Consumer Segmentation for a National Telco provider
Thinking Machines identified key consumer segments in the Philippines' digital economy using consumer transaction data involving millions of dollars.
A pioneer in the mobile money and payments market, and developed the first prepaid online payment app in the Philippines. It is backed by two national giants in the wireless and telecom industry.
Illustrated by Robbie Bautista
Using big data to tap into the digital economy
PayMaya, the first prepaid online payment app in the country, serves as an easy-to-acquire, mobile-based passport to the digital economy for “uncarded” Filipinos. The country’s untapped mobile money and payments market has 40 million potential consumers, which calls for deep insights into consumer behavior to potentially generate millions of dollars in revenue from thoughtfully-targeted products and services.
With Paymaya’s business model lending itself to easy, quick generation of consumer transaction data, they sought to leverage their big data by bringing in the Thinking Machines team to extract data-driven insights such as market segmentation based on consumers’ spending patterns.
Consumer segmentation using network analysis
Thinking Machines conducted a network analysis on Paymaya merchants to identify merchant groups. Treating each merchant as a node in a network, two merchants will be connected if at least 20 PayMaya users transacted from both those merchants.
The network was then divided into clusters using the Louvain Modularity algorithm, which detects groups of merchants that were more densely connected to one another than to others.
As a result, the network analysis identified 8 types of digital consumers with distinct, specialized interests ranging from gaming to travel.
Model interpretability using exploratory analysis
For better interpretability of results, distinguishing behavioral patterns for each consumer type were teased out through an exploratory data analysis on merchant and consumer transaction data.
For instance, gamers had a distinct pattern for transaction frequency, which was then validated by external research as peak days coincided with a games sale. The analysis also looked into the relationship between transaction frequency and transaction amount, with urban navigators having the most frequent transactions and jetsetters having the highest transaction amount.
This solution is one of the many use cases of exploratory data analysis and machine learning to extract insights from consumer transaction data, which Paymaya can leverage to create consumer-customized products and services.