Data analytics has become a core part of Online PSB Loans

Digital credit infrastructure provider, Online PSB Loans (OPL) develops and integrates advanced technologies to automate and digitize the lending processes for both borrowers and lenders by creating a completely scalable system. Founded in 2015, the marketplace portal, www.psbloansin59minutes.com, claims to provides digital approval within 59 minutes or less if the applicant can input/upload the necessary data. Online PSB Loans, now has 95% of the country’s commercial banks on its platform and helps them to streamline approvals for several loan types, including business loans for MSMEs and retail loans

The company has built expertise in solving the pain points of credit delivery to the MSMEs and retail borrowers. OPL rolls out unique analytics and algorithms for AI/ML-based credit products and provides high-end Fintech solutions to credit institutions. OPL is closely working on the process of opening its infrastructure for various other players. The usage will be for bank schemes, government schemes, Industry ecosystems, etc. This will help the ecosystem achieve operational efficiencies and better ROI.

Expediting digital loan application 
The digital loan application process for MSMEs requires collecting Goods and Services Tax (GST) data, Income Tax Return (ITR), and Bank Statements (6 months). The GST data is sourced in JSON format, ITR data in XML, and the bank statements are in pdf format. The challenge from a business perspective was handling varied versions of GST, ITR, and Bank statement formats. The GST APIs go through continual changes and all new releases need to be accounted for in terms of the completeness of data being captured on the marketplace platform. Similarly, ITR forms change with changes in taxation amendments almost every year.

With the launch of the new Income Tax filling system last year (2021), the ITR format has been changed to JSON from XML. And with bank statements, the pdf formats vary due to the non-standard content across banks. On the technical front, these semi-structured data pose difficulty in getting relevant insights unless they are parsed and standardized along with the business process flow. The team went through analyzing attributes that would make sense to parse and standardize for insights. Additionally, the processes needed to be scalable to cater to large data sizes and cloud operations.

“The team has overcome these challenges and standardized the processes to the extent possible over the past three years. We now have our in-house pdf parsing and bank statement analyzer tool along with a third-party service API that we use. Our entire solution of the platform, data lake, and Tableau is on the cloud infrastructure,” informs Paresh Ashara, President of Analytics, Online PSB Loans.

Why Tableau solution

Before the Tableau Embedded Analytics solution was put in place, the necessary data analysis and dashboard/reports creation was an offline process and time-consuming with limited scope. The question the company had when initially started thinking about putting the embedded analytics into the platform was whether to custom-build or go for a COTS.

“Given the effort and timeline involved in building the solution ground up, we decided to go for a COTS. However, in determining the ideal tool to use, it was apparent basis prior knowledge and experience that Tableau would fit the bill. We also had thought through in terms of futuristic features like creating automated insights commentary for some of the performance dashboards that we wanted to create on a quarterly/half yearly/ yearly basis. This support is available in Tableau and we are now working on leveraging the feature,” informs Paresh Ashara.

With the implementation of Tableau prep, desktop, and server solution uses information captured on the digital platform for the borrower journey and receiving updates from banks on sanction and disbursement amount, date, etc. The analysis covers the business operations performance, product performance, and productivity insights. It provides a relative ranking of the bank with respect to top-performing banks across multiple critical KPIs.

The online reports/dashboards are useful for operations as well as mid to senior management consumption. Apart from these, the solution also share quarterly/half-yearly/yearly reports that contain additional insights and commentary. These reports are delivered via email as pdf attachments. Since the launch of the initial set of reports/dashboards, we have added several new dashboards that are shared on a monthly basis with a handful of PSU Banks.

Finalizing data strategy 
The different datasets that are getting captured as a part of the loan application process on marketplace portal required a defined data strategy to parse and standardize the data for generating insights. So, most of the effort went into identifying the data points available and how to leverage them to provide insights. However, it took OPL around three months to release the first batch of dashboards and reports for use.

“We spent a considerable amount of time getting our data architecture right to support the insights that we wanted our Embedded Analytics solution to deliver. The platform caters to two broad categories of loans viz. business loans and retail loans,” he says.

Under business loans, there are two subtypes viz. term loan and working capital loan. These loan products cater to MSME and Mudra segments. Under retail loans, there are three products viz. auto loan, home loan, and personal loan. So, depending on the application for the loan product, the combination of data that needs to be collected varies. There were challenges in terms of the quality of data and consistency of output for doing comparative period performance analysis. Once those challenges were solved, it was easy to build the reports/dashboards in Tableau and integrate Embedded Analytics into the platform.

The Tableau solution has certainly helped both the external and internal stakeholders in terms of satisfaction. The banks can get their performance insights periodically basis which they can decide the next steps. The internal team is happy that the manual grunt work of data crunching has been drastically reduced and they get to spend more time on analysis and insights creation as well as identifying newer analyses in form of new dashboards/reports. In the past 8-9 months, the team has delivered a newer set of performance dashboards/reports for both internal and external stakeholders.

Catering to future needs
The current solution of providing insights commentary is manual for the quarterly/annual dashboards/reports that we generate for the banks.  “We would like to automate this commentary generation through programming logic supported by Tableau (which we have recently started working on) as the immediate next step and maybe later explore a new solution component of tableau,” says Ashara.

Recently it launched another marketplace portal, www.jansamarth.in, which caters to priority sector and interest subsidized schemes of government across education, Agri Infrastructure, business activity, and livelihood needs. This new portal has all public sector banks and the majority of Cooperative Banks onboarded to offer relevant loan products.