Your PAS Migration Questions – Answered by Owen Greenwood, Insurance Client Partner.
If your PAS migration is exciting, something has probably gone wrong. The best migrations are deliberately uneventful: the data has been understood, the difficult decisions have been made early and the migration has been tested so many times that nothing is left to chance on cutover day.
In this FAQ, Dufrain’s Insurance Client Partner, Owen Greenwood, answers some of the most common questions insurers ask when planning and delivering PAS migrations — from deciding what data should move to reducing risk through a repeatable migration engine approach.
1. How do we decide what policy data actually needs to migrate to a new PAS
Start with the migration strategy, not with an assumption that the entire legacy estate needs to move.
For each PAS, understand what data is held, what the future-state system needs and what the migration scope should be. In some cases, the full book may need to move. In others, some books may be allowed to run down, while others may need further assessment because of the quality or condition of the data.
The important thing is to establish clarity on what needs to migrate, how it should move, when it should move and in what sequence before migration execution begins.
Discover further insights. https://www.dufrain.co.uk/not-all-policy-data-should-move-the-pas-migration-decision-insurers-often-miss/
2. How do we know whether our legacy policy data is fit to migrate?
Assess the source data in detail before fixing the migration approach.
That means understanding its structure, data model and quality. Insurance organisations often know their data is complicated, but can underestimate just how complicated it is until they try to move it.
That assessment can materially influence the migration decision. One source may be relatively straightforward, while another may contain quality issues significant enough that the organisation first needs to understand whether — and how — it should migrate.
3. Do we need to migrate every legacy policy book?
Not necessarily. Across a larger PAS estate, some books or systems may be suitable to run down rather than migrate. Others may be clear migration candidates, while some may need further assessment because the data quality is particularly poor.
The key is not to blindly attempt to migrate everything. The migration strategy should determine what genuinely needs to move and the right approach for each part of the estate.
4. What makes our insurance policy data difficult to migrate?
Policy data is time-dependent, not static.
A policy may include information relating to quotes, renewals, endorsements, cancellations, premiums, tax and risk. Those elements need to be understood and handled correctly as data moves from source to target.
The complexity can increase further across specialty risk and less familiar insurance products. That is why PAS migration needs people who understand the insurance data itself, not just the technology holding it.
Find out more in this recent article PAS Transformation: Why Data Complexity Makes Migration Harder
5. We have several PAS platforms, how should we decide what moves first?
Sequencing should form part of the upfront migration strategy.
Where several PAS platforms are involved, the organisation needs clarity not only on what should migrate, but also when it should move, in what sequence and with what scope.
The whole estate should be considered before individual migrations are treated as isolated projects. That gives the programme a clearer view of the wider migration landscape and the path into the future-state environment.
6. How do we map legacy insurance data into a new Policy Admin System?
Start by understanding both the source and the target.
The source data model, structure and quality need to be assessed, alongside the future-state target model. From there, individual data elements can be mapped from source to target and the required transformations defined.
Some transformations may be relatively simple, such as converting a date into the format expected by the target system. Across more complex insurance data, the transformation requirements can be much more involved.
Those mappings and transformations then become part of the migration engine.
7. We have already completed one PAS migration, what can we reuse for the next one?
Potentially a significant proportion of the migration engine.
Where the target remains the same, parts of the engine may be reusable across subsequent migrations. The next source will still have its own structures, quality issues and nuances, so it still needs to be understood properly.
But those differences do not automatically mean returning to the beginning and rebuilding the entire approach. Where appropriate, existing mappings, transformation logic and components can be reused and adapted for the next migration.
8. How much efficiency can we realistically gain from a repeatable PAS migration approach?
It depends on the variation between the sources and whether they are moving into the same target. In a scenario where multiple sources are moving into the same target and there is roughly 30% variation between the source datasets, around 70% efficiency can be a reasonable estimate.
That should not be interpreted as a universal 70% time saving or cost reduction. The value comes from not having to recreate the entire migration capability from the beginning where target knowledge, mappings and parts of the migration engine can be reused.
9. How much testing should happen before PAS cutover?
The migration should be run repeatedly before final cutover. Once the mappings and migration engine have been built, the migration can be executed again and again during testing. By the time the final migration takes place, it may effectively have been run hundreds of times already.
The purpose is to identify problems, refine the approach and reach cutover with a very high level of confidence in the source-to-target migration.
10. How do we reduce the risk of discovering major data problems late in the programme?
Understand the source data early and test the migration repeatedly. The source structure, quality and data model should be assessed before the migration approach is built. From there, source-to-target mappings and transformations can be developed and then proven through repeated test migrations.
That gives the team repeated opportunities to expose issues before the final move, rather than relying on cutover to reveal problems that could have been found earlier.
11. What should a PAS migration partner understand beyond the technology?
Quite simply, Insurance data. PAS migration needs a combination of deep data expertise and insurance-sector understanding. Policy data varies across lines of business and products, and the people moving it need to understand what that data represents as well as how it is structured technically.
Experience matters too. Dufrain has specialists with extensive experience across insurance data and PAS migration, including team members who have completed more than 100 migrations.
12. Does experience with the target PAS make a difference?
Yes it can. Dufrain has worked across major policy administration environments including Acturis, Open GI and Insly, with familiarity across their target data models and frameworks.
That target knowledge is useful because PAS migration depends on understanding both the source being left behind and the environment the data needs to enter. Where several sources are moving into the same target, that knowledge can also support greater repeatability across the wider programme.
How can Dufrain help?