LENDDO: EXPANDING FINANCIAL INCLUSION THROUGH MINING SOCIAL MEDIA DATA

by Oleksiy Anokhin, adapted from a case study by Dean Caire

Introduction

Digital Financial Services provide an enormous opportunity to deliver formal financial services to underserved individuals.  Large obstacles remain toward meeting this goal, such as customers who lack identification cards; or, for example, on the national level, inadequate credit bureaus.   These types of barriers push users to seek financial services from the informal sector, which can carry higher risks and costs. Data analytics can help close these gaps, to build bridges between the industry needs and existing business solutions.  Services that help bring underserved segments into the formal sector, reducing these risks and costs, also promote the principles of responsible finance.

This blog post describes a data analytics solution, which allowed a team to overcome some formal barriers for potential customers, using modern data driven tools and techniques. Lenddo combined social media data of clients with information, received from survey data received from loan applicants.  The data were refined to drive models; and the outputs resulted in increased efficiency of processes as well as reduced fraud risks and costs for the lender. In addition, it ensured the growing transparency of customer protection together with proper sensitive data management.

Lenddo co-founders Jeffrey Stewart and Richard Eldridge initially conceived the idea while working in the business process outsourcing industry in the Philippines in 2010. They were surprised by the number of their employees regularly asking them for salary advances and wondered why these bright, young people with stable employment could not get loans from formal FIs. The particular challenge in the Philippines was that the country had neither credit bureaus nor national identification numbers. If people did not use bank accounts or services – and less than 10 percent did – they were ‘invisible’ to formal FIs and unable to get credit. In developing their idea, Lenddo’s founders were early to recognize that their employees were active users of technology and present on social networks. These platforms generate large amounts of data, the statistical analysis of which they expected might help predict an individual’s credit worthiness. Lenddo loan applicants give permission to access data stored on their mobile phones. The applicant’s raw data are accessed, extracted and scored, but then destroyed (rather than stored) by Lenddo. For a typical applicant, their phone holds thousands of data points that speak to personal behavior:

  • Three Degrees of Social Connections
  • Activity (photos and videos posted)
  • Group Memberships
  • Interests and Communications (messages, emails and tweets).

More than 50 elements across all social media profiles provide 12,000 data points per average user:

Across All Five Social Networks:7,900+ Total Message Communications:

·       250+ first-degree connections

·       800+ second-degree connections

·       2,700+ third-degree connections

·       372 photos, 18 videos, 27 interests, 88 links, 18 tweets

·       250+ first-degree connections

·       5,200+ Facebook messages, 1,100+ Facebook likes

·       400+ Facebook status updates, 600+ Facebook comments

·       250+ emails

Data Usage

Confirming a borrower’s identity is an important component of extending credit to applicants with no past credit history. Lenddo’s tablet format app asks loan applicants to complete a short digital form asking their name, DOB, primary contact number, primary email address, school and employer. Applicants are then asked to onboard Lenddo by signing in and granting permissions to Facebook. Lenddo’s models use this information to verify customer identity in under than 15 seconds. Identity verification can significantly reduce fraud risk, which is much higher for digital loan products, where there is no personal contact during the underwriting process. An example from Lenddo’s work with the largest MNO in the Philippines is presented below.

Lenddo worked with a large MNO to increase the share of postpaid plans it could offer its 40 million prepaid subscribers (90 percent of total subscribers). Postpaid plan eligibility depended on successful identity verification, and Telco’s existing verification process required customers to visit stores and present their identification document (ID) cards, which were then scanned and sent to a central office for verification. The average time to complete the verification process was 11 days. Lenddo’s SNA platform was used to provide real-time identity verification in seconds based on name, DOB and employer. This improved the customer experience, reduced potential fraud and errors caused by human intervention, and reduced total cost of the verification process. In addition to its identify verification models, Lenddo uses a range of machine learning techniques to map social networks and cluster applicants in terms of behavior (usage) patterns. The end result is a LenddoScore™ that can be used immediately by FIs to pre-screen applicants or to feed into and complement a FI’s own credit scorecards.

Conclusion

This case study demonstrates that formal barriers of financial inclusion for potential customers can be overcome sometimes with the help of modern data driven solutions. Analyzing activity of future clients indirectly through contemporary approaches and tools in certain environments with less traditional formats of access to financial services create opportunities for all interested parties. Such solutions help lenders mitigate their risks, decrease costs, and improve own efficiency, following the best practices in customer protection regulation and its transparency, responsible pricing and respectful treatment of own clients; in turn, borrowers, receive access to financial resources which were not previously available to them due to strict formal rules of a traditional financial sector.

Adapted from a case study (prepared by Dean Caire, IFC)  and additional content presented in the Data Analytics and Digital Financial Services Handbook (June, 2017), this post was authored by Oleksiy Anokhin, IFC-Mastercard Foundation Partnership for Financial Inclusion, for the Responsible Finance Forum Blog June, 2018.