A lifelong journey

Insurance companies are faced with many generic - and specific - challenges when it comes to data. The generic challenges face all companies as they seek to enhance the quality of their data and their competitiveness.

The second, specific challenge, relates to the way insurance companies deal with data. Insurers understand inherently that data forms the underpinning of their agreements with policyholders.

This data can have validity for five decades or more. They all follow the ICO (input check output) model of data management, as it provides the correct underpinning for long-term security and integrity. This model finds expression in data acquisition, quality management and exploitation or utilisation.

At a conceptual level, this translates to:

1. Data profiling and mapping
2. Data quality
3. Data governance

This is the process of assessing/analysing the data and creating data element mappings between two separate, distinct data models. It is a foundational activity, typically viewed as the first step for many data tasks, including data integration.

This step is crucial, as data tends to reside in multiple systems, across many databases and in many applications, and the business needs to be given a single, co-ordinated view of this data.

Many millions - even billions - of words have been written regarding data quality in recent decades. Yet the problem persists, to the extent that it seems intractable. Data quality needs an ongoing, day-to-day commitment, one that spans management changes. It needs to be driven from the boardroom, and ownership needs to reside not with the IT department but with the business stakeholders. The reason is clear: without such clear and direct ownership, individual departments can produce sub-standard data quality.

Data quality is not a destination: it is a lifelong journey, once begun, never ended. It needs an extraordinary level of commitment.

Without such a passionate level of commitment, people who already have policies will be offered another; cross-selling and up-selling opportunities will be missed; CRM will flounder; and more.

Data quality needs an ongoing, day-to-day commitment.

Mervyn Mooi is director of Knowledge Integration Dynamics.

Corporate and IT governance (and so too, data governance) are no longer viewed as “nice-to-haves”. These practices are now indispensable to business. While it is an emergent discipline, it has so far evolved to the point that it embraces data quality, data management, data policies, business process management (as this is the process through which much information about data and processes are gathered), and risk management as it relates to data management.

Just as there has to be answerability, culpability and liability when it comes to corporate governance, so data governance needs to be taken seriously. In the US, for example, there are serious consequences (including jail time as per the dictates of Sarbanes-Oxley) if a company can be shown not to have suitably discharged all its obligations in terms of its future profitability and sustainability. Similarly, all organisations are required to show that they have chosen the right supplier and made the right technical and architectural decisions.

In terms of data governance, there must be a structure that supports the long-term goals of data quality, master data management and data integrity.

Data governance empowers data custodians and data stewards to discharge their duties. It ties their tasks directly back to the boardroom, and allows management to look for constant improvements, and ensure that it happens.

As data governance is an emerging discipline, it is too early to say exactly how it will look in the future, but it is vital to factor in the following issues when it comes to data governance: as per Wikipedia, it is a quality control discipline for assessing, managing, using, improving, monitoring, maintaining, and protecting organisational information. It is a system of decision rights and accountabilities for information-related processes, executed according to agreed-upon models which describe who can take what actions with what information, and when, under what circumstances, using what methods.

Margins are tight, times are tough, defaults are high, competition has never been more intense, and new competitors are eating established organisations' lunch - especially those that offer insurance via the Web, which encourage and facilitate self-service, and which take advantage of new process enablement technologies. Any insurer that does not understand the importance of data in dealing with this threat stands a real chance of losing market share.