Modified recency-frequency-monetary model for determining feeder schools as geomarketing promotional target in Indonesia

• Main Author: Ernawati
• Format: Thesis
• Language: English; English
• Published: 2023
• Subjects: Q Science (General); QA Mathematics
• Online Access: http://eprints.utem.edu.my/id/eprint/28292/; https://plh.utem.edu.my/cgi-bin/koha/opac-detail.pl?biblionumber=124029

Promotion to the right target market helps higher education institutions (HEIs) increase student recruitment success and save costs. High schools are HEI student feeders. Thus, promoting to the schools is critical. However, due to many feeder schools and their vast distribution, particularly in Indonesia, HEI management should choose high-value feeder schools as a target market. The most common method to identify the target market is the Recency-Frequency-Monetary (RFM) model combined with data mining techniques. The original RFM model measures customer value by business profitability and focuses more on getting customers with high financial gains. However, educational institutions have different profit values. Therefore, this study adapts RFM model to the HEI marketing context to quantify feeder schools' value as a target market. Besides, geographical information is vital in marketing because knowing customers' geographical distribution enables businesses to devise more effective marketing strategies, resulting in more targeted marketing strategies. Therefore, this study includes geographic information. Besides, only a small amount of research, especially in the education field, has combined RFM model with geographic information and analyzed it spatially. Therefore, this study proposes a framework for determining feeder schools as a geomarketing target market using a modified RFM model combined with geographic information. This study modified RFM model into the Recency-Length-Frequency-CGPA-Persistence (RLFCP) model and combined it with the geographic information about the potential of districts (D) derived using spatial analysis and K-means clustering to obtain RLFCP-D model. Based on a case study at an Indonesian private university, this study chose two clusters with the highest values as the target segment. The target segment has three different enrollment patterns based on time series clustering. The decision tree model was used for profiling. The model's accuracy was 93.19% and 82.65%, respectively, based on validation and testing. This study recommends 141 feeder schools for prioritized target markets. These schools are 6.35% of the university feeder schools, contributing 54.95% of enrolled students. They are in 25 cities and 13 regencies from 19 provinces. The university should prioritize promotion to private schools while also targeting public schools in specific districts. This study contributes a data analytics framework that can help decision-makers choose feeder schools as the target market while paying attention to the location for developing a geomarketing strategy.