Mapping the insurance data landscape end to end

Claims management systems. Track a claim from First Notice of Loss (FNOL) through settlement. Guidewire ClaimCenter and Duck Creek Claims dominate the US market. Claims data is the golden record for loss amounts, reserves, and settlement dates, even though policy data feeds into it.

Billing and premium accounting systems. Track what was actually invoiced and collected, which frequently diverges from what the policy says was due, due to mid-term changes, cancellations, or grace periods.

Reinsurance systems. Track ceded business, meaning the portion of risk an insurer passes to a reinsurer (a company that insures insurers) like Munich Re or Swiss Re. These systems own the golden record for treaty terms and recoverables (amounts owed back to the ceding insurer).

Third-party bureau data. External enrichment sources: credit bureaus (Equifax, TransUnion), motor vehicle records, CLUE reports (Comprehensive Loss Underwriting Exchange, a US claims history database run by LexisNexis), and catastrophe models from vendors like Verisk or Moody's RMS. None of these are golden records for internal fields, but they're often the *only* record for external risk facts, like a prior undisclosed claim.

Telematics and IoT data. Usage-based insurance (UBI) programs like Progressive's Snapshot or State Farm's Drive Safe & Save generate raw driving-behavior streams (braking, speed, mileage) that get aggregated into risk scores before hitting the warehouse.

Distribution and CRM data. Broker and agent portals, quote engines, and call center logs. Often the messiest source, because entry quality varies by intermediary.

Regulatory data layers worth knowing

In the US, NAIC (National Association of Insurance Commissioners) statutory filings force standardized reporting on loss reserves and premiums, which is why US carriers maintain a parallel "statutory data mart" distinct from their GAAP (Generally Accepted Accounting Principles) reporting data.

In Europe, Solvency II (the EU's risk-based solvency regulation, in force since 2016) requires granular data lineagedata lineageData lineage maps how data moves and transforms across systems, from origin to consumption, showing where it came from, what changed it, and where it goes.View full definition → documentation for anything feeding capital calculations, which is why European insurers invest heavily in data governancedata governanceData governance is the set of policies, roles, and processes that ensure data is accurate, secure, well-defined, and used responsibly across an organization.View full definition → tooling specifically for regulatory reporting, overseen by EIOPA (European Insurance and Occupational Pensions Authority).

Who owns the golden record?

This is the single most useful skill in this lesson: for any given field, know which system is authoritative.

| Field | Golden record owner | Why |

|---|---|---|

| Coverage limit | Policy admin system | It's the contract |

| Claim reserve amount | Claims system | Adjusters update it live |

| Customer address | Usually CRMCRMCustomer Relationship Management: software and strategy to manage and analyse customer interactions throughout their lifecycle.View full definition →, sometimes PAS | Depends on carrier's MDMMDMMaster Data Management (MDM) is the discipline of creating and maintaining a single, consistent, trusted version of an organization's core business entities like customers, products, and suppliers.View full definition → setup |

| Prior claims history | Bureau (CLUE) | Internal systems only see their own history |

| Ceded loss amount | Reinsurance system | Treaty terms live there |

| Credit-based insurance score | Credit bureau feed | Computed externally, ingested read-only |

When two systems disagree, this table is your first troubleshooting step. A mismatch between PAS premium and billing system premium isn't necessarily an error; it might mean a mid-term endorsement hasn't synced yet.

How it consolidates: the pipelinepipelineAll active sales opportunities across the stages of the sales process, together with their combined potential value and probability of closing.View full definition →

Most large carriers now run a variant of this flow:

CDC (Change Data Capture) is increasingly replacing nightly batch ETLETLETL (Extract, Transform, Load) is a data integration process that pulls data from sources, reshapes it into a consistent format, and writes it into a target system.View full definition → (Extract, Transform, LoadExtract, Transform, LoadETL (Extract, Transform, Load) is a data integration process that pulls data from sources, reshapes it into a consistent format, and writes it into a target system.View full definition →), because claims adjusters and underwriters now expect near-real-time visibility. A basic validation rule that runs at the landing zone stage might look like this:

This kind of rule catches integration bugs (a claim linked to the wrong policy version) before they poison downstream reserving.

Data qualityData qualityThe degree to which data is fit for purpose: accurate, complete, consistent, timely, valid and unique. Poor quality data undermines analytics, reporting and AI.View full definition → metrics that matter

Generic data-quality frameworks (completeness, accuracy, timeliness, consistency) apply here, but insurance gives them sharp, measurable teeth.

Completeness rate: percentage of required fields populated. A common target for critical claims fields (loss date, cause of loss) is above 98 percent; below that, actuarial reserving models start producing unreliable output.

Match rate in entity resolution: when MDMMDMMaster Data Management (MDM) is the discipline of creating and maintaining a single, consistent, trusted version of an organization's core business entities like customers, products, and suppliers.View full definition → tries to link "John A. Smith" across CRMCRMCustomer Relationship Management: software and strategy to manage and analyse customer interactions throughout their lifecycle.View full definition →, PAS, and a bureau feed, carriers typically track what fraction resolve automatically versus need manual review. Industry practitioners often cite automated match rates in the 80 to 95 percent range as healthy, though this varies enormously by carrier and is not a universal benchmark, treat any specific figure as an estimate.

Lag time / latency: how long between an event (a claim payment) and its appearance in the warehouse. Regulatory reporting under Solvency II or NAIC statutory statements typically demands monthly or quarterly finality, but internal reserving increasingly wants daily or near-real-time feeds.

Reconciliation break rate: the percentage of records that don't tie out between two systems that should agree, for example, premium booked in PAS versus premium recognized in the general ledger. Finance and actuarial teams track this monthly; a rising break rate is an early warning of an integration failure.

Lineage coverage: the percentage of reportable fields with documented, auditable lineage back to source. Solvency II examiners in Europe specifically test this during model validation.

A simple worked example: if a carrier processes 500,000 claims a year and completeness testing finds 9,500 missing a required "cause of loss" code, completeness rate = (500,000, 9,500) / 500,000 = 98.1 percent. That sits right at the common internal threshold, worth flagging to the data governancedata governanceData governance is the set of policies, roles, and processes that ensure data is accurate, secure, well-defined, and used responsibly across an organization.View full definition → committee rather than ignoring.

Governance structures behind the metrics

Most carriers formalize this with a Data Governance Council, typically pairing a Chief Data Officer with actuarial, compliance, and IT leads, and a data stewardshipdata stewardshipA business-side owner responsible for the quality, consistency and appropriate use of data in their domain.View full definition → model where named business owners (not IT) are accountable for specific domains, like "claims data stewarddata stewardA business-side owner responsible for the quality, consistency and appropriate use of data in their domain.View full definition →" or "policy data stewarddata stewardA business-side owner responsible for the quality, consistency and appropriate use of data in their domain.View full definition →." This is standard practice at large carriers and is increasingly required, not optional, given both Solvency II documentation demands in Europe and heightened NAIC model governance expectations in the US following the growth of AI-based underwriting tools.

For a practical primer on data qualitydata qualityThe degree to which data is fit for purpose: accurate, complete, consistent, timely, valid and unique. Poor quality data undermines analytics, reporting and AI.View full definition → dimensions applicable across industries, see the DAMA-DMBOK framework overview, a widely referenced standard in data management.

Key Takeaways

• Insurance data comes from structurally different sources (policy admin, claims, reinsurance, bureau, telematics), each with its own golden record; disagreements between systems are often timing issues, not errors.
• Build the habit of asking "which system owns this field?" before troubleshooting any data discrepancy; it's the fastest diagnostic shortcut in the industry.
• Data qualityData qualityThe degree to which data is fit for purpose: accurate, complete, consistent, timely, valid and unique. Poor quality data undermines analytics, reporting and AI.View full definition → metrics that matter most operationally: completeness rate, entity-resolution match rate, latency, and reconciliation break rate; track these monthly, not annually.
• Regulatory frameworks (NAIC in the US, Solvency II and EIOPA in Europe) increasingly mandate documented data lineagedata lineageData lineage maps how data moves and transforms across systems, from origin to consumption, showing where it came from, what changed it, and where it goes.View full definition →, which is reshaping governance from a nice-to-have into an audit requirement.