Page 269 - The Digital Financial Services (DFS) Ecosystem

ITU-T Focus Group Digital Financial Services
Ecosystem

Currently, relatively little is known about the lives of the BoP, since there are few paper and electronic records.
This lack of data hurts the BoP. For example:

• Obtaining credit is difficult without a credit history file or electronic/accessible proof of income.
• Businesses have less insight into consumer needs.
• Medical professionals react less quickly to changing health conditions. For example, Google searches can
reveal the outbreak of a virus.
Social networks are uniquely positioned to collect this data:

• Deep insight: Besides simple demographic information like age, gender, and location, social networks
collect ‘softer’ data, like interests, behaviours, and attitudes. Social networks also capture all dimensions
of a user’s life: leisure, social, work, and school activities.
• Heavy user engagement: Because users visit social networks often, the data are fresh.

MNOs have been touted as a good source of data, particularly for making credit decisions. For example,
research has shown that calling behaviour (frequency, length of call, number of contacts, locations, etc.)
indicates different levels of credit risk, but, social networks are likely better positioned than MNOs. As more
MNO voice and messaging traffic shifts to Voice-over-Internet-Protocol (VoIP)/chat platforms, MNOs will lose
their insights, with social networks picking up the difference. Social networks can then add these new insights
to their already deep user profiles.

4.5.1 BoP use cases and benefits

Consumer credit

BoP consumers have trouble obtaining formal credit. One reason is lack of information (verifiable income,
repayment reputation, etc.). Social networks can help solve this problem. First, wallet activity provides valuable
data points about income and expenses. Second, social network data (attitudes, friendship references, etc.)
create a broader view of an individual’s propensity to pay.

Commercial credit

Social networks can also help businesses borrow. As with consumer credit analysis, data will likely include
transaction histories and supplemental data. But, analyses will likely employ different metrics. For example, a
downward trend in "Facebook likes" may indicate growing credit risk.

Targeted advertising

While the BoP do not fit the stereotypical ‘big brand micro-targeting’ use case, local merchants may find
targeting very valuable. Even simple attributes like location are useful. Someone selling a bicycle may target
individuals within a 10 km radius.

Targeted communication
Non-commercial communication also needs targeting. For example, a government may deliver health messages
to a drought-stricken region.

New product and program identification
Through sentiment analysis and other techniques, commercial companies, NGOs and governments can identify
new products/services and helpful community interventions.

241