Dane Richards, chief technology officer of JMR Software, says the trick to data migration in the insurance (or any other) sector is the same as with anything else: you need to start with a plan. Then you need to follow that plan to ensure a successful project ensues.
Richards says: "There's plenty of technological innovation coming into the insurance space, particularly in the property, health and life insurance spaces. However, before companies can make the most of these innovations, which include automating activities such as submitting claims or onboarding new clients, they need to digitalise their data and move away from legacy processes and onto a platform that makes this data available to use."
He refers to a study by McKinsey and Company that says big data and machine learning is the most adopted technology in the insurtech sector, followed by usage-based insurance, IOT, gamification and robotics.
"One of the biggest challenges faced in the insurance industry is that a lot of existing players are far behind when it comes to aligning their data to the present. As much as they're aware that they need to move forward with these new technology innovations if they want to grow their business and improve the customer experience, they simply can't do so using their existing legacy systems."
As legislation requires that certain insurance policy data be maintained for a set number of years, this makes migrating from old to new platforms a mammoth and expensive exercise. However, the upside is that as a result, the insurance sector has access to huge amounts of internal data, which is a valuable asset that can be used to drive growth and customer experience, making it all the more important to migrate it quickly and accurately to the new platform.
The McKinsey study highlights growth and a desire to reduce costs as two key drivers behind the adoption of innovation in the insurance sector. "Insurance companies have existing data sets that they've accumulated over the years, but they need to be able to access this data, organise it and combine it with newer data in a meaningful way if they want to capitalise on big data initiatives such as analytics."
Richards identifies two challenges faced by players in the insurance space that are aiming to compete with the new insurtechs:
1. Companies in this sector traditionally have large and complex data sets; and
2. The data is in various formats and is spread across several platforms.
It's no surprise that a study by the Bloor Group reveals that that 80% of data migrations will run over time and over budget.
"There are several reasons for businesses resisting migrating their data, including fears that key data will be lost. Companies are also concerned that project delays will increase the cost of data migration, and they dread the downtime that could result if the system crashes during the migration," explains Richards.
However, there are steps that will help mitigate these and other risks during a data migration project. According to a data migration study by JMR Software and ITWeb, 38% of respondents support allocating internal resources to the project as they are familiar with the data and systems. A further 26% of respondents favour outsourcing to a data migration specialist. It's clear that collaboration with data migration experts is thought to contribute significantly to the success of any data migration project.
Richards discloses five key lessons learnt regarding data migration:
1. Plan your migration. This is, surprisingly, an oft-neglected point. While you might not know all of the details around exactly how your data will move across to the new platform, you still need to include it in your project plan so you can plan around it. You also need to include the experts on your system and the target system from the outset, so they can identify any risks earlier on rather than later. It's essential to do multiple iterations of your migration through the various stages; don't leave it until the testing phase. The load strategy also needs to be planned to account for any physical constraints that might inhibit the ability to move source data to the target platform.
2. Do a detailed mapping specification of where your data is coming from and where it is going to. This needs to be signed off to create a baseline for the project migration to which everyone is held accountable. You also need to allocate a small group of experts who represent the business and IT as both are affected by the migration.
3. Define measurable success criteria. You want to know upfront what exactly are your success criteria, so that once you've done your migration run, you have something rock solid to go back to, to prove that you've migrated your data correctly. It's also important that you don't skip any steps in the testing cycle.
4. Consistency. You need to consider how the migration will be executed and do it in such a way that you trigger the relevant business rules, flagging any potential issues before you go live. It's also advisable to reduce the human interaction cycles and automate the entire migration process, if possible.
5. Collaboration. Migration projects are very complex and there's a high likelihood of something overrunning or failing. You need to have a collaborative culture within your team. The project could not be seen as the target system versus the source system; everyone has to have the same goal.
Share