Improve Customer Service and Intelligence with MDM and Data Virtualization
- By Ravi Shankar, Salah Kamel
- March 16, 2017
When my son entered college, I advised him that using a credit card for his purchases would be a better choice than using his ATM card. He agreed and applied online for a credit card with the bank where he has had savings and checking accounts since he was 10 years old. After two weeks, he received a rejection letter with the reason: "Insufficient credit history."
Why did that happen? He has had thousands of dollars in his savings account -- earnings from his paid internships. Further, his account is linked to my account, which I have had for more than 25 years with a solid balance, and between cash accounts, past credit cards, and mortgages with that bank, I have had transactions totaling hundreds of thousands of dollars.
It's unlikely the bank didn't have this information -- financial institutions are notorious for hanging on to historical data, so how did they miss this crucial link? Was it a failure of intelligence or just a missed opportunity?
Not an Isolated Problem
Unfortunately, my son's experience is all too common and is the result of a widespread problem of storing and using the same customer's data in different systems across different business units, in this case consumer banking and credit. Many financial institutions have embraced big data and cloud technologies to better serve customers; however, they often inadvertently isolate data into separate datasets, possibly in diverse locations. With customer information and associated transactions siloed in core banking, loan origination, and risk systems, it can be difficult to create a unified customer view.
This often goes unnoticed, particularly when transactions are limited to a single business unit. However, when multiple business units are involved, these operational problems surface because the bank had not created a holistic view of its customer.
Some systems haven't kept pace with today's multichannel banking environment, and as a result, even financial transactions related to a single business unit can live in different systems across the enterprise, providing yet another siloed or partial view of the customer.
This inability to truly understand the customer comes with a huge price tag -- it increases the burden on customer service time to address problems or questions and it erodes consumer loyalty, eventually affecting the enterprise's bottom line.
A Complete, Contextual View of Customers
When a company like the bank described above lacks a single version of the truth, whether for customer, product, or service data, what should it do? Despite advances in data capture and analytics, many organizations also lack the ability to capture the real-time customer transaction data needed for a complete, in-context view of the customer.
Two technologies used in tandem can address the issue: data virtualization and master data management (MDM). Data virtualization connects data located in physically diverse locations, and MDM provides a unified view based on an organizing concept such as customer, product, or location.
Master Data Management
MDM systems provide an integrated view into disparate datasets by automating the resolution of conflicts and establishing master records. Until recently, the process of creating, integrating, and managing these types of solutions was burdensome and expensive. Outdated offerings from large enterprise software vendors or heavily customized systems developed internally didn't fulfill the need for fast, efficient, and flexible technology.
Today, innovative new software applies an agile approach to MDM, overcoming the limitations of legacy systems (which implement different solutions for product, customer, and supplier data). An agile approach also allows organizations to adjust to changing business requirements quickly and implement incrementally in order to test and deliver value as they go.
Data virtualization handles the second part of the puzzle: accessing transaction data in real time, capturing minute-to-minute changes. Compared to conventional data integration, this technology doesn't require data to be moved before use. No matter where it lives, a virtualized view of the data means that it can be available to any system in real time with low latency.
A query to the data virtualization layer is resolved by accessing multiple sources, but to the user it feels as if the data comes from a single repository. It does this by capturing the data about the data (metadata) in each source without moving anything. This layer sits above the disparate systems, regardless of whether they are in the cloud or on premises.
Data virtualization has some additional benefits as well. Because no data is replicated, companies can avoid the costs and risks of maintaining multiple live instances of a single data set. Also, new sources can be added relatively quickly, so companies can always expand the breadth of customer views without investing in costly, time-consuming modifications to the infrastructure.
Modernization Without the Pitfalls
With MDM and data virtualization working in concert, companies can gain a complete view of the customer, including his or her household and related transactions. If my bank had this perspective, the credit department would have realized that my son had an established account on the banking side. The bank could have even reached out to me, the long-standing, loyal customer, and asked whether I wanted to cosign for the new account.
Leveraging data virtualization and MDM together, companies can avoid subjecting their customers to the pains of disparate systems with poor integration. The bank would have benefited from a more efficient use of resources, and this would have resulted in improved customer loyalty. In addition, it would have been able to engage in much more powerful, comprehensive analytics and cross-selling opportunities. It could have taken advantage of this opportunity to add me as a joint account holder and cross-sold me a credit card as well! After all, the more products (read: connections to the bank) I have, the less likely I am to switch banks.
Savvy consumers and data stewards at modern organizations demand more -- their operational systems need to deliver real-time, unified views of data, regardless of where the data resides. This is a reality today and it is saving organizations considerable time and energy while delivering a superior customer experience.