The last five years have seen an extraordinary number of poor people starting to use digital financial services via their cell phones in one way or another. My work in specific countries and with businesses makes me optimistic that this trend will continue. And, as we all now know well, in that same time period, even more people have acquired their first phone and SIM card and joined the ranks of the over five billion cell phone users.
All this cell phone use generates data – from basic call data records
, to mobile money transactional data, to data from social media usage and so on -- that leaves what can be called a ‘digital footprint.’
Whether we call it digital footprint or something else, the truth is that the existence of this data is quite extraordinary for those of us interested in developing services for the poor and people with little or no formal financial access. In fact, in a number of countries, basic cell phone usage data may be the only source of information on vast sections of the low-income population that is both electronic and available in aggregate form in one or two data warehouses. In other words, it is available in a way that can be analyzed.
My colleague Kim Muhota and I describe in this recent CGAP publication how as long as consumer interests are protected and security, privacy and ethical use concerns are addressed, these data can become a useful way to reach unbanked poor with a range of financial services.
We describe how the data can be considered either passive (information generated from a call I made) or active (information I actively provided in response to survey or on social media sites). While active is potentially pact with insight, even passive data can be powerful as MIT researchers pioneering the field of “reality mining” have shown
. Not all countries specify data retention rules so the duration and quality of data may vary. The potential for innovation is with data that is non-financial in nature (the duration of a call, for example) because it covers more people and has not been explored so far.
We expect all sorts of internet usage to eventually become part of this digital footprint. In developing countries, people across all income groups are leapfrogging how they access the internet (via mobile) and what they access (social media). In Kenya, for instance, where there are a reported 25 million mobile subscribers and 17 million mobile money users, 90 percent of the country’s internet usage is via mobile and 31 percent of internet users are on Facebook. In fact, it will be interesting to see how Facebook Zero
, Facebook’s text-only service takes off in markets like Ghana
. More on social media in a forthcoming post by my colleague.
In our publication, we briefly highlight the opportunity in using both predictive and propensity models to deliver a range of financial services. Most of the recent experience has been with credit and companies like Cignifi and Experian Microanalytics, profiled by our colleagues at GSMA, lead the charge. Credit is also an area where US-based companies are experimenting with all sorts of data sources including the social graph. An interesting company is DeMyst.Data which mines a wide variety of data – their demo stood out among others at a FINNOVATE event last fall.
There is evidence from these companies and elsewhere that alternate sources of data can impact credit access. A recent study in the United States by PERC also found that those earning $20,000 or less annually saw a 21% increase in acceptance rates when non-financial payment data was used in credit profiles.
You can download the publication here