A large, 60 year old financial services company, with 5+ million policy owners and with portfolios worth billions of Rands.
The company provides asset management, long- and short-term insurance and investment products with a footprint in several African countries.
The Business Challenge
Over the years, the company has bought out smaller companies and products, and attempted to integrate this data with varying success.
As a result, customer data has become split across 18 BSS systems with various products, rules & formats –
Various SQL databases, flat files, proprietary formats from old mainframes, and semi-structured XML & JSON data.
Customer data was poorly integrated and understood, and the staff and systems have changed over the lifetime of the company.
Furthermore, the company wants to leverage customer insights from social media, call centre recordings, and credit bureau data.
1. Provision a cost-effective, modern data warehousing platform with Big Data capabilities.
2. Model and implement a “Single Customer View” and related data marts, to enable simple reporting and insight into the entire customer base.
3. Identify and group unique customers across the different datasets.
A modern Big Data platform was designed, installed and configured to ingest structured and semi-structured data and serve as the data platform.
The platform was integrated with existing IT infrastructure, such as Active Directory for authentication and BMC Control M for handling workflow.
An EDW-like environment was built to enable development and deployment of 100+ batch jobs to load 300+ datasets daily and monthly.
This allowed the data engineering team to ingest the various data mentioned above, using any Big Data tools in a standardised and supportable way.
Modelling and development of 10+ datalakes was done to provide “data mart” capability, including “Single View of Customer”.
Business users were provided with a SQL interface to access the Big Data platform, so they can leverage their existing skillsets and BI tools.
With the principles of machine learning, unique customers were identified across the different datasets and “grouped” together
This provides a holistic picture of each customer’s portfolio across the company, enabling further use cases.
The Value Proposition
The software was effectively free to use, and commodity-grade hardware was used.
By comparison, a previous Big Data proof-of-concept had required expensive monthly licencing for both software and hardware.
Benefit – Single View of Customer:
• Improved customer experience via the IVR (call centre); the agents have a complete customer picture across the BSS systems.
• Customer self-service reduces call centre load
• Anti-Money Laundering & Fraud use cases, plus regulatory requirements in this regard
• Actuarial modelling and analytics to understand the impact of environment variables (such as VAT changes) on the customer portfolios.
Benefit – Big Data platform:
The company was able to generate real value from Big Data, with minimal risk.
They now understand the place of Big Data, and have identified suitable future use cases for the platform.
No posts were found.