Digital platforms for financial services are one way to increase financial inclusion. In 2013, the Financial Services for the Poor team at the Gates Foundation released a new framework for digital money innovation, exploring the possibilities within digital financial services by asking some big “what if” questions. Analytics, which cut across other innovation areas while accelerating and improving financial services for the poor, are key in this framework.
With what data?
Before analytics can happen, there must be a data source to analyze. A persistent problem in development is there is scant – if any – digital data on the lives and behaviors of the poor. This is especially true when the data needs to be real-time and robust.
However, with the diffusion and adoption of mobile phones, there are now billions of data-generating devices across emerging markets. In July the Gates Foundation launched a report titled “Using Mobile Data for Development,” where we explored the ability to use cell phone data to serve a variety of development causes, ranging from improved public health programs to improved provisioning of financial services.
We learned that mobile data can serve as a strong base layer for capturing robust information on the lives of the poor. And, when combined with other sources such as satellite, agricultural, or banking data, can support powerful analytical applications.
Photo Credit: Mahasweta Mazumder
How can data analytics enhance financial services?
Advanced analytics can support financial services for the poor in many ways. These include simplifying activation processes, accelerating customer product adoption, improving providers’ mobile money service offerings, strengthening fraud detection and monitoring systems within countries, and providing key inputs into product design decisions. Following are two use cases that demonstrate how advanced analytics can solve current problems in digital financial services: agent network optimization and marketing effectiveness.
Agent network optimization
Mobile money networks in emerging markets still rely on an agent network to allow individuals to make “cash-in” deposits and “cash-out” their money. Cash-In Cash-Out (CICO) networks are a critical component of mobile money networks, as they provide customers with the opportunity to deposit, transfer, and access their financial assets through agent terminals rather than having to rely on far-away bank branches. Despite their convenience, agent networks are inadequately planned – often they run out of money or are not optimally placed. Managing cash within these networks is a recurring challenge for providers. The lack of predictability of cash needs due to system shocks and instances of agent sharing drive both customer frustration and the cost and risk of increased cash holdings among agents.
As discovered in our September 2013 study on the economics of payments for financial inclusion Fighting Poverty Profitably, cash transactions can cost almost 90% less when completed by agent networks instead of brick and mortar bank branches. However, approximately two-thirds of such cost savings rely on the ability to improve network operations by optimizing locations, coordinating agent routes, forecasting cash needs, and reducing overall liquidity needs. Advanced analytics using mobile transaction data, social data and mobility data could help achieve these operational efficiency gains throughout the CICO network to both lower the cost of maintaining agent cash balances and to improve customer satisfaction.
Singapore’s DBS bank underwent an exercise where they used “big data” analysis to optimize their ATM network placement and float. Their analytics and the resulting implementation resulted in a 95% decrease in empty ATMs, a 40% decrease in excess cash, and 30,000 hours of client time saved by not having to wait for cash at bank branches.
Improved marketing effectiveness
Customer adoption and dormant accounts continue to be major roadblocks to the scaling of digital financial inclusion efforts. Sixty-six percent of all mobile money accounts globally are inactive, indicating no usage in the past 90 days. Though providers are strongly incentivized to measure, understand and improve marketing efforts to drive both customer adoption and usage, few are able to effectively target mobile money offers and messages to the poor. But by using large mobile and social data, it is possible to create improved methods for understanding and segmenting customers, leading to improved messaging and increased customer adoption and usage.
There is already promising work in this area. Telenor, a Norwegian-based global mobile operator, completed a study with researchers from MIT to use mobile call records as a basis for targeting marketing in the developing world. They set up an experiment – Telenor’s marketing department would target an offer according to traditional methods while the MIT researchers would target the same offer based on analysis of call records and social graphs. The MIT researchers were 13 times better at converting customers to purchase the products. Of those who purchased “from MIT,” 98% retained the service compared to 37% who purchased from Telenor.
If mobile money products and innovations can be better targeted, user uptake will increase. This will enable providers to migrate consumers onto digital wallets, ultimately improving the diffusion of financial services for the poor.
In order for advanced analytics to support financial services for the poor, one must satisfy the commercial incentives of providers in exposing the data for social good, as well as create appropriate standards and regulations around data privacy and security.
First, there is an opportunity for this sector to partner with data furnishers in order to demonstrate how these applications both improve digital financial services and expand access for the poor, thereby creating proof cases for the industry.
Second, many stakeholders – including standard setting bodies, multilaterals, and industry leaders – are concerned with and carefully considering appropriate levels of security and protection for sensitive datasets. Rules and procedures must be established regarding data security, access controls, and user permissions in order for applications to flourish.
If these elements can be brought together, exciting financial service products and innovations will follow, and contribute to improving the lives of the poor.