Summary

A leading NBFC operating in digital commerce needed real-time visibility into customer journeys, from lead generation through conversion, to improve segmentation, underwriting, and risk management. In fast-paced digital commerce, having this kind of real-time visibility is essential for boosting conversions and making informed business decisions delayed data means delayed action, and in lending, delayed action shows up directly as lost conversions and higher risk.

Bajaj Tech.AI designed and built an OBT (One Big Table) Data Warehouse on the AWS tech stack, consolidating data from web traffic, CRM systems, underwriting platforms, payment gateways, and customer interaction tools into a single, near-real-time view. The result: a 15% increase in lead-to-customer conversion, a 20% reduction in underwriting approval times, and a 10% reduction in high-risk loans.

Business Challenge

The client, operating in digital commerce, faced difficulties tracking customer journeys from lead generation to conversion in near real-time. They required a comprehensive view of customer interactions across multiple platforms to improve segmentation, underwriting, and risk management. However, their existing data processes weren't agile enough to consolidate information quickly, and the gap between when customer behavior happened and when the business could see it caused three connected problems:

• Delays in identifying sales bottlenecks, since fragmented data made it hard to see where prospects were dropping off
• Missed personalized marketing opportunities, because customer behavior data wasn't available quickly enough to act on
• Slower decision-making across sales, underwriting, and risk teams, all of whom were working from incomplete or delayed information

These issues negatively impacted conversion rates, operational efficiency, and customer satisfaction all downstream consequences of the same root problem: no one had a real-time, unified view of the customer journey.

Solution Approach

The OBT Data Warehouse consolidated data from multiple sources including web traffic, CRM systems, underwriting platforms, payment gateways, and customer interaction tools into one holistic table (an approach known as “One Big Table,” where related data is deliberately denormalized into a single structure for faster, simpler querying, trading some storage efficiency for significantly reduced query complexity). The platform was built entirely on the AWS tech stack, enabling real-time data integration, processing, and analysis rather than the batch-oriented approach the client had relied on before.

Key AWS components used were:

• Amazon S3 for scalable storage of both raw and processed data.
• AWS Glue for data cataloguing and ETL (Extract, Transform, Load) processes, the pipeline that moves and reshapes data from source systems into the warehouse.
• Amazon Redshift as the data warehouse itself, optimized for large-scale analytical queries.
• Amazon Kinesis Data Streams for handling real-time streaming data as it arrived, rather than waiting for a batch process.
• Amazon QuickSight for data visualization and business intelligence dashboards accessible to non-technical stakeholders.
• AWS Lambda for event-driven data processing and automation, triggering actions the moment relevant data arrived.

These services worked together to allow seamless data integration from multiple sources, giving the client the ability to act on real-time insights rather than yesterday's data.

Business Impact & Results

The OBT Data Warehouse had a substantial, measurable impact across the organization, particularly in tracking digital platform performance and customer behavior. By creating a real-time, mirror view of customer buying behavior, teams across sales, underwriting, analytics, and tech were able to monitor fluctuations in lead volume and funnel performance and intervene in time to matter rather than discovering a problem days later in a weekly report.

• Lead-to-customer conversion: A 15% increase in conversion rates was achieved by identifying bottlenecks in the sales funnel and optimizing customer engagement strategies.
• Underwriting efficiency: Approval times were reduced by 20% due to real-time access to enriched customer data, rather than waiting on batch-processed reports.
• Risk mitigation: A 10% reduction in high-risk loans was achieved by using historical data to identify potential credit defaults earlier in the process.
• Sales and marketing insights: A 12% increase in repeat customers was achieved through personalized marketing strategies, made possible by real-time behavioral data.
• Operational efficiency: Time spent on manual data handling and query optimization was reduced by 25%, thanks to the consolidated data structure replacing fragmented, manual processes.

This wasn't a marginal improvement to reporting speed, it directly changed how quickly the business could detect a problem and how much risk it was carrying at any given moment.

Key Takeaways

• A One Big Table architecture trades some storage efficiency for dramatically simpler, faster querying across a full customer journey
• Real-time data access improved underwriting speed by 20% and reduced high-risk loans by 10% simultaneously speed and risk control aren't necessarily a trade-off
• Consolidating web, CRM, underwriting, and payment data into one view was what made a 15% conversion lift possible
• Event-driven processing (via AWS Lambda and Kinesis) is what turns a data warehouse from a reporting tool into a real-time operational system
• This kind of real-time customer data foundation is closely related to the AI-ready data foundations that banks and NBFCs need for reliable, production-grade AI

Conclusion

For this NBFC, the OBT Data Warehouse didn't just streamline data management, it directly moved the metrics that matter: a 15% lift in conversion, a 20% cut in underwriting time, and a 10% reduction in high-risk loans, all from the same underlying platform. Financial services organizations facing similar visibility gaps across sales, underwriting, and risk can draw a direct lesson: consolidating customer journey data into one real-time view often unlocks improvements across several metrics at once, not just one because the same fragmented data was quietly costing the business in more places than any single team had visibility into.

This kind of data engineering work pairs naturally with efforts like cloud database cost optimization, since real-time data platforms only stay sustainable if the underlying infrastructure costs are managed just as deliberately as the data architecture itself. It also reflects the broader shift toward digital lending built on real-time, well-governed customer data.

Looking to solve a similar business challenge? Connect with our experts to explore the right solution for your organization.