As financial services digitize, more consumers are bringing their experiences online as well, using social media channels like Facebook and Twitter to raise concerns and issues with specific products or services. Social media data could shed new light into the issues that affect digital consumers and how providers are responding to these complaints. This project will collect and analyze user comments on financial service provider pages on Facebook, Twitter, and Google Play from Kenya, Nigeria, and Uganda, to explore their relevance and reliability as data sources for monitoring consumer experiences in near-real-time.

Policy Issue 

There is evidence that many digital finance consumers may not use formal complaint channels when issues such as fraud or improper conduct from providers occur. Consumers may perceive formal complaint processes as lengthy, time-intensive, and unlikely to deliver the desired resolution. This also poses a challenge for regulators of financial services, as the volume of these complaints may be underreported in the official records that providers are mandated to deliver. As consumers become more familiar with online platforms, they have also begun to voice their complaints and issues on the social media channels of individual financial service providers (e.g., Facebook, Twitter, etc.), creating a new place where institutions could monitor trends in consumer sentiment around the quality of services and consumer experience. However, making sense of this very large number of comments in a systematic manner remains a challenge for regulators and providers alike. 

Context of the Evaluation 

Kenya, Uganda and Nigeria have seen a lot of growth in digital financial services (DFS) in recent years, but this growth has come alongside anecdotal accounts of a surge in the incidence of reported fraud and other risks to consumers. The low penetration of financial services in Uganda and a lack of trust in providers Nigeria has led to a dearth of reliable data sources to monitor consumer risks in a systematic way.1 In a previous study in Kenya, researchers created and tracked responses to the twitter handle @pesastory over four months to understand consumer responses around experiences with digital financial services. Researchers found responses and insights across six consumer protection themes, including “hidden and excess charges,” “unauthorized bank debits (fraud),” and “limited responses to consumer complaints,” among others.

In this study, researchers are trying to explore the possibilities of leveraging social media data to build a more systematic tool that helps financial institutions and supervisory bodies identify and monitor consumer complaints in real-time. 

 
Details of the Intervention 

Note: this study is not a randomized trial

This project is providing the first proof of concept about whether social media channels can provide reliable data to monitor consumer complaints. Researchers, in collaboration with experts from each country, compiled a list of financial institutions with public pages on Facebook, Google Play Store, and Twitter, and a list of consumer protection topics to monitor, such as operational failures, customer care, fees and charges, lending, and fraud. For each topic, the team produced a list of seed words related to each consumer protection topic, that were likely to be associated with the chosen consumer protection topics. The algorithm then used these words to identify other words associated with each topic, resulting in a list of more than a million relevant social media posts across Kenya, Nigeria and Uganda.

The research team will report on how this social media listening tool fared and whether the volume and variety of data sources can be a reliable source of information for government regulators and financial institutions working to improve consumer protection outcomes. Areas of measurement include the volume and relevance of the posts identified to the consumer protection topics, ability to identify replies from providers to measure responsiveness, and ability to segment data based on factors such as location, gender and date to identify patterns of consumer protection experiences across consumer and market segments Based on findings, the research team may also recommend additional interventions to test their experimental effects on consumer complaints volumes and topics.

Results and Policy Lessons 

Project ongoing. Results forthcoming.

Sources

[1] European Investment Bank. “Digital Financial Services in Africa: Beyond the Kenyan Success Story,” 2014. https://www.eib.org/attachments/country/study_digital_financial_services_in_africa_en.pdf.