Segmenting donors to improve the fundraising efficiency of a humanitarian NGO
No Thumbnail Available
URL
Journal Title
Journal ISSN
Volume Title
School of Business |
Master's thesis
Authors
Date
2022
Department
Major/Subject
Mcode
Degree programme
Information and Service Management (ISM)
Language
en
Pages
64
Series
Abstract
While humanitarian crises are on the rise, NGOs are facing increasing competition for the time, money and effort of people. In the era of personalization, it is growing fundamental to know the customer in order to retain the customer, which can be extrapolated to donors in the context of the third sector. However, the third sector has been vastly neglected in terms of customer analytics, which has led many organizations to operate on gut feelings and traditional ways of working. This thesis approaches donor understanding from a practical point of view: the one-off donors of a non-governmental organization that raises funds in the Finnish market are segmented as an empirical case study. The study aims to find an effective and meaningful partition between the donors, as well as explain the composition of the segments through analyzing their similarities and differences. The segmentation is evaluated through ease of interpretation and application, which leads this thesis to also focus on ways to incorporate the findings into the work of the case organization. The segmentation is conducted with RFM variables derived on an empirical dataset of donors and donations from a two-year time period, which are then used as input for a k-means clustering model. The results are validated with the elbow method, Davies-Bouldin index, silhouette analysis, as well as nonparametric Kruskal-Wallis and Mann-Whitney U tests. Lastly, the results are validated qualitatively with an expert from the case organization to ensure correct focus and applicability. As a result, five statistically differential segments were formed from the donor base. Three of the segments contain active donors, while two consist of dormant population. The most potential segment contains active donors that respond to almost all of the marketing they receive. The least potential segment on the other hand consists of people that have only donated once over a year ago and are very unlikely to donate again. The results aid the organization in prioritizing donors, allocating resources more efficiently and designing targeted messages for different types of donors. This will enable decreasing the costs of fundraising, and potentially improving the incoming donations, leading to an improved cost-efficiency of fundraising at the case organization.Description
Thesis advisor
Malo, PekkaKeywords
NGO, segmentation, donor segmentation, RFM, clustering, k-means