How to Run a Value Stream Map in Insurance

The short version

A value stream map in insurance traces one thing — a claim, an application, a policy change — through every step, queue, and handoff from request to resolution. The point isn't a tidy diagram; it's a shocking number: how little of that elapsed time the work is actually being worked. This guide shows how to map a claims value stream end to end — what to record at each step — including the quality and rework metrics a factory map leaves out — a worked example with the real math, and the future-state moves that take days out of the cycle. You'll finish able to map one value stream this week and see exactly where your lead time hides.

Why value stream mapping is different in insurance

Almost everything written about VSM maps a factory: steel moving through presses, parts stacking up between machines. An insurance operation breaks that mental model, and copy-pasted factory VSM stalls because of it:

• There's no material — you map information. The "product" is a claim or application moving through people, inboxes, core systems, and approval queues. You can't see it pile up on a floor, which is exactly why the waste stays invisible.
• The waste is waiting, not motion. In a plant, waste looks like work-in-process inventory. In insurance it looks like a claim sitting in an adjuster's queue, a file parked pending more info, an email waiting two days for a reply. The muda is real — it's just quiet.
• Rework is the hidden tax. A claim that arrives incomplete bounces between desks for missing documents or a wrong code. Factory VSM tracks machine uptime; insurance VSM has to track percent complete & accurate — did the work arrive good enough to use without sending it back?
• The headline metric is flow efficiency. Every services value stream you map produces the same gut-punch: the claim spends almost all its life waiting. A file that takes two weeks to settle might get two hours of actual work. That ratio — work time over elapsed time — is the number that changes the conversation.

So the goal isn't to make adjudicators type faster. Borrowing the Toyota lens: efficiency is the consequence, not the goal. You attack the queues and the rework loops, and speed, cost, and a better customer experience follow.

A claims value stream: small work steps, big queues waiting between them.

“

The one-line test for every box on the map: is the claim being worked, or is it waiting? In insurance the honest answer is "waiting" — by a landslide — and that's where your lead time is hiding.

Run those numbers on a stream of your own before reading on:

What a value stream map of a claim actually captures

A factory VSM records cycle time and machine uptime. An insurance map keeps the boxes but changes what goes in the data box under each step. Here's what to capture as you walk the stream.

1Process time — cycle time: the few minutes of real work

For each step, record the hands-on time — how long the task takes once someone actually starts it. Adjudicating a straightforward claim might be 20 minutes; intake, 10. This is almost always tiny, and that's the point: process time is the small number you'll compare everything else against.

2Wait time: where the days actually go

Between every step is a queue — the time the claim sits before anyone picks it up. Record it honestly (a day, three days, a week). On most insurance maps the sum of wait times dwarfs the sum of process times — often by 20-to-1 or worse. Waiting is the waste you're really hunting; the work itself is rarely the problem.

3Percent complete & accurate (%C&A): the rework loop you can't see

At each step, ask the people downstream: what fraction of the work arrives complete and accurate enough to use without sending it back? That's %C&A. A 70% intake means three of every ten claims bounce back for missing info — a rework loop that adds more delay than any single slow desk. Multiply the %C&A across all steps and you get the first-pass yield of the whole stream — often a sobering number.

4Queue size — the stack of claims between desks

Note how many claims are waiting in front of each step. A queue that keeps growing marks your bottleneck — the step that sets the pace of the entire stream, exactly as the slowest machine does on a line. Big queues are where you look first.

5Demand and takt time: the pace the stream has to hold

Count how many claims arrive per day. Divide your available working time by that demand and you have takt time — the pace the stream must sustain to keep up. Takt turns "we're behind" into a number you can design against.

How to run your first claims value stream map (step-by-step)

Pick one value stream — one claim type, end to end — not the whole department. Auto first notice of loss (FNOL) through payment is a good first map.

1. Pick the stream and its endpoints. Name where it starts (claim submitted) and ends (claim paid or closed). One product family, one map.
2. Walk it at the gemba. Follow a real claim through the system with the people who touch it. Don't map from a conference room — map what actually happens, including the workarounds.
3. Draw the current state. Box per step, a data box under each (process time, wait time, %C&A, queue size), and the queues between them. Draw the rework loops right on the map, too — the arrows where a claim bounces back upstream for missing or wrong information. Those loops are usually where the lead time is hiding.
4. Add the timeline and do the math. Lay process time and wait time on a stair-step line at the bottom. Sum each. The ratio of total process time to total lead time is your flow efficiency — the headline.
5. Find the waiting and the rework. Mark the biggest queues and the lowest %C&A steps. These are your targets.
6. Design a future state. Where can you cut a handoff, fix accuracy at the source, pull work instead of pushing it, or level the intake? Draw it, then pick the first kaizen to get there.

A worked example: an auto claims value stream

Here's how a first map tends to look. The numbers are an honest hypothetical — the shape is what you'll actually find.

You walk an auto claim from first notice of loss to payment and record each step:

Add it up: about 2 hours of real work, spread across roughly 11 working days (about 88 working hours) of elapsed time. That's a flow efficiency near 2% — the claim is being worked about one fiftieth of the time it's in your shop. And the rolled first-pass yield (0.70 × 0.90 × 0.85 × 0.95 × 0.99) is around 50% — only half of all claims travel the stream clean; the rest loop back for missing or wrong information.

Read that map and the targets are obvious, and they're not "adjudicate faster." The five-day queue before adjudication is the bottleneck. The 70% intake %C&A is the rework loop feeding it — fix the FNOL form so a claim can't be submitted incomplete and you stop creating the bounce-backs that swell the queue. The days come out of the waiting and the rework, not out of the people. Get a clean claim to flow one at a time instead of moving in batches between desks, and the timeline collapses while the work content barely changes.

Now build your own. Drop in the steps, set each one's process and wait time, and the tool draws the timeline and calculates flow efficiency for you — current state first, then a future state:

A quick value-stream health check

You don't redraw the whole map every month. Walk the stream and answer five questions:

1. Flow efficiency: Do you know your process-time-to-lead-time ratio? Is it trending up?
2. The big queue: Which step has the longest wait in front of it, and is it shrinking?
3. %C&A at the front: Is intake arriving complete and accurate, or still feeding rework downstream?
4. Handoffs: How many times does the claim change hands or systems — and have you removed any?
5. Pace: Is the stream keeping up with takt (demand), or is the backlog growing?

Build and share the map itself — current and future state, with the timeline math done for you — using the free value stream map tool.

Common insurance VSM mistakes

• Mapping the system, not the work. The process diagram in the procedure manual is the official story. The gemba — how a claim actually moves — is the real one. Map reality.
• Chasing process time. Shaving minutes off a 60-minute adjudication is rounding error against an 11-day lead time. Attack the wait.
• Ignoring %C&A. A map without complete-and-accurate data hides the rework loops that cause most of the delay. It's the metric office VSM most often skips and most needs.
• Mapping every product at once. One claim type, end to end, beats a tangled map of all of them. Pick the highest-volume or most painful stream first.
• Drawing a future state with no kaizen behind it. A beautiful future-state map that nobody is assigned to build is wall art. End with one improvement and an owner.

Templates & tools

• Free value stream map tool — build current- and future-state maps, with process/wait time and the flow-efficiency math handled for you. No login required.
• Spaghetti diagram — trace how a single file bounces between people and systems; a strong companion to the VSM.
• Related guides: How to Use Takt Time in Manufacturing for the demand-pace math, and How to Do 5S in a Warehouse for organizing the work the stream surfaces.

FAQ

What is value stream mapping in insurance? It's tracing one work item — typically a claim or application — through every step, queue, and handoff from request to resolution, recording process time, wait time, and percent complete & accurate at each step. The map exposes how little of the elapsed time the work is actually being worked, and where the delays and rework hide.

What's the difference between mapping a factory and mapping an insurance process? A factory map follows material through machines and tracks cycle time and uptime. An insurance map follows information — a claim through people and systems — and tracks wait time and percent complete & accurate. The dominant waste shifts from work-in-process inventory to claims waiting in queues.

What is flow efficiency in a claims process? Flow efficiency is total process (hands-on) time divided by total lead (elapsed) time. In insurance it's often around 2-5% — meaning most of the cycle is queue time, not work. Raising it is mostly about removing queues and rework.

What is %C&A and why does it matter? Percent complete & accurate is the share of work that arrives at a step good enough to use without sending it back. Low %C&A — especially at intake — creates rework loops that add far more delay than any slow desk. It's the quality metric transactional VSM lives on.

Which value stream should I map first? Pick one claim type, end to end — the highest-volume or most painful one. One product family mapped deeply beats every product mapped shallowly.

Related concepts & guides

• Dictionary: Value Stream Mapping · Value Stream · Cycle Time · Takt Time · First-Pass Yield · Bottleneck Analysis · Pull System · Muda
• Guides: How to Use Takt Time in Manufacturing · Value stream mapping in healthcare (coming next)

Sources

• Lean Enterprise Institute, Lean Lexicon: Value-Stream Mapping
• Keyte, B. & Locher, D., The Complete Lean Enterprise: Value Stream Mapping for Administrative and Office Processes