Digital Financial Services (DFS) hold the promise to accelerate and deepen financial inclusion – especially in markets where traditional banking has not yet reached large segments of the population. DFS providers have acquired millions of new customers. But many digital services ecosystems experience intermittent use among the customer base and are working to develop active ecosystems. Data analytics provide an opportunity to help address this activity problem to keep customers engaged. New data analytic methods, combined with responsible use of customer and provider data show strong results for more robust digital financial ecosystems and better financial inclusion.
Geospatial data is one such example. Making use of geospatial data can be especially useful to visualize relationships that would otherwise not come out so clearly and to provide further basis for evidence-based decision making about where to focus strategies and resources for financial inclusion and customer engagement. This blog post explores two ways in which geospatial analytics can be used to drive financial inclusion.
It is key that financial service providers adapt and implement responsible finance practices throughout their operational processes. Customizing their product offering and services based on customer data provides huge potential to improve access to financial services as examples in this blog post show. Doing so responsibly requires managing related risks and educating consumers about the use of their information and data, obtaining their consensus as well as putting procedures in place that allow users to opt out and not share their information is essential and should be a pre-condition for any use of individual data by financial service providers. Or simply not using individualized data at all.
Mapping dimensions of financial inclusion for operational decision making
Uptake and usage of digital financial services
Tigo Cash Ghana was launched in Ghana in April 2011 and quickly established itself as the second-largest mobile money provider in terms of registered users. But despite high registration rates, getting customers to do various transactions through mobile money remains a key challenge and focus. GSMA estimates activity rates on mobile money accounts worldwide to be around 30% on average only.
IFC partnered with Tigo Ghana to deliver a predictive analysis of mobile money adoption and usage. Six months’ worth of call detail records and mobile money transaction data was analyzed using the model and applying it in day-to-day operations. The data driven approach was successful. It helped Tigo to adopt new data-driven approaches more generally; and the one-off analysis helped Tigo to acquire 70 000 new additional customers – active customers, importantly.
Predicted adoption rates from the modelling were used to develop a district level heat map of potential mobile money adoption. This heat map of potential mobile money adoption was combined with the existing mobile money adoption rates in respective districts resulting in a map of districts that shows where the gap between actual and potential adoption is the highest, presenting also the highest potential for financial inclusion in these districts. Using geospatial analytics at a district level ensures responsible data use by visualizing information at an aggregate level that is disassociated from any individual or even groups of individuals.
The predictive analysis that was conducted in collaboration with Tigo Cash Ghana is also referenced in the Handbook on Data Analytics and Financial Inclusion.
In this example, the geospatial mapping helped identify new strategic areas for Tigo to engage customers – particularly those outside the urban centers in Accra. Using a related but different approach, the following example with another IFC project with Cal Bank also illustrates the benefit of using geospatial data analytics.
Identifying areas of biggest need for digital financial services
Mapping exercises cannot only be useful to improve service adoption and usage after they have been launched. Going one step back, they can also be a tool to inform the roll out and choice of locations for new services to be introduced, such as agent banking.
CAL Bank in Ghana is about to launch their new agent banking network with an objective to specifically cater to the unbanked. In order to find out where the unbanked customers are concentrated, the IFC partnered with Fraym, a data analytics company, for a mapping of the unbanked as well as target households for agent banking all over Ghana. This approach avoids customer data issues by not using any sensitive data at all. Rather, public survey data, machine learning models and satellite maps allow geo-localized demographic data to be interpolated to gain strategic insights.
The resulting maps drew attention to areas with high concentration of unbanked that the bank was previously not aware of and led them to make informed choices on where to locate agents to maximize the availability of financial services in these areas.
Heat map of Ghana depicting high concentrations of unbanked households – Developed by fraym
Geospatial data analytics can employ responsible data principles and deliver insights that improve operations and drive financial inclusion. These examples illustrate this through using predictive modelling in the first case, whose results are non-specific to individuals. And in the second case, using publicly available and national statistics information to achieve insights on market demographics.
Good quality and reliably geo-coded datasets linked to a broad range of socio-economic indicators are only slowly becoming available, especially in developing and emerging markets or if available, they are often not used to their full potential. Looking forward, more granular geospatial data is likely to become available – either from DFS providers directly, third-party surveys, or other sources. Granular data can introduce risks or individually-identifiable patterns. These risks must be managed, to ensure customer data protection is at the forefront of the analytic approach. As this blog post illustrates, data can be properly managed, non-specific to groups or individuals and still yield powerful insights.
Adapted from a case study and additional content presented in the Data Analytics and Digital Financial Services Handbook (June, 2017), this post was authored by Sinja Buri, IFC-Mastercard Foundation Partnership for Financial Inclusion, for the Responsible Finance Forum Blog April, 2018.