Home > Insurance > Reducing Claims Leakage: Where to Start and What to Measure

A practical guide for claims and operations leaders at mid-to-large insurers and MGAs.

Perceptive Analytics Point of View: You Cannot Fix What You Cannot See
Most claims leaders know they lose money to leakage. Very few know where those losses happen, how large they are, or how much they can recover. When a VP of claims enters a board meeting with a combined ratio over 100, directors ask where the money went. Without a clear measurement framework, the only honest answer is: we are not sure.
Perceptive Analytics works with insurers and MGAs who have tried the obvious fixes — tightening sign-off limits, setting up a special investigations unit, changing vendors. These actions help on the edges. To cut leakage permanently, you need a stronger foundation: a data baseline, clearly defined leakage categories, and regular measurement that surfaces the hidden problems. This guide shows you how to start.

A regional property and casualty insurer recently asked its claims team why its loss adjustment expense (LAE) ratio rose four percentage points over three years when claim volume stayed flat. The team assumed rising litigation costs caused the spike. They were wrong. A formal audit showed that one-third of the extra costs came from missed subrogation, stale reserves left open, and inconsistent coverage checks at settlement. None of these issues appeared in the monthly operational reports.

Claims leakage is an old issue, but the pressure to address it is growing. U.S. property and casualty insurers lost $26.9 billion in underwriting in 2022 — the worst result since 2011. Incurred losses and expenses grew 14.1% while premiums grew only 8.3% [Deloitte, 2024]. Combined ratios remain above 100 for most personal and commercial products. In this environment, recovering even one percent of loss payments makes a significant financial difference.

This guide is written for claims, operations, and finance leaders who need to measure and reduce their leakage. It covers the main causes, the claim stages where losses accumulate, the metrics you should track, and a 10-step plan for getting started without buying expensive new software. Perceptive Analytics brings together advanced analytics consulting, AI consulting, and data infrastructure expertise to help insurers and MGAs build the measurement foundation that makes leakage reduction durable. You can explore our broader insurance analytics approach in our data-driven blueprint for growth in the insurance industry and our analysis of how AI is rewiring the insurance claim process.

$26.9B U.S. non-life underwriting loss (2022) — largest since 2011 (Deloitte Global Insurance Outlook, 2024)	~10% P&C insurance losses attributable to fraud annually (Deloitte 2025)	20–40% Potential savings from fraud-related leakage through multimodal analytics (Deloitte 2025)

Talk with our consultants today. Book a session with our experts now. → Schedule Your Free 30-Minute Session with Perceptive Analytics

1. Understanding the Main Drivers of Claims Cost Leakage

Claims leakage is the gap between what you pay on a claim — including payments to claimants and administrative costs — and what you should have paid if you followed your policy rules and claims guidelines correctly. This gap stems from several distinct issues. Treating them as a single problem leads to ineffective fixes and misdirected investment.

Indemnity Leakage

Indemnity leakage happens when you pay too much to settle a claim. This includes paying excessive medical bills due to inadequate independent medical examinations, failing to adjust settlements for pre-existing injuries, or ignoring asset depreciation during property valuations. In liability cases, adjusters often fail to calculate shared fault properly. In property lines, poor damage assessments at the start lead to expensive mid-repair adjustments that are difficult to reverse once committed.

Expense Leakage: LAE and Vendor Costs

Loss adjustment expenses are frequently undermonitored. You lose money when you pay defense lawyers who do not follow agreed rate structures, hire expert witnesses without verifying necessity, or manually manage simple claims that could be automated. Insurers with undertrained staff see operational costs rise by roughly 12%, while unskilled adjusters pay out up to 20% more in claims settlements [Deloitte, 2025]. Perceptive Analytics’ Talend consulting and data engineering consulting practices build the data pipelines that make vendor billing and LAE monitoring visible in near real time — rather than surfacing in a quarterly financial review after the money has already been paid.

Fraud and Build-Up

Built-up claims — where claimants pad real losses with fabricated expenses — are far more common than completely invented claims. You will struggle to catch these without data filters applied at the initial notice of loss. According to the Coalition Against Insurance Fraud, insurance fraud steals at least $308.6 billion a year from American consumers. The challenge is that most fraud detection still operates reactively — after payment — rather than at the point where intervention is cheapest and most effective. Perceptive Analytics’ AI consulting practice builds the pre-payment scoring models that shift fraud detection upstream, reducing both the frequency and the cost of fraudulent payouts.

Subrogation and Salvage Recovery Gaps

Subrogation leakage happens when you pay for damage caused by a third party but fail to recover those funds. This occurs because staff did not identify the recovery opportunity early, or because the recovery was deprioritized and forgotten once the claim was closed. Property salvage is another consistently missed area — particularly when organizations do not have systematic processes for recovering value from total loss vehicles or water-damaged assets.

Coverage and Data Errors

Paying claims that exceed policy limits or fall under the wrong coverage section is entirely preventable — but it happens regularly when policy data at intake is incomplete or incorrect. MGAs often carry coverage errors in their spreadsheets after binding, leaving adjusters making payout decisions based on bad policy data. Poor data at intake makes every subsequent step of the claims process less accurate and more expensive. Our how automated data quality monitoring improved accuracy and trust across systems case study documents how systematic data quality monitoring at intake prevents exactly these downstream errors.

2. Claims Process Areas Most Prone to Inefficiencies

Leakage does not happen evenly across the life of a claim. Identifying the high-risk stages helps you focus your review effort where the financial return is greatest.

FNOL and Intake

The intake stage is where you set up the claims file — and where most preventable errors begin. Failing to verify coverage immediately, record initial statements accurately, or collect liability details at the start creates confusion that increases downstream costs at every subsequent stage. Collecting structured, validated data at intake improves every decision that follows. Perceptive Analytics’ chatbot consulting services support the intelligent intake layer that captures structured FNOL data automatically — reducing the dependency on adjuster discipline for information that should be system-enforced.

Triage and Segmentation

Without systematic sorting at intake, complex and simple claims end up in the same work queue — assigned by claim type alone rather than by complexity, fraud risk, or recovery potential. A minor injury claim with clear fraud indicators may go to an adjuster who typically handles straightforward home damage. When complex claims lack the right resources, they take longer, escalate to litigation, and cost significantly more to settle. Simple claims assigned to senior staff waste operational budget that could be directed at high-risk files.

Reserving

IBNR reserve adequacy is an actuarial discipline, but case reserve accuracy is a claims management discipline. Leaving old estimates open on claims — especially for long-term injury cases — hides your true cost exposure and makes budgeting unpredictable. When adjusters lack the tools to update reserves as new facts emerge, expenses accumulate invisibly until the fiscal year-end actuarial review surfaces the problem when it is most difficult to correct.

Medical Management and Litigation

For bodily injury and workers’ compensation claims, close medical management is essential — reviewing treatment plans, checking provider network compliance, and organizing independent medical reviews to control indemnity costs. Your litigation rate is one of the clearest signals of claims management quality. Once a claimant retains legal representation, defense costs escalate rapidly. Intervention protocols at defined intervals — typically 90 and 180 days — are a standard practice among top-performing claims operations.

Settlement and Payment

Signing off on settlements without clean data or formal review authority creates permanent financial losses. Payment errors — duplicate checks, uncollected deductibles, weak release language — seem minor in isolation but accumulate to significant sums across thousands of files. These errors are also largely invisible in aggregate financial reporting, which is why they persist.

Perceptive Analytics Point of View: Triage Is Not a Technology Problem
Many insurers believe they must install a new core system before they can improve claim routing. They do not. The companies that reduce leakage fastest start by grouping claims using data they already own — claim type, severity indicators, coverage flags, and prior claims history — then enforce this sorting through staff guidelines rather than software.
Technology speeds up and scales a strong workflow. Putting a machine learning model on top of poor intake processes only helps you make bad decisions faster. Fix the process first, establish the baseline, then automate the work. This approach builds a claims department that improves steadily — rather than one that oscillates between costly IT projects and temporary fixes.

3. How Leading Insurers Tackle Claims Cost Overruns

Top-performing insurers do not use complex or mysterious methods. They execute basic operational tasks exceptionally well, supported by data analytics that make discipline scalable rather than dependent on individual adjuster judgment.

Analytics-Driven Segmentation and Triage

Insurers that score claims at intake — assessing complexity, fraud risk, litigation probability, and recovery potential — outperform those relying entirely on staff judgment. Updating claims systems and workflows can cut claim costs by up to 30% and raise customer satisfaction by 20% [McKinsey, 2021]. Data tools improve operational efficiency by 25% across insurance operations [KPMG, 2023]. Perceptive Analytics builds the scoring infrastructure and BI reporting that makes segmentation-driven triage operational. Our Tableau consulting and Power BI consulting teams build the claims management dashboards that surface these scores at the point of assignment — not buried in a report that claims managers check weekly.

Special Investigations Unit Effectiveness

Top performers treat their fraud units as active warning systems rather than reactive investigators. They apply predictive models to live claims to trigger fraud alerts before making payments — a fundamentally different operating model from the traditional post-payment SIU referral approach. Nearly 60% of insurers plan to increase fraud software spending to lower loss ratios [Verisk, 2024]. Real-time analytics could save P&C insurers $80 to $160 billion by 2032 through improved fraud detection [Deloitte, 2024]. Perceptive Analytics’ advanced analytics consulting and Looker consulting capabilities support the model development and real-time monitoring layer that makes pre-payment fraud detection operational.

Vendor and Counsel Management

Legal panels and preferred vendor networks only generate savings if you audit their bills and track their performance consistently. Leading claims teams grade law firms on defense costs, cycle times, and cost per case. Auditing invoices from body shops, repair contractors, and medical providers catches unnecessary charges that otherwise accumulate unnoticed across thousands of files over time.

Continuous Reserve Review

Top claims teams review reserve adequacy on a strict schedule — typically at 30, 90, and 180 days for any claim above a defined dollar threshold — comparing cases against historical averages to identify outliers by adjuster or product line. Identifying reserve issues early, rather than at year-end, is one of the clearest operational signals of a disciplined claims organization. Perceptive Analytics’ Tableau development services and Power BI development services build the reserve adequacy dashboards that make these structured review intervals automatic rather than dependent on individual manager initiative.

Capability	Standard Practice	Leading Practice	Leakage Risk (Standard)
FNOL Triage	Manual queue assignment by claim type	Predictive scoring at intake; auto-segmentation	High: complexity misjudged at intake
Fraud Detection	Adjuster flags; post-payment SIU referral	Real-time analytics; pre-payment referral triggers	High: build-up paid without review
Reserve Adequacy	Annual actuarial review	30/90/180-day structured case review + IBNR alerts	Medium-High: stale reserves mask true exposure
Subrogation ID	Adjuster intuition at close	FNOL-triggered recovery scoring; automated pursuit	Medium: recovery opportunities missed
Vendor Billing	Invoice payment on receipt	Rate-schedule validation; automated audit rules	Medium: LAE inflation over time
Settlement Authority	Fixed dollar thresholds by adjuster grade	Data-informed guidelines; real-time reserve check	Medium-High: authority misapplied on complex claims

4. First Steps to Identify and Quantify Your Claims Leakage

Always begin with a baseline. Before you buy software, restructure your team, or hire consultants, you must know where your leakage happens and how much it costs. You can build this picture in three steps.

Step 1: Data Inventory and Quality Assessment

Most insurers have significant volumes of claims data — but very little of it is clean or analytics-ready. Start by auditing your claims systems, billing logs, adjuster notes, and third-party data feeds to verify that you have the fields your measurement models require. This work is tedious but necessary. It routinely reveals that staff leave intake forms partially blank, reserve histories are fragmented, or recovery data sits in an isolated system with no connection to the main claims environment. Perceptive Analytics’ Snowflake consulting and Talend consulting teams conduct exactly this kind of structured data inventory as the first phase of any claims analytics engagement — treating it as a prerequisite for measurement, not an optional diagnostic.

Step 2: Leakage Audit and Sampling

To perform a leakage audit, pull a random sample of 200 to 500 closed claims for each major product line and review them against clear standards covering indemnity payouts, expenses, subrogation, fraud indicators, and coverage rule compliance. This audit gives you a defensible leakage percentage to present to your board and CFO — one that reflects actual file-level review rather than aggregate financial inference.

Step 3: Identify Quick-Win Focus Areas

Not all leakage categories are equally addressable within a 90-day window. Rewriting litigation guidelines requires extensive staff retraining and months of implementation. Other changes — setting up recovery flags in your existing claims system or establishing a structured reserve review cadence — can show results in weeks. Rank your leakage categories by financial impact and implementation ease, then target the highest-value, lowest-friction areas first to build momentum and credibility with leadership before tackling the more complex changes.

Perceptive Analytics Point of View: The Audit Is Not the End — It Is the Beginning
Many organizations treat a leakage audit as a one-time project. They order the audit, read the report, agree on a few changes, and file it away. Six months later, the same bad habits return because no monitoring process was established to sustain the improvement.
An audit identifies your losses — but it does not stop them. Insurers that successfully reduce leakage use audit results to build a permanent tracking system: metrics matched to each leakage type, clean dashboards for managers, and quarterly reviews to check progress against baseline. This continuous tracking is what separates organizations that control leakage from those that simply study it periodically.

5. Benchmarks and KPIs to Track Claims Costs and Leakage

Selecting the right KPIs is less about breadth and more about alignment. Choose metrics that map to your specific leakage risks from the audit. A 30-metric dashboard that does not prompt management action is less useful than five or six focused metrics that do. Perceptive Analytics builds the claims KPI dashboards — through Tableau implementation services, Power BI implementation services, and marketing analytics frameworks — that make these metrics operationally visible to claims leadership in real time. Our frameworks and KPIs that make executive Tableau dashboards actionable and answering strategic questions through high-impact dashboards articles explain the design principles that make claims dashboards trusted and used — rather than built and ignored.

KPI	Definition	Industry Benchmark	Primary Leakage Signal
Loss Ratio	Incurred losses / earned premiums	< 65% P&C personal lines target; 2022–2023 industry average ~72–75%	Overall claims cost control
LAE Ratio	Loss adjustment expenses / earned premiums	Varies by line; upward trend is a red flag	Expense leakage; vendor and litigation costs
Combined Ratio	Loss ratio + expense ratio	< 100 for underwriting profit; industry ran 101.7–102.5 (2022–2023)	Aggregate efficiency signal
Average Paid Severity	Total paid losses / number of closed claims	Benchmark against prior year trend and NAIC line data	Indemnity leakage; settlement quality
Claim Cycle Time	Days from FNOL to claim closure	Personal auto: 15–30 days straightforward; complex: 90–180+	Process efficiency; litigation risk
Litigation Rate	% of open claims with legal representation	BI personal auto: industry range 15–30%; target < 20%	Complexity, fraud, settlement quality
Subrogation Recovery Rate	Subrogation recovered / eligible paid losses	Leading performers: 8–12% of applicable losses	Subrogation and salvage leakage
Leakage Rate (Audit)	Leakage identified / sampled paid losses	Industry audit findings: 5–15% of paid losses	Direct leakage quantification
Reserve Development Ratio	Paid vs. initial reserve at close	Target < 10% adverse development	Reserve adequacy; IBNR accuracy
SIU Referral Rate	Claims referred to SIU / total claims	Rising trend signals growing fraud exposure	Fraud and build-up leakage

Not every KPI in this table needs to appear in your monthly management pack from day one. A practical starting point is three to five metrics aligned with your highest-priority leakage categories from the audit — establish the baseline, and build from there as your measurement discipline matures.

6. Bringing It Together: A 10-Point Starter Checklist

Use this checklist as a planning tool to guide your team through a 90-day diagnostic and baseline phase. The goal is clear measurement — not perfection.

Category / Step	Action
Data (Step 1)	Audit your claims system, billing logs, and vendor feeds. Find where data is missing at intake and fix those gaps first.
Data (Step 2)	Define your leakage categories (indemnity, expenses, fraud, subrogation) and list the specific database fields needed to track each one.
Process (Step 3)	Map your claims process to identify the three or four steps where staff are most likely to make payment or administrative errors.
Process (Step 4)	Review how you assign incoming claims. If you assign by claim type alone without scoring for complexity, make this your first process fix.
Measurement (Step 5)	Pull a random sample of 200 to 500 closed claims per major product line. Audit them using a single set of scoring rules to establish your baseline leakage cost.
Measurement (Step 6)	Select five to seven metrics from the KPI table above. Calculate your baseline using 12 months of historical data and set a first-year target of 10% improvement.
Technology (Step 7)	Check whether your current software can generate the metrics you need. You rarely need a new system to run a diagnostic review — you need clean data and a governed reporting layer.
Governance (Step 8)	Schedule quarterly leakage reviews and assign each leakage category to a specific manager responsible for hitting the target.
People (Step 9)	Evaluate adjuster skills in key areas — reserving, coverage limit review, and third-party recovery identification — and design training around the specific gaps the audit reveals.
Quick Wins (Step 10)	Identify two or three immediate changes — auditing contractor invoices or adding recovery alerts in the existing claims system — to demonstrate early savings and build momentum.

Perceptive Analytics helps insurers and MGAs design and execute claims leakage diagnostics, build the measurement frameworks that make leakage visible, and develop the analytics capability to track and close the gap over time. Our full delivery capability for claims analytics programs spans advanced analytics consulting, AI consulting, Snowflake consulting, Talend consulting, and BI delivery through Tableau expert, Power BI expert, Tableau developer, and Tableau partner company capabilities. Our insurance sales dashboard, unified CXO dashboards in Tableau, and data observability as foundational infrastructure resources provide the operational reference points for what a mature claims analytics environment looks like when it is fully operational.

Perceptive Analytics Point of View: The First 90 Days Are About Visibility, Not Transformation
Resist the temptation to solve the problem before you have fully defined it. The instinct to move quickly to solutions — new software, reorganized teams, outsourced functions — is understandable under board pressure. But launching solutions before you have a verified baseline is the fastest route to spending money on the wrong things.
The value of the first 90-day diagnostic is not the leakage number itself. It is the credibility that number gives you to drive change. A well-executed audit, with clear methodology and defensible data, converts a board conversation from “we think we have a leakage problem” to “we have identified $X million in recoverable leakage in these specific categories, and here is our phased remediation plan.” That is a fundamentally different conversation — and it gets a fundamentally different response.

Every insurer has claims leakage. The ones that outperform their peers over time make it measurable, assign ownership, and close the gap systematically. That begins with a baseline. It continues with the discipline to track, review, and improve — quarter after quarter, not just after the next board presentation.