The conversation about data quality in insurance has a framing problem.
For as long as most people in the industry can remember, the case for investing in better data has been made primarily in defensive terms. Better data reduces errors. Better data improves compliance. Better data reduces the risk of regulatory action. Better data means fewer problems.
All of that is true. And none of it is particularly compelling to a leadership team weighing up where to allocate a finite budget against a set of competing priorities, because “fewer problems” is a difficult thing to put a number on and an even more difficult thing to celebrate.
What doesn’t get said often enough – clearly, directly, with the commercial weight it deserves – is that better data in insurance doesn’t just protect the business. It generates revenue, improves margins, and creates competitive advantages that compound over time in ways that are genuinely significant.
The defensive framing has undersold the case for years. It’s worth making the commercial one properly.
What bad data actually costs – beyond the obvious
Before making the positive case, it’s worth being honest about the full cost of poor data quality, because it tends to be significantly larger than the visible parts suggest.
The visible costs are the ones that appear in audit findings, regulatory correspondence, and remediation projects. They’re real and they’re measurable, which is why they tend to dominate the conversation. But they represent the surface of a much larger iceberg. Gartner’s research puts the average annual cost of poor data quality at $12.9 million per organisation across all industries – a figure that includes both the visible remediation expense and the harder-to-trace costs of decisions made on incomplete information[1].
The invisible costs are in the decisions that get made on incomplete or inaccurate information – and in insurance, where the quality of decisions about risk selection, pricing, and portfolio construction has direct and immediate P&L consequences, the financial impact of systematically poor data quality is substantial.
An underwriter pricing a risk without accurate claims history attached to the account is making a decision with less information than the organisation theoretically has. The premium that results may be too low to reflect the actual risk profile – not because the underwriter made an error, but because the data environment didn’t give them what they needed to make a fully informed decision. Multiply that across a book of business and the cumulative impact on loss ratios is not a rounding error.
A portfolio manager trying to understand concentration risk from data that isn’t properly integrated across systems is working with a partial picture. The concentration they’re managing against may not be the concentration that actually exists. The exposure they believe they’ve diversified away from may still be there, just invisible in the data.
A renewals team working from customer data that hasn’t been properly maintained is making pricing and retention decisions on a profile that may be significantly out of date. The customer they think they’re renewing and the customer they’re actually renewing may be meaningfully different – with consequences for both retention rates and loss experience.
These aren’t theoretical scenarios. They’re the operational reality of running an insurance business on data that hasn’t been properly managed. And the financial consequences – in premium inadequacy, in unexpected loss emergence, in retention decisions that optimise against the wrong variables – accumulate quietly and persistently in ways that are difficult to trace back to their source without looking carefully.
The pricing advantage that better data creates
Pricing is where the commercial case for better data in insurance is most direct – and where the advantage of having genuinely good data over merely adequate data is most tangible.
Insurance pricing is fundamentally a statistical exercise. The accuracy of the price depends on the quality of the information used to set it. Better information about a risk – more complete claims history, more accurate exposure data, more reliable information about the factors that correlate with loss – produces prices that more accurately reflect the actual risk profile. Which means, in a competitive market, the ability to price more accurately than competitors is a direct commercial advantage.
The numbers behind this are striking. According to Capgemini’s World Property and Casualty Insurance Report, 83% of insurance executives believe predictive models are critical to the future of underwriting, and carriers using predictive underwriting models report 15-25% improvements in loss ratios within the first 18 months[2]. Industry research has also pointed to 3-5 percentage point reductions in loss ratios, 10-15% increases in new business premiums, and 5-10% improvements in retention rates for profitable segments where digitised underwriting is properly implemented[3].
This works in two directions. On risks where your data tells you the profile is better than the market believes, you can price competitively and win business that is genuinely profitable. On risks where your data tells you the profile is worse than the market believes, you can avoid the business that looks attractive on the surface but will deteriorate on the loss account. Over time, the portfolio that results from systematically better pricing information is materially different from the one that results from pricing on data of average quality.
The insurers that have invested in data quality – in making sure claims data is properly attached to the right accounts, that exposure information is accurate and current, that the enrichment data used in pricing models is reliable and well-maintained – are not just reducing errors. They are building a pricing capability that produces better outcomes than competitors working with inferior information.
In a market where pricing accuracy is a significant driver of underwriting profitability, that’s not a marginal advantage. It’s a structural one.
How better data improves retention – and why that matters commercially
The connection between data quality and retention is less obvious than the connection between data quality and pricing, but it’s equally significant – and in some ways more valuable because the commercial impact of retention compounds in a way that pricing improvements don’t always.
Retention decisions in insurance are based on a combination of loss experience, relationship quality, and pricing adequacy. All three of those inputs depend on data. Loss experience needs to be accurately attributed to the right account. Relationship quality depends on having a complete and current view of the customer across all touchpoints. Pricing adequacy can only be assessed if the underlying risk data is reliable.
When data quality is poor, retention decisions get made against an incomplete or inaccurate picture. The account that looks like a good retention risk on the basis of the data available may have a loss history that isn’t properly reflected. The customer who looks price-sensitive may actually be in a different risk category than their current pricing reflects. The relationship that looks straightforward may have complexity that isn’t visible in the data.
The consequence is a retention strategy that optimises against the wrong variables – retaining business that will deteriorate and losing business that would have been profitable, because the data needed to distinguish between them wasn’t good enough.
Better data doesn’t just improve the accuracy of retention decisions. It changes the profile of the book that results from them. Over several underwriting cycles, the difference between a retention strategy built on good data and one built on poor data shows up clearly in loss ratios and in the quality of the portfolio that remains after natural attrition.
The fraud detection dividend
Fraud is one of the areas where the financial case for better data is most concrete and most immediate – and where the investment in data quality pays back most quickly.
The numbers tell the story. The Association of British Insurers reported £1.16 billion of detected fraudulent general insurance claims in 2024, with insurers uncovering over 98,400 fraud-related cases – a 12% rise on the previous year[4]. The Insurance Fraud Bureau estimates that the true figure including undetected fraud exceeds £3 billion annually, adding around £50 to every policyholder’s premium[5].
What that scale represents, beyond the obvious cost, is opportunity. Insurance fraud is disproportionately concentrated in organisations where the data environment makes it easier to perpetrate and harder to detect. Fraudulent claims exploit the gaps between systems, the inconsistencies in data, and the absence of the cross-referencing that would make patterns visible.
A well-integrated, well-structured data environment creates detection opportunities that a fragmented one doesn’t. Patterns across claims that share characteristics – in timing, in the parties involved, in the circumstances described – become visible when claims data from different systems can be analysed together. Connections between claimants, intermediaries, and loss events that would be invisible in siloed data become apparent when the data is properly integrated. Anomalies that would be lost in the noise of a poorly organised dataset surface more clearly when the underlying data is clean and consistent.
The carriers that have invested in this infrastructure are pulling ahead of those that haven’t. Aviva detected 14% more fraudulent claims in 2024 than the year before, uncovering more than 12,700 suspect claims worth £127 million; in the first half of 2025 alone it stopped over 6,000 fraudulent claims totalling £60 million[6]. Allianz UK detected £92.6 million of fraudulent activity in the first half of 2025, a 34% year-on-year increase[7]. What links these results is not the analytics platform itself but the integrated, structured data environment that the analytics depends on.
The financial return on improved fraud detection is direct and measurable in a way that some of the other benefits of better data are not. Claims that would have been paid fraudulently don’t get paid. The cost of investigating suspicious claims reduces because the signals that trigger investigation are more reliable. And the deterrent effect of demonstrably better detection capability reduces the frequency of attempts over time.
For insurers that have invested in data integration and quality specifically with fraud detection in mind, the returns have in many cases justified the investment many times over – which makes it one of the most straightforward commercial cases for better data that exists in the industry.
Portfolio management that sees what others miss
The cumulative effect of better data on portfolio management is perhaps the most strategically significant commercial benefit – and the hardest to articulate to stakeholders who haven’t experienced what genuinely good portfolio data looks like.
Most insurance portfolios are managed against data that is, to varying degrees, incomplete. Exposure data that doesn’t fully reflect current risk profiles. Aggregation data that doesn’t capture all the relevant concentrations. Performance data that reflects what the systems show rather than the full picture of what the portfolio actually contains. Decisions about where to grow, where to reduce, and where to hold are made against this partial picture – which means they’re optimising against an approximation of reality rather than reality itself.
The portfolios that are managed most effectively – that consistently achieve better risk-adjusted returns over time – are typically the ones where the management team has the most complete and accurate view of what they actually hold. Not the most sophisticated models, not the most experienced underwriters, not the most conservative risk appetite – the best data. Because the quality of every other input into portfolio management decisions depends on the quality of the data those decisions are based on.
This shows up in practice in the ability to identify deteriorating segments earlier, to allocate capacity to genuinely attractive risk rather than apparently attractive risk, and to respond to market changes with a more accurate understanding of how the existing portfolio is positioned than competitors who are working from inferior information.
Where Dajon helps insurers make the commercial case real
At Dajon Data Management, we work with insurance organisations on the data challenges that sit underneath the commercial outcomes described above – making the connection between data quality and financial performance concrete rather than theoretical.
That means integrating data across the systems that have historically operated in silos, so that the complete picture of risk, customer, and performance that better decisions require is actually available. It means cleansing and standardising data that has accumulated inconsistencies over time, so that the models and analyses that depend on it are working from a reliable foundation. And it means building data environments that maintain quality over time rather than degrading gradually back towards the state that prompted the remediation in the first place.
The goal isn’t cleaner data for its own sake. It’s the pricing accuracy, the portfolio visibility, the fraud detection capability, and the retention intelligence that cleaner data makes possible – and the commercial outcomes that follow from them.
The framing this deserves
Data quality in insurance has been sold as risk management for too long.
It is risk management. But it is also – and more compellingly for the people who control the budgets – a commercial capability. One that improves margins, generates revenue, and creates sustainable competitive advantages in a market where the quality of decisions about risk is the most direct driver of financial performance.
The insurers that have understood this – that have invested in their data not just to avoid problems but to gain advantages – are not in a different category of sophistication from those that haven’t. They’ve just made a different decision about how to frame the value of what they were building.
Better data doesn’t just protect the business.
In insurance, it builds it.
Is your data quality a defensive measure – or a commercial asset?
Dajon Data Management helps insurance organisations build the data foundation that turns data quality from a compliance consideration into a commercial advantage. Get in touch to understand what better data could mean for your business specifically.
References
- Data Quality: Why It Matters and How to Achieve It Gartner[↩]
- Data Analytics for Insurance: A Strategic Guide Vantage Point[↩]
- The Increasing Significance of Analytics in the Insurance Industry Decerto[↩]
- Fraudulent insurance claims continue to top £1 billion ABI[↩]
- Insurance fraud AI: UK detection systems & legal impact Browne Jacobson[↩]
- Fraud on the rise but fraudsters facing the consequences Aviva[↩]
- Insurance fraud in UK climbs to record £1.16 billion – ABI Insurance Business[↩]