two separate claims as one. 18
  • Probably not a substantial claim. 24

Create a free account to see these claims.

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

1 of 1

Citations

Cites

Show all 57 references

Cited by

See all 142

Metadata

Authors
Donkers, B., Verhoef, P. C., & de Jong, M. G.
Affiliations
  • Erasmus University Rotterdam (NL)
  • University of Groningen (NL)
Corresponding author
Donkers, Bas
Venue
Quantitative Marketing and Economics
Published
2007-06
Volume
5
Issue
2
Pages
163-190
ISSN
  • 1570-7156 (ISSN-L)
  • 1573-711X
OpenAlex
W2066061252
Cited by
142
References
57
Citations by year
2026: 2 · 2025: 5 · 2024: 6 · 2023: 4 · 2022: 5 · 2021: 4 · 2020: 5 · 2019: 8 · 2018: 6 · 2017: 13 · 2016: 8 · 2015: 6 · 2014: 15 · 2013: 10 · 2012: 6
FWCI
27.17
Type
journal article
Language
en
Open access
Licence
Locations
  • Quantitative Marketing and Economics (publishedVersion, cc-by-nc)
  • University of Groningen research database (University of Groningen / Centre for Information Technology) (publishedVersion, cc-by-nc)
  • EUR Research Repository (Erasmus University Rotterdam) (submittedVersion)
  • RePEc: Research Papers in Economics (submittedVersion)
  • University of Groningen research database (University of Groningen / Centre for Information Technology) (submittedVersion, cc-by-nc)
Funders
  • Rijksuniversiteit Groningen
  • Westfälische Wilhelms-Universität Münster
Topics
  • Social Sciences › Business, Management and Accounting › Marketing › Customer churn and segmentation
  • Social Sciences › Business, Management and Accounting › Marketing › Customer Service Quality and Loyalty
  • Social Sciences › Business, Management and Accounting › Marketing › Consumer Market Behavior and Pricing
Keywords
customer lifetime value, insurance industry, customer relationship management, cross-selling, customer retention, choice models, duration models, Tobit model, model comparison, customer segmentation

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

A language model drew the claims from the paper on 2026-10-11, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-10-11 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Modeling CLV: A test of competing models in the insurance industry. https://app.cochat.ai/papers/10.1007/s11129-006-9016-y
The paper
Donkers, B., Verhoef, P. C., & de Jong, M. G. (2007). Modeling CLV: A test of competing models in the insurance industry. Quantitative Marketing and Economics. https://doi.org/10.1007/s11129-006-9016-y

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.