As with every booming industry, Mobile Money Providers (MMPs) within the Fintech industry must stay ahead of the curve to remain competitive by integrating with new technologies, expanding their customer base, and reducing their overall financial risks. On top of minimizing revenue leakage with robust reconciliation controls and applying a comprehensive fraud management framework, MMPs must prioritize data analytics strategies to predict, segment, and manage their portfolio. While it is easier to convince executive teams and secure budgets for Revenue Assure and Fraud Management systems, justifying investments in isolated data analytics tools to support AI/ML use cases is a harder sell.
We at LATRO believe in a converged approach to reconciliation (RA) and fraud detection (FM) controls within a joint system enhanced with AI/ML capabilities. A unified RAFM Data Analytics system leverages common data sets to extend beyond standard reconciliation and fraud controls to provide actionable data analytics findings from bad debt prediction and liquidity management insights. By leveraging the vast amounts of transaction logs, customer profiles, and partner files already processed for RAFM controls, a unified solution empowers MMPs with data informed decision making that enhance operational efficiency, increase customer lifetime value, and optimize Mobile Money margins.
Here are four key data analytics use cases, covered by Assure Fintech, that are crucial for MNO-led Mobile Money Providers to invest in:
1. Customer Segmentation
2. Bad Debt Prediction
4. Liquidity Management
Customer Segmentation
One of the most straightforward use cases for data analytics in Mobile Money is the need for customer segmentation. When it comes to understanding your customers do you really know their preferences, pain points, and behaviors beyond their gender, registration date, and age? Do you have an in depth understanding of your VIP customers possessing the top transaction volumes? What about your entire customer base? And how will you continue to segment your customer base as registered mobile wallet accounts grow +20%, year over year? Often, the biggest challenge with segmentation is handling large volumes of data and continuously updating segment groups.
When it comes to Mobile Money, the diversity in target markets is extremely unique and poses new challenges that most other industries do not face – including telecom. Since the inception of mobile money, providers have been driven to identifying segments of their targeted region who are sending money and what their motivations are behind their transfers. Early in the launch of mobile money this included segmentation into different socioeconomic groups such as migrant workers or demographic groups by gender and age. Now with many MMPs thriving across the globe, customer segmentation has a new importance as providers seek to expand their Fintech services into new verticals such as microloans and microinsurance, all while remaining competitive in a crowded marketplace. Maintaining an in-depth understanding of customer segments will help MMPs make data driven decisions for targeted campaigns more effective and increase customer engagement and making strategic investments into new fintech verticals.
Bad Debt Prediction
Throughout 2023 the most popular adjacent vertical for MMPs were credit products, offered by 45% of global MMPs. The introduction of microloans and credit have opened the door to new financial opportunities to core unbanked demographics such as small business owners, rural farmers, and women. With an increase in credit products comes an increase in risk for mobile money providers. Now more than ever, MMPs require bad debt prediction models. While the capped value of credit and microloans is limited and thus reduces the risk of a single customer owning debt, the number of unique customers who are receiving MoMo loans is continuing to increase. This poses a need for MMPs to take proactive measures to adjust credit limits, modify repayment plans, or reduce the volume of loans given to reduce overall financial risk.
With the implementation of Bad Debt Prediction models, MMPs can gain insights into predictive indicators that are indicative of upcoming reduced cash flow from loan and credit repayments. When occurring on a mass scale, across many customers, this might call for a revised credit offering and a risk mitigation strategy. Some predictive indicators that are offered through bad debt prediction models include anomaly transaction behaviors such as a sudden drop in transaction activities by a wallet, a high number of small value transactions following inactivity, and other irregular behaviors.
Churn Prediction
As new players continue to enter Mobile Money markets across the globe, MMPs are at risk for increased customer churn due to competitors offering more appealing services, or lower transaction fees. This is increasingly important in markets where non-MNO led MMPs are offerining drastically reduced transaction fees. Many non-MNO MMPs are backed by venture capital funding and are less dependent on a transaction fee-based revenue model compared to MNO-led providers. MNO-led MMPs must proactively recognize customers at risk for churn before it’s too late. Identifying churners in mobile money is not as straightforward as tracking churners on the telecom side for two key reasons: unpredictability of spending behaviors and lack of wallet closure.
Compared to mobile CDR usage (voice, SMS, or data) mobile money transaction events can be irregular. These irregular spending habits make it challenging to establish precise churn status predictors, depending on the customers’ inactivity over a set period. In some cases, churned customers might be flagged by an MMP through using a simplistic logic such wallets with no transaction activity in X number of days. However, with this approach customers are likely to have already “churned” and now the MMP is faced with additional customer acquisition costs to replace churned customers.
Similar to the pre-paid SIM Card market, many mobile money customers might also have multiple wallets with different providers. The customer might favor one provider to another based on a promotional transaction fee rate or convenience of an agent’s location. Since there are no subscription costs to keeping a mobile money wallet account active, many customers may be keeping their accounts active but no longer actively use them. This complicates the detection of churn since there is not an official “churned” or “unsubscribed” status for wallet accounts.
Advanced churn prediction models can be used to identify customers at risk for churning before it’s too late so that Mobile Money providers can offer targeted marketing campaigns to try and secure their customers and extend their customer lifetime value.
Liquidity Management
A core function of mobile money providers is managing the flow of e-value or digital currency, this is also commonly referred to as “float”. This task is challenging when paired with a pool of geographically disbursed agents who perform varying levels of cash-in and cash-out transactions daily. Since the volume and monetary amount of transactions can drastically vary, many agents become cash-rich (thus increasing risk for fraud attacks) or cash poor (thus turning away business transactions because they lack cash). There is a fine line MMPs must balance—cash availability versus high-risk targets. There is also a great need to understand inward and outward float across the mobile money ecosystem. Data Analytics can be leveraged to predict cash needs and identify high usage areas to proactively route cash deliveries. Machine learning techniques can also be used to identify anomalies in agent liquidity, such as unusual patterns of cash withdrawal. One example is withdrawing large amounts of cash in a region where transaction volumes do not justify it.
Liquidity management has an immediate ROI by ensuring that agents have enough cash on hand to perform transaction requests, ensuring that they are not at risk for having too much cash on hand, and detecting liquidity-related fraud anomalies.
Conclusion
Over the coming years MMPs (especially MNO-led) will need to capitalize on data driven insights to prevent customer churn, expand their footprint, and optimize liquidity management. The strong use cases for data analytics within the Mobile Money space justify investment in a dynamic tool like LATRO’s Assure Fintech which converges Revenue Assurance, Fraud Management, and Data Analytics within a unified system that not only improves decision-making but also maximize the use of shared resources, leading to increased operational efficiency and ROI. As the mobile money ecosystem expands into new fintech verticals, MMPs that invest in robust data analytics capabilities will be better equipped to manage risks, retain customers, and drive sustainable growth in the long term.
