Migration, like fitting a cube through a round hole

annika-edit-1-scaled-7700943
Annika Karppinen, Evitec Life Product Manager

When an insurer plans a system renewal, the primary focus is usually on how the new system supports needs today and in the future. However, few insurance companies start from scratch. Especially within life insurance, policies may be older than 50 years. Therefore, the migration of run-off portfolios usually pops up at some point during the renewal project.

Older systems often have an “uncontrolled flexibility”, a feature that originally was regarded quite handy. Individual policy details could be modified in many ways, and not all information had a designated place or format. Thus, over time, users may have entered the same information in different places and, for example, dates in different formats. Older policies also do not always have all the information required by the new structure; in which case the policy information needs to be enriched. Not to mention file formats, which have changed over the years. There are certainly many more examples. And now, 15–20 years later, when this rather mixed data should be adapted to the structures of the new system, we are faced with a data cleaning task. The scope of a migration project can often be bit of a surprise, but luckily there are tools available to help.

The power of collaboration

In data migration the cooperation between the insurance company and the system supplier is key. The insurance company knows its old products and can foresee some of the challenges in the data structures. The system supplier on the other hand, knows inside out the logic and structure of the new system. When a mechanism for checking the quality and consistency of the data is created in between, even a difficult migration becomes easier.

The three phases of migration

Data migration can be divided into three phases. In the first phase, the migration is planned, and the portfolios are studied to the smallest details. First steps are taken with smaller test data and the creation of data mapping rules starts. At the same time, the insurance company often considers whether some product portfolios can be combined to simplify the management of portfolios in the future.

In the next phase, our conversion tool will take centre stage. It is used to check whether the data to be migrated is consistent and compatible with the new system. Rarely, if ever, is older data ready at once. The conversion tool provides feedback on differences and inconsistencies, such as data fields that cannot be matched in the new structure, missing data fields, or data in an inappropriate format.

This is where the actual data cleaning begins. The same data may be run through the conversion tool several times until it can be stamped as OK. Finally, a policy lifecycle testing will be done to ensure that everything matches in the future as well. For the work to progress promptly, the conversion tool is made available also for the insurance company. Hence, the actual experts of the portfolios and those working on data cleaning can independently test the changes and updates. All in all, a time-consuming phase, but the work is rewarded in the last phase.

The actual migration is often the fastest phase. When the old data has been processed and its compatibility has been verified, this is largely a technical routine, where the converted policies smoothly float into the new system. As a final check, the outcome is reconciled with the source data.

Extensive experience of migrations

In addition to the conversion tool, Evitec Life‘s accurate description of the data structure makes migration work significantly easier. The description gives the customer a clear view of which information is needed and in which format.

At Evitec we have carried out system migrations for several decades. We have converted nearly one hundred portfolios and hundreds of thousands of policies. So, it’s fair to say that our experience has built up over time and our migration process and tools have been put to the test in many demanding projects.