Value Stream Mapping for Insurance Underwriting: From Application to Policy Issuance Without the Stale-Rate Waiting Game

In insurance underwriting, the customer often experiences the process as a single question: How long will it take to receive a decision and policy? Internally, however, the application may move through brokers, intake teams, underwriting assistants, data providers, risk specialists, approvers, compliance reviewers and policy administration systems.

That fragmented journey creates a familiar pattern: modest hands-on work surrounded by extended queues. Applications age, rates become stale, information must be revalidated and underwriters spend time locating work instead of assessing risk.

Value Stream Mapping (VSM) makes this hidden system visible. It connects the Voice of the Customer: speed, clarity and certainty: with the Voice of the Business: profitable risk selection, regulatory compliance and efficient capacity. It also reveals where the Voice of the Process shows a mismatch between demand and performance.

This guide presents a worked VSM example for a commercial insurance underwriting stream, from application receipt to policy issuance.

1. Select the Right Underwriting Scope

The fundamental purpose of scope selection is to create a value stream narrow enough to measure and broad enough to explain the customer outcome.

Avoid mapping “all underwriting.” Instead, define:

• Product family: SME property insurance
• Submission channel: Broker portal
• Start point: Complete or incomplete application received
• End point: Policy issued or application formally declined
• Customer: Broker and insured
• Demand profile: 120 applications per working day
• Primary CTQs: Decision turnaround, quote validity and first-time-right documentation

Exclude unrelated streams such as claims, renewals and endorsements unless they directly affect the selected policy-issuance journey.

A strong project charter should also define stale applications. For this example:

Stale application: An open application older than 15 business days without a final decision.

This operational definition allows the team to measure the problem consistently rather than debating individual cases.

For broader guidance on project framing, use the Lean Six Sigma Practitioner’s Guide and the Project Scope Boundary Calculator.

2. Build the Current-State Map

The mapping team conducts a gemba walk with brokers, underwriting assistants, underwriters, approvers and policy administrators. The map should show the actual process, including manual workarounds, email loops, batch reviews and waiting between steps.

A typical current-state flow is:

Application received → Completeness check → Triage → External data collection → Risk assessment and pricing → Senior approval → Offer communication → Policy issuance

At each step, collect:

• Touch time or cycle time
• Queue and wait time
• Work in Process (WIP)
• First Pass Yield
• Rework percentage
• Number of staff assigned
• Systems and handoffs involved

Do not map only the documented procedure. The customer experiences the real process, including weekends, overnight queues, approval delays and missing-information loops.

3. Worked Current-State Example

The following example uses a hypothetical SME property underwriting team processing 120 applications per working day.

Available productive time is:

• 7.5 hours per employee per day
• 450 minutes per employee per day
• Demand: 120 applications per day

Therefore, the pacemaker takt time is:

Takt time = Available time ÷ Customer demand
450 minutes ÷ 120 applications = 3.75 minutes per application

Takt time is not the same as cycle time. It establishes the rhythm required to meet demand. If intake requires eight minutes per application, the team must use parallel capacity, automation or redesigned work to maintain flow.

Current-state process data

The process time is:

140 minutes = 2.33 hours

The total lead time, including rework and elapsed queues, is approximately:

35 business days = 16,800 minutes

The process cycle efficiency is therefore:

PCE = Process time ÷ Lead time
140 ÷ 16,800 = 0.83%

Less than 1% of the elapsed time is hands-on work. The remaining time is principally waiting, queueing, information retrieval and revalidation.

The current staffing model includes:

• 2 underwriting assistants
• 12 cross-trained underwriters
• 2 senior approvers
• 2 policy administration specialists

Although the team has sufficient total headcount to process the work, capacity is poorly aligned with demand. The approval team is only required for complex submissions, while intake and data collection create large queues because work arrives in batches.

A monthly snapshot shows 1,200 open applications, of which 216 are stale, producing an 18% stale rate. Stale submissions are particularly costly when the original quote depends on market rates, property information or external reports that must be refreshed.

The Process Cycle Efficiency Calculator can help teams calculate this process-time-versus-lead-time baseline.

4. Identify the Eight Wastes

The eight DOWNTIME wastes appear clearly in this underwriting stream:

1. Defects: Missing schedules, incorrect risk data and policy-document errors create rework.
2. Overproduction: Applications are pulled into underwriting before required information is complete.
3. Waiting: Applications wait for data providers, underwriters and approval windows.
4. Non-utilised talent: Skilled underwriters spend time chasing documents and correcting duplicate entries.
5. Transportation: Information moves between broker portals, email, spreadsheets and policy systems.
6. Inventory: Open applications accumulate as WIP, including stale submissions.
7. Motion: Staff search across systems for attachments, notes and approval history.
8. Extra-processing: Multiple people recheck the same information without adding risk insight.

The principal bottleneck is not necessarily the longest touch-time step. In this example, external data collection is the constraint because it combines high waiting time, manual coordination and frequent rework.

5. Design the Future State

The future-state map should not simply accelerate every existing step. It should redesign the flow around demand, risk complexity and controlled decision rights.

Countermeasure 1: Create two underwriting lanes

Separate submissions into:

• Fast lane: Complete, low-complexity applications eligible for rules-based processing
• Expert lane: Complex, high-value or unusual risks requiring specialist judgement

This prevents straightforward applications from waiting behind cases that require extensive assessment.

Countermeasure 2: Improve completeness at the source

Use portal validation to check mandatory fields, document types and declared values before submission. Provide brokers with a standard checklist and reject incomplete submissions electronically within four hours, rather than allowing them to enter a long underwriting queue.

Countermeasure 3: Automate data retrieval

Where governance permits, connect the underwriting platform to approved property, business and risk-data sources. The objective is not to automate professional judgement; it is to remove repetitive searching and copying.

Countermeasure 4: Establish pull-based WIP limits

Set a WIP limit of:

• 40 applications in intake and triage
• 80 applications awaiting data
• 60 applications in risk assessment
• 20 applications awaiting approval

When a queue reaches its limit, new work is not pushed forward automatically. The team resolves the constraint or redirects capacity.

Countermeasure 5: Reduce approval layers

Define approval thresholds based on exposure, risk class and deviation from standard rules. For example:

• Standard applications: underwriter approval
• Moderate exceptions: one senior approval
• Material exceptions: specialist committee review

Formal approval checkpoints support governance, but unnecessary checkpoints create bottlenecks. The control should be proportionate to risk.

Countermeasure 6: Use visual management

A daily Andon-style digital signal can alert the team when:

• An application has had no movement for 24 hours
• Required external data is approaching expiry
• A queue exceeds its WIP limit
• An application reaches 10 business days without a decision

This makes delay visible in real time and supports rapid escalation.

6. Future-State Worked Numbers

After piloting the redesign, the target future state is:

• Average lead time: 8 business days
• Process time: 100 minutes
• Stale rate: 3%
• First Pass Yield: 94%
• Straight-through or fast-lane processing: 45%
• Average WIP: 960 applications
• Staffing: 2 assistants, 12 underwriters, 2 approvers and 2 policy administrators, with cross-trained flex capacity during demand peaks

The improved process cycle efficiency becomes:

100 minutes ÷ 3,840 minutes = 2.60%

That may appear modest, but it is more than three times the original PCE. In transactional service processes, the largest opportunity often remains queue reduction rather than shaving seconds from touch time.

7. Current-State Versus Future-State Data

8. Kaizen Sequencing Plan

Improvement should be sequenced so that each action stabilises the next. Attempting a major system implementation before understanding the process can simply automate existing waste.

Phase 1: Baseline and stabilise

Weeks 1–2

• Confirm operational definitions
• Validate timestamps
• Measure lead time, PCE, WIP and stale rate
• Create a daily visual queue board
• Stop avoidable batching where governance allows

Phase 2: Remove preventable demand

Weeks 3–4

• Introduce portal completeness checks
• Standardise broker submission requirements
• Create a first-time-right checklist
• Track missing-information defects by source

Phase 3: Pilot flow lanes

Weeks 5–8

• Launch fast and expert lanes
• Set WIP limits
• Define approval thresholds
• Pilot automated data retrieval for one product segment

Phase 4: Control the improved process

Weeks 9–10

• Create standard work for triage, data retrieval and approval
• Monitor lead time by lane
• Use control charts or run charts for stale rate and First Pass Yield
• Establish a response plan for queue breaches

Phase 5: Sustain and expand

Weeks 11–12

• Review results with the process owner
• Refresh the Process FMEA
• Train replacement and flex staff
• Extend the future state to another product family only after the pilot remains stable

Build the Capability to Lead the Change

Value Stream Mapping is powerful because it connects customer requirements, process data, staffing, governance and financial outcomes in one view. In insurance underwriting, it can show that the central issue is not individual effort; it is the design of the flow.

A trained practitioner can turn this map into a complete DMAIC project: define the business case, measure the baseline, analyse bottlenecks and variation, pilot improvements, and control the result.

Pursue Lean Six Sigma certification with Lean 6 Sigma Hub to learn how to lead data-driven improvement projects using practical simulations, case studies and CSSC-accredited self-paced training. Explore the Lean Six Sigma Green Belt Online Training or review the full Lean Six Sigma Certification pathway.

Kaizen. Kai-Care. Kai-Done. ( Lean Six Sigma)