The Hidden Tax on Transformation

Written By Daniel Pears

Why Insurance Migration Costs Shape More Decisions Than We Admit

Every insurance modernisation conversation begins in the same place. A clean platform. Better analytics. Something cloud-shaped. The future is always twelve to eighteen months away, on a slide with a gradient background. The conversation that decides whether any of it actually happens is the other one. The one about what it costs to move. It tends to take place in a smaller room, with worse biscuits

Policy migration, the work of lifting books of business from one system to another, carries a transaction cost the industry is oddly polite about. In the UK especially, where legacy estates run deep and regulators expect every data field to be traceable, that cost does not just dent project budgets. It quietly shapes strategic decisions. Which is how modernisation turns into million-pound archaeology.

The Problem in Plain Terms

At its core, policy migration means extracting every active and historic policy record, cleansing it, transforming it, validating it, and loading it into something new. Every endorsement. Every claims linkage. Every underwriting note. Every workaround someone built in 2007 because the original spec did not contemplate a product that later became 40 percent of the book. For a mid-size UK insurer carrying a few hundred thousand policies, that's not a weekend job. It is not a quarter job. It is the kind of job where people develop hobbies they did not have when it started.

The industry numbers are grim. Gartner research suggests 83 percent of data migration projects fail, overrun their budgets, or overshoot their timelines. Bloor Group puts average cost overruns at 30 percent and time overruns at 41 percent. These are cross-sector figures, but insurance is where they go to retire. Policy data here carries decades of business logic, bespoke product rules, and regulatory audit trails that cannot be approximated. You cannot round off a claims linkage. Someone will notice, and they will almost always work at the FCA.

Large insurer transformations can run north of $5 million and stretch across 12 to 18 months. That is when things go relatively well. The UK has its own cautionary tales, programmes where investments north of £300 million were quietly shelved without delivering on their promise. Those are not budget overruns. Those are expensive, late-career educations in what was actually sitting inside the business.

What High Migration Costs Actually Do to Decision-Making

High, unpredictable migration costs do not slow transformation down. They bend it. The destination stays the same on the slide. The route to it gets rerouted quietly, in the margins of the operating plan.

Carriers stick with platforms they have outgrown because the perceived cost of leaving is too daunting. Product innovation stalls because launching on a new system means migrating the old book first, and no executive wants to be the one who signed off on that particular business case. Mergers and acquisitions turn awkward the moment someone asks, with real curiosity, what it would actually cost to integrate the target’s policy estate. In the London Market, that question has ended more than one deal and delayed many others. Legacy estates there sit in stratified layers, each generation of underwriting preserved like a fossil record. Fascinating if you like that kind of thing. Terrifying if you have to move it.

The London Market is the case study. Lloyd’s Blueprint Two, designed to move the market from paper-driven legacy processes to a digital backbone, has identified over £800 million in potential cost reductions. DXC’s recent migration of London Market infrastructure involved moving 70 billion rows of data and more than 200 business applications. Those numbers explain why the plan exists, and why executing it is so hard.

McKinsey’s finding that insurers with modernised technology stacks report 41 percent lower IT costs per policy points to a frustrating irony. The destination is clearly better. The journey to get there carries so much cost and risk that many carriers never set off. They keep maintaining what they have and hope the problem ages out. The problem does not age out. The problem gets a pension and goes on mainframe cruises.

The Economy Drivers Behind Migration Cost

So what actually drives the cost? A handful of factors converge, and naming them matters, mostly so we can stop being surprised by them every eighteen months.

  • Data complexity and quality. Insurance policy data is not a clean table. It is a sediment of business decisions encoded by the system that made them and comprehensible to nothing else. Experian research shows 64 percent of data migration projects go over budget. Data quality is almost always the iceberg, and the business case is almost always the Titanic.

  • Regulatory and compliance requirements. FCA oversight means migrated data must be demonstrably complete and accurate. You cannot cut corners on policyholder records when the regulator expects full traceability. That means validation cycles, sign-off gates, and specialist resource. None of which come cheap.

  • Platform and vendor lock-in. Guidewire, Sapiens, Duck Creek and the rest operate in a market projected to reach $31.6 billion by 2032. Extracting data from one ecosystem to move it to another is rarely a straightforward export. Proprietary data models turn what should be a data problem into an engineering problem (and, on a sufficiently long programme, an HR problem).

  • Undocumented business logic. Rules, workarounds, the “ask Janet, she knows” knowledge that lives only in code paths and in Janet, who retires in March. This is the line nobody puts in the business case and everybody discovers in sprint nine.

  • Organisational readiness. Migrations are not just technical. They require business sign-off on data mapping, product rationalisation, and the inevitable gap between what the old system held and what the new one can represent. This is the cost nobody quantifies because it shows up as tired people making rushed decisions, which historically has worked out fine.

What Comes Next

We will explore these drivers in depth in a follow-up post. They deserve more than a paragraph each. The companies navigating migration well are not the ones with the biggest budgets. They are the ones doing discovery first, before they commit, because they would like to know what they are walking into. Strange, perhaps, but it seems to help.

Metamorphic was built on a belief that is only mildly controversial. Legacy systems are not the enemy. They are where the business logic lives, where the decisions that made companies successful quietly sit, and where almost nobody is looking. The job is not to destroy and replace. The job is to understand what is there and make it usable, so that when migration does happen, it is informed, targeted, and far less likely to join the 83 percent that do not go to plan.

If you are staring down a migration decision, or halfway through one that is proving harder than the slides suggested, we would welcome the conversation. No hype. No hand-waving. Worse biscuits, perhaps, but better outcomes.

‍ ‍

References and Sources

Dan Pears is a solution architect and co founder of Metamorphic Services. His work sits at the intersection of legacy systems, data, and the very human reality of organisational change. After years spent helping large enterprises modernise the right way, he now focuses on making AI useful rather than theatrical. He writes about technology with one simple belief in mind: transformation starts with understanding, not disruption.

Daniel Pears
Next

AI First Doesn’t Mean Human Last