Dismantle your legacy gradually with a modular approach
A barrier to success
Insurance policies, especially for life insurance products, can be longer than 25 years in term. This results in an excessive, long-term preservation of the legacy systems that are needed to administer these policies. A study by software quality firm Cast, who analyzed 250 back-end systems from 38 organizations in 8 countries, found that 69% of insurers still run legacy Cobol applications which were originally built in the 1970s and 1980s. Insurers still rely heavily on these systems for their daily business operations. Whether it’s to bind new contracts, process endorsements, track premium payments or manage claims. All such daily tasks, already increasingly complicated due to the diversity and complexity of insurance products, are supported by decades old software.
These legacy environments are costly to change or even maintain. Much of the data and processes covered by these systems are difficult to analyze and integrate, due to the attrition of most internal knowledge. Digital transformation becomes costly, and risky business. A study performed by Genpact in collaboration with ACORD confirms this, with 68% of insurance companies considering their legacy systems as the biggest obstacle to digital transformation. Even within the subset of firms that views themselves as being positioned ahead its pears competitively, 61% of them still view legacy systems as a barrier to success. It is clear that the majority of firms in the industry are in need of a way to eliminate these barriers, as to enable effective use of digital technologies to deliver positive business outcomes.
The big bang debacle
In an attempt to overcome the challenges imposed by legacy systems, the replacement of these systems is prioritized by leaders on the architectural roadmap, resulting in insurance carriers taking disproportionate risks by embarking on enormous core system replacement tracks. But is it really justified to spend multiple years and millions of euros on such risky endeavors?
The most typical challenge with these transformation programs is the lack of fitment of the target system to absorb the legacy policy data and business processes. Throughout the years, legacy systems have consolidated an inordinate amount of custom business logic to keep up with the continuously evolving requirements mandated by market realities and regulatory bodies. There is no silver bullet solution available that can absorb all these functionalities. This isn’t necessarily a result of gaps in the software market, but due to lack of clear modular boundaries within these legacy monolithic applications. There is a lack of accurate documentation, and the internal knowledge of the system has dissipated over the years.
Another major challenge is hidden data quality issues, causing the target platform to reject the data set. This often results in unanticipated effort from business stakeholders to perform the necessary data cleansing and data enrichment in a timely manner.
Once the target application is finally live – optimistically assuming that this milestone is reached – the resulting solution is often perceived by business stakeholders to be more stringent, less flexible and causing lower efficiency compared to the legacy predecessor.
So, if completely replacing the legacy is too costly and often proves to be counterproductive, is there a better option?
The modernization layer
Insurance carriers need to move away from the notion of replacing legacy systems completely. Improving the way of working and enhancing the agility of the landscape doesn’t necessitate that one must change everything at once. Boarding such huge transformation tracks often distracts the organization from its real priorities, raising the question what business problem they are trying to solve.
A better approach is to dismantle the legacy system and replace only those functionalities which make sense to replace at the time, while keeping the rest unhanged. A modernization layer can be positioned between the front-end and the legacy systems as a foundation for digitalization. This is a hybrid layer combing gradually evolving modernized functionality (often modules of packaged solutions) with custom services containing data and process orchestration capabilities that expose legacy functionality.
This effectively de-couples the fast-paced digital roadmap from the constraints imposed by the lagging legacy layer. Solving the legacy problem can then be realized through small increments, each disentangling a portion of the legacy monolith into a functional building block on the modernization layer. This puts the focus back on the customer journey and allows insurers to have their legacy replacement roadmap driven by business value, while delivering digital initiatives and launching new products and services today.
Slow changing versus fast changing
Features that typically change rarely (and are often particularly complex) are better not placed first in the modernization roadmap. Consider for example the complex calculation routines in classic life insurance, external interfacing in non-life (e.g. RDR and Informex), or financial and actuarial analysis. For these use cases it may be better to retain the legacy at first, thereby freeing up resources to transform differentiating capabilities instead.
Best placed ahead on the roadmap are value-adding features that require flexibility and fast-paced change. Typical candidate building blocks may be pricing simulators or risk acceptance engines that require the flexibility to be tweaked instantly, preferably without the need to involve IT. This can be realized by positioning a rule-based decision engine in the modernization layer. And because this component is built in such a modular way, on top of a modern technology stack, this decision engine can be easily enhanced to integrate with machine learning models in the future.
Another candidate component for the modernization layer would be an adaptive data store that is capable of persisting unstructured, semi-structured, and structured data. This component is characterized by its ability to cope with the rapidly changing data capturing requirements imposed by mobile front-ends and other digital initiatives. For example, consider interactive questionnaires in the context of AML, MiFID and medical or financial risk acceptance (in life insurance), as well as the first-notice of loss use case (in P&C insurance).
Re-platforming as a tactical cost driven approach
On the downside, a gradual dismantling of the legacy may defer its final decommissioning. This may not always be desirable due to the high costs involved in keeping these systems and their underlying hardware operational. However, this doesn’t necessarily mean that pulling the plug is the only option. When combined with a well-thought-out IT spend strategy, operational costs can be minimized while still keeping the lights on. Maintenance can be outsourced to third parties whose expertise is to standardize procedures in order to leverage economies of scale. As for infrastructure, expensive mainframe costs can still be eliminated by re-platforming the application to commodity hardware, on-premise or in the cloud. Such re-platforming projects are low risk, require minimal involvement of internal IT and practically no involvement of business stakeholders. In addition, they are non-disruptive and are able to realize a ROI of 12 to 18 months.
Our insurance architecture quick scan to the rescue
Curious how your organization can unlock the potential of a modular legacy dismantling approach? As thewave is involved at various insurers to put in practice this approach, we can help you define or refine your architectural transition roadmap, starting with a 4-week quick scan performed by our insurance industry-focused architects. The result is an independent review with actionable insights of how your organization can best modernize its legacy landscape.