Anurag Jain, CIO, Bajaj Housing Finance Ltd., talks about how the company is harnessing the power of data and continuously refining its data strategy to stay ahead in the dynamic housing finance industry.
There’s a lot of talk these days about data-driven businesses, and rightfully so. The ability to adeptly manage torrents of data has emerged as an indispensable factor in determining the success of companies.
Anurag Jain, Chief Information & Technology Officer at Bajaj Housing Finance Limited has been at the forefront of harnessing data effectively to drive business growth. In this interaction, he shares his views on how the company is leveraging data at every stage of its value chain, optimizing customer experiences and what it will take to overcome the challenge of ‘data insufficiency.’ Edited excerpts:
Q. What does ‘Driven by Data’ mean for your business and your customers?
For us, the ‘Driven by Data’ approach is a broader concept encompassing various stages of mortgage lending—right from customer identification and targeting to customer on-boarding, servicing, and monitoring.
With rising macro-economic complexities, using the right mix of lending information about the customer, including loan borrowing and payment behavior and alternative data, becomes imperative.
This is where ‘analytics’, as a domain, comes into play. Manoeuvring and interpreting such big data using AI and ML models helps every organization offer the right proposition to the customer while monitoring early warning signs.
Q. How is Bajaj Housing Finance Limited (BHFL) leveraging data at every stage of its value chain?
At BHFL we believe in using customer data insights to redesign and revamp business processes and models.
‘Right customer selection’ is core to our business, which is done by leveraging the data of a large customer franchise.
The journey starts with a scoring model where we rely on risk profiling to offer each customer the best product suite with the most optimal BHFL loan offer. Campaign orchestration, post leveraging the attributes of the model created, helps to enhance the customer journey, thereby defining the best channel for reaching out to the customer.
It is our constant endeavor to have a robust technological infrastructure to ensure that most prospective borrowers after they have expressed an interest in our loan offering, are duly assisted in their mortgage journey – facilitating journey completion within an optimum time interval.
While the mortgage loan lending space has traditionally been paper-heavy, with increasing digitization and technological advancements, it is indispensable to create a workflow where we can offer a delightful customer experience. Various digital initiatives, including E-sanction letters and E-agreements, have helped us a long way.
Our journey with the customer does not end with the loan disbursement. Using various data analytics techniques, a continuous monitoring of Net Promoter Score (NPS) is done to achieve the highest levels of customer satisfaction.
Q. What are some of the challenges that you face around data?
Targeting the ‘right customer at the right time’ for a mortgage loan continues to be our aim, and of course, there are challenges along the way.
Some of these include identifying customer clusters (right customer) with a propensity for a mortgage loan, and aligning with BHFL risk parameters, while considering the crucial factor of timing (right time), is essential.
Another challenge is ‘data insufficiency.’ Tapping customers’ digital footprints outside our ecosystem is a significant problem. Further, a substantial volume of unstructured data remains available in the internal ecosystem, which in today’s scenario needs to be leveraged more efficiently.
Q. Is too much data causing analysis paralysis?
In today’s world, where more and more complex data sets are being used by organizations with significantly higher volumes, frequently refining data strategies can help avoid the problem of ‘analysis paralysis.’
It is imperative that each data source is assigned a separate confidence score, which is a derivative of variables, including the recency of data, the frequency at which the data source can be updated, and the source from which this data is extracted.
Once a confidence score is assigned using appropriate AI and ML techniques and models, it is crucial to define a specific end-use for each data source. This allows for data prioritization, ensuring that the appropriate set of data is inferred at the right time, and is consumed in the most optimal manner by the end-user.
Q. Why would you call your business Driven by Data?
We want to create a differentiation in customer sourcing and customer servicing in the mortgage lending domain. To achieve this, we are continuously building, monitoring, and refining the forecasting ability of our AI and ML models by capturing the non-linear relationship between the available scenario variables and ultimate behavioral customer outcome.
All efforts are being directed towards adopting a ‘data at core’ strategy, which we understand will directly result in better accuracy, optimized operations, richer data, and a better customer experience.
