The Bill & Melinda Gates Foundation believes that access to digital financial services is fundamental to enable poor people to become more economically stable, prosperous and resilient. Digital financial services – payments, credit, savings and insurance offered through mobile phones or other technology – are reaching millions of people around the world that have never been reached before. Despite their promise however, digital financial services are still not reaching the bulk of the world’s unbanked poor. In many contexts, data has proven a very powerful input to product development and service delivery strategies. That said, the true potential of data and analytics has not been fully explored in the context of digital financial services for the poor. Sophisticated data collection and analysis tools are often lacking, creating an opportunity for innovation in this area.
Develop an innovative analytics or data capture solution to improve the delivery and use of digital financial services in developing countries. The focus of this call is on solutions that are relevant to commercial strategies and likely to make large improvements in performance outcomes for commercial deployments. Solutions should either create completely new approaches or create improvements an order of magnitude lower cost, faster, higher quality, less risky, greater transparency/auditability and/or more reliable than existing approaches. Requested solution areas include:
- Service delivery location placement and management - Retail network management is the highest cost center for many developing country financial services deployments. In large, distributed mobile money agent networks, agent placement, activity monitoring, demand prediction, customer activity tracking, and other prediction tasks are currently done in an ad hoc and off line manner and could be dramatically improved with analytic approaches.
- User segmentation and credit scoring: Lenders, banks, mobile money deployments, insurers, and other DFS providers have limited client data and poor analytics capability. They often lack insights into consumer and household usage patterns and behavior,customer or agent perceptions of products and brands, financial habits, rates of adoption of different products, and product usage patterns (including of enabling devices like mobile phones).This limits product design efforts and defeats credit and insurance risk profiling.
- Insurance and weather indexing applications: remote sensor data, satellite imagery, measures of agricultural yield, or other analyses to inform agricultural insurance or index products. Additionally, novel ways to collect or analyze ground truthing data on actual yields and link it to satellite imagery analysis would be especially valuable.
- Fraud detection solutions: Going forward there will be increasing risks to both consumers and providers from counterfeiting, petty fraud, and unauthorized third party transactions. Fraud solution applications should solve a problem specific to low resource, developing country settings or mobile money systems.
Many potential approaches will be considered
Successful applicants will fall into the category of devices, software, algorithms, or business models, including:
- New analytics approaches tailored to the context of financial services to very poor populations in the developing world
- New approaches to the above challenges that leverage new forms of digital data that were not previously available or leverageable (e.g. mobile CDR, smart phone data)
- Novel use of satellite data or other remote sensing technology related to digital financial service implementation (e.g. market intelligence, agent coverage and activity, credit risk management, weather data for index insurance);
- Technologies that facilitate fraud detection and mitigation in the context of financial services to very poor popluations in the developing world;
- Data applications that provide incentives to low income consumers to engage with or use financial services in ways that contribute to client welfare (e.g. through personal money management applications or “gamification”)
Winning Proposals must:
- Identify the data it proposes to capture and explain how such data can be used, by what stakeholders, and the anticipated impact the data will yield. (Explaining the value of the application for specific stakeholders is a critical part of the application.)
- Describe how personally identifiable information will be protected if a proposal involves the collection of such information, e.g., the collection of data from households and individuals.
- Describe in detail the analytic approach to be used and why it is superior to other approaches.
- For solutions that rely on mobile technology, the proposal must demonstrate an understanding of the cellular infrastructure in the targeted geographical area, and propose solutions that do not require connectivity beyond the current system. For example, solutions cannot assume that most people have smartphones or that connectivity is fully reliable in rural areas of Africa or South Asia.
- Set forth a clear hypothesis regarding the expected improvements in speed, cost, accuracy, etc. against existing alternative collection approaches. Also a clear test of this hypothesis must be proposed.
- Explain how the proposal would be sustainable and scalable in the developing world context (solutions developed for South Asia or Sub-Saharan Africa will have priority).
- Explain the hypothesized usefulness and commercial value of the proposed solution
- Explain how the usefulness and commercial value will be rigorously tested and validated
Ideas we will not consider funding:
- Basic methodologies or approaches without clear relevance to the Challenge.
- Proposals that do not include a plan for measuring and demonstrating to the foundation the improvement in performance.
- Proposals that would not work in developing country contexts (especially rural).
- Proposals that are highly dependent on permissions to obtain data that are not likely to be granted in a commercial context
- Proposals that do not address the target population (especially women, girls, and populations living on less than $2 a day).
- The development of technical solutions that will provide only modest or incremental improvements in financial inclusion and/or provide benefit in non-strategic populations.
- Proposals that lack a hypothesis or an innovation that can be tested, at least in part, during the initial round of the call for proposals.
- Automation of existing tools without a clear advantage in cost or reach into rural, poor and underdeveloped areas.
- Minor or low-impact improvements to existing approaches.
Applicants must not disclose personally identifiable information or other sensitive data to the Foundation without the prior written consent of the Foundation. Applicants with proposals that involve the collection of personally identifiable information or other sensitive data may be subject to laws or other rules governing the collection, management and protection of such information. Applicants should consult with your own advisors to determine whether laws or other rules apply to your project.