Master Data Management vs Data Integration
Data mastery and intelligent workflows have the strongest impact on efficiency, revenue growth, service quality, and customer satisfaction - Deloitte insights on digital maturity and financial performance.
Insurance companies are in different stages of digital transformation. Whether it is automated workflows, AI or IoT, they all have one commonality - data.
Taming data challenges in the insurance sector
The increasing digital interactions with customers are generating new types of data and most of it is unstructured. To put this into perspective, the Internet of Things is resulting in the creation of 2.5 quintillion bytes of data every day.
IoT insurance data is on its own a massive repository of information that will find increasing usage to improve risk assessment, claims leakage, and product pricing.
This complex treasure of Big Data is worthless if it is scattered across different systems without one consolidated view. Any data management system is worth its investment only if it is an effective interface between the database it controls and the applications that access them. This includes providing clean analytical information to drive operational decision-making.
More often than not, each insurance product in a company may have its own methodology to capture and store data about customers. Customers across product lines are rarely recognized as a single entity and that leads to not being able to leverage customer value.
According to a KPMG report, 56% of CEOs have concerns when it comes to the integrity of their data. Building confidence requires a unified data strategy and there are two routes to remove the opaqueness of diverse data silos.
Master Data Management: It is a complete software system that centralizes the data to one master copy which is then synchronized to all applications.
Data Integration: Unlike MDM, this process combines data, that remains in various source locations, for a unified view.
Pros and Cons of a Master Data Management System
A typical auto claims experience for a customer is them calling their insurer’s call center. They then have to keep punching buttons until they are finally directed to an appropriate customer service rep. And then go through a long process of providing their insurance information and claim. Now compare this with a system that employs Master Data Management (MDM). In this instance, an MDM approach can pull customer’s data and share it with customer-facing systems. Customer service can immediately access the customer profile, access their GPS locational data, and provide personalized service.
It is not just customer data, a master data system will link all critical business data to a single point of reference and deliver a ‘master record’. It does this by linking an organization’s data using a common architecture. This is a complete end-to-end approach that ensures data quality and can be encapsulated as:
Data collection from different sources
Application of business rules to ensure data quality
Data transmission to different workflows and verification of data during migration
The technology allows insurance systems to cover a single domain or multi-domain. A multi-domain data management platform will support different data assets, such as Customer Data Management and Product Information Management.
The disadvantages of a Master Data Management (MDM)
It is very complex and requires the involvement of too many stakeholders. Due to the complexity, the budgets for MDM implementation are too costly to be easily justified.
MDM finds it almost impossible to reconcile externally sourced data, machine-generated or unstructured data. That puts it at a decided disadvantage because the insurance sector is moving towards IoT and third-party data in its digital transformation journey. The final clincher is that MDM may end in creating another data silo in the business ecosystem.
A Data Integration Platform
Data silos need not be the devil they are made out to be. Data silos in insurance allow carriers to comply with different state regulations. Additionally, some data MUST be stored separately to comply with legal regulations. Data might also be too valuable in specific business operations to permit it to be consolidated or normalized to suit MDM data rules. This is why many insurtech companies advocate a data integration platform to create a unified view.
A data integration layer is not a single solution like MDM. It is an architecture design that operates at the compute layer and can therefore connect data wherever it resides - in the cloud or on-premise. Above the compute layer, sits the EKG Layer (Enterprise Knowlege Graph), this layer uses a semantic graph that maps entities, their metadata as well as the relationships between them. This involves a data virtualization approach to connect to different physical data stores. By creating a virtualized view of the underlying physical environment it does away with the need to physically move the data as in the case of an MDM system.
An insurance ecosystem can run up to 500 applications and this number will continue to grow. The point made here is that most of these applications cannot easily communicate with each other. It is not feasible to eliminate silos through data migration and consolidation. The more practical way to approach the communication hurdle is by an integration layer that prevents the creation of copies of the data. Data integration platforms can also enable cleansing and monitoring data so that it can comply with data governance rules.
The biggest benefit of data integration layers is that they can bring into their ambit, data that powers digital transformation. As long as insurers are dealing only with structured information, a database management system will work acceptably in an enterprise data landscape. That landscape though is already changing, the increasing relevance of external data sources, the emergence of IoT, and the shift to multi-cloud environments necessitate a more responsive data strategy. Data integration platforms like SimpleINSPIRE, combine the flexibility of the cloud along with computational power to deliver data and analytics to the frontline.
Topics: System Architecture