This article first appeared in Fintech Innovation.

‍

Data science skills continue to be in high demand for fintechs, banks, and financial institutions because of the tremendous opportunity to unlock the power of alternate data in addition to the burgeoning data from traditional sources. Sources of alternative data include social media, telco information, psychometric surveys, and smartphone device-generated information, all of which form major pillars in the rapidly evolving and growing field of alternative credit-scoring.

An increasing number of banks, lenders, digital lending platforms, insurers and financial intermediaries are looking beyond the traditional credit scores that incumbent credit scoring agencies and credit bureaus provide. They are enthusiastically embracing sources of alternative data, which is quickly becoming an integral and complementary component of credit underwriting, fraud management, and risk profiling.

Though each alternative data source has its merits, smartphone-based data has risen to prominence because of its stability and higher predictive power. The predictive power of social media, for example, has waned. Scraping data from a person’s Facebook, Twitter, and other profiles not only represents a narrow view of an individual but intrudes into the personal domain and can also easily be gamed. Psychometric surveys, for their part, are time-consuming and invasive, two characteristics which do not align with the customer journey in the era of digital lending and financial services. Smartphone-based data, in contrast, is more accurate in its behavioral assessment, while providing the digital era “user experience” that consumers are now accustomed to.

The industry transition to smartphone-based data is a watershed case study for all entrepreneurs in fintech, as it highlights the growing need for privacy to be embedded into a product from the get-go rather than added on in an ad hoc manner. Smartphone-based credit-scoring maintains this privacy-consciousness from beginning to end, starting with the anonymized extraction of behavioral attributes and concluding with these being used to credit score the particular profile. Never at any point in the process is an individual’s personal data mined or stored.

Smartphone based behavioral data has been researched to be stable, individualistic, and highly predictive. This makes it an ideal source for credit underwriting, risk evaluation, and early fraud detection. The multiple attributes and combinations evaluated through machine learning algorithms make it practically impossible to game.

Banks and financial institutions not already harnessing alternative data from smartphones usage have an exciting window of opportunity to evaluate the benefits of this complementary approach to traditional scoring methodologies and credit risk frameworks. This provides a unique opportunity to evaluate the demographic segment on the fringes of financial inclusion while also strengthening the view and assessment of individuals and micro-SMEs that have traditionally been part of the financial ecosystem.

Data privacy, controls, and regulations will continue to influence this approach and solutions harnessing the power of alternate data. Fintechs and large enterprises that handle end customer data in an anonymized, ethical and responsible way will have a clear head start as this solution becomes more mainstream. Alternative credit-scoring, in other words, is fast outgrowing its own name because it is probably no longer an alternative but a must have.

‍

Read the original article here.