System redundancies – challenges, roadblocks, and a smarter approach
Given current business scenarios, data is of immense importance. From manufacturing to insurance, data (or specifically, relevant and clean data) is pivotal to operational and customer excellence across multiple industries.
Leading insurance providers, as a result, are rethinking core systems – undertaking incremental changes to ensure infrastructure is in-line with available data and analytics tools.
However, the above landscape does pose several challenges.
An end-to-end transformation is an exorbitant and protracted process, leading insurers to delay the transition pathway, staying on with outmoded legacy systems.
Despite the enormity of cost and timelines involved, the systemic change could drive significant benefits: pre-emptive and proactive decision-making structured around clear and transparent data.
The road ahead is, therefore, built not only on core system rewiring, but on a larger strategic change of direction. Companies must endeavor to craft a robust future-proof reporting and analytics blueprint outlining how bolstered analytics capabilities could work in tandem with a comprehensive systemic transformation.
In fact, insurance companies have been struggling with legacy complexities for a while now, impeding their business agility and retarding a possible competitive edge.
Corporate MIS (Management Information Systems) in insurance also carries a similar history.
The analytics flowchart for MIS is based on outdated structured relational database technologies, incapable of managing the sheer magnitude of volume, voracity, or range of the data in operation today.
In general, insurance organizations possess multiple data marts and data stores; these store data in a variety of formats with minimal adherence to standardization or consistency.
Obviously, data duplication is a common occurrence with a possible reconciliation often arduous. Data governance is a major challenge, enterprise reporting accuracy is hindered and end-users are unable to access necessary data with minor alterations requiring extended timelines.
This large-scale data blockage and rigidity leads to business users creating their fragmented reports with a parallel ‘shadow IT’ culture at work within the organization.
Dichotomies – two paths and the lack of ‘oneness’
Before one resolves the above issues, it is important to understand the flowchart for the data movement and processing. Here’s a brief breakdown of how data moves ‘into and out of’ systems in the quest for clarity:
- Reporting data is integrated into an enterprise data warehouse (EDW) for enterprise reporting purposes
- A number of subject specific data marts (claims, policy, billing, etc.,) are put in place to manage operational reporting needs or support specific analytical requirements
- Data is accessed via one or more IT-controlled business intelligence tools and interfaces - Business Objects, Cognos, and Micro-strategy
- The landscape is governed by a dedicated MIS or BI department, part of the larger IT infrastructure
- MIS or BI divisions control the data warehouses/data marts enabling various tactical and strategic data necessities
The architecture for Advanced Analytics utilization is different.
Data scientists (or insurance actuaries) possess their own data infrastructure models with specific toolkits that can envision actuarial and predictive data frameworks.
Data scientists are known to often bifurcate away from the MIS functionalities (arguments range around the arbitrariness, rigidity, and inflexibility of MIS) and gravitate towards the creation of individual ‘sandboxes’. These ‘sandboxes’ can extract additional data from several external sources and develop clearly delineated analytical models.
The above duality of systems is an aberration and needs immediate resolution. Duplication, discrepancies, corporate governance inefficiencies, and data quality shortcomings are rampant – leading to the absence of a unified source of truth.
Singularity – immense possibilities, deep complexities
How do organizations identify a workaround against the above issues?
A singular, enriched, and connected ecosystem is essential — an approach that contains and synergizes the various data streams, tools, and capabilities into a single whole, supporting the needs of the entire organization. This is a difficult process – and would require a layered and multi-levered game plan.
Actuarial and predictive model creation needs strong computing abilities and data storage potential – often beyond the ambit of a corporate reporting environment. Conventional ETL tools, relational databases, and local area networks cannot easily process voluminous data sets with several streams of structured and unstructured data.
Additionally, ‘power user’ communities and advanced users who operate within business areas are steadily rising in number — with their demands for fast and flexible data access often difficult to satisfy.
Reduced IT budgets have also pushed internal MIS organizations to respond to the best of their abilities through a number of solution sets:
- Data warehouse platforms
- Visualization tools
- Data quality/Data governance mechanisms
These are, of course, workarounds against a fundamental architectural incapability. The separate structural platforms for MIS reporting and advanced analytics continues to remain a steady challenge.
Only a few insurance firms have the capacity to streamline developing, testing, and auctioning analytical models, while effectively managing the burgeoning demands of user communities. New technologies such as Hadoop and Big Data are often unclear and vague to these enterprises with the lack of a smoothened migration pathway.
An ideal decision support framework would be seamlessly embedded into the operational matrix with guides on best practices/action areas and alerts for critical areas requiring attention.
Interactive dashboards could add differentiated value — individualized to specific organizational needs — as against template reports largely in use today.
The way forward may be fraught with challenge and change, but could over the long term revolutionize data management for the insurance industry.