Borrower Profiles: The Data Dilemma


For microfinance institutions, understanding the borrower base of the informal economy is essential to tailor products for different needs and scales, and to do so without incurring undue risk.  Unlike the digital economy where numerous data sources– ranging from bank account information to social media profiles –are available to help drive more intelligent credit underwriting and product innovation, we operate in a ‘data dark’ environment where virtually every data point has to be collected from the borrower. Working with ‘Dark Data’ rather than the Big Data of the digital world poses very unique challenges for us, and the industry as a whole.

Given that much of the data about the borrower is collected by field staff and requires time commitments from both the company and customer, it is crucial to be clear about its relevance.Why is it collected?  Can it help in the credit decision or product design?


In our data spotlight above we provide a very simple snapshot of our borrower base. We serve largely middle-aged women, of whom 85% have a 10th pass education or less and 75% have total assets less than Rs. 1 lakh. Nonetheless, from the perspective of credit needs, there are multiple consumer segments.

Age is a very simple data point. However, even this can be problematic in the dark data world. Many people don’t know their exact date of birth, resulting in vague or incomplete responses (‘could be 40 or 50; I don’t know’).  When the reported age of an individual deviates by 10 years or more from the actual age, is it a useful metric for decision making?


Read More about Dark Data vs. Big Data


Occupation can be even more problematic. In these communities, people often spend a few months of the year as daily wage labourers, and during other times may make and sell products like snacks or fruit and vegetables that are grown on their land. Are they employed? Self-employed? Retailers? Agriculture workers? Traditional classifications don’t work well in these cases. How does one then extract meaningful occupational patterns that can have value for decision making?

These are challenges that we will have to tackle as individual companies and as an industry.

Borrower Profiles: The Data Dilemma

Leave a Reply

Your email address will not be published. Required fields are marked *