I — Datum

Value Stream Mapping with Simulation

Draw the map, then run it. A value stream map in SVEND is not a drawing — it is a model. The same steps, inventory and cycle times that make the map feed a discrete-event simulator, a scaled floor layout, and the pull-sizing maths, and the improvements you find come out the other side as tracked projects rather than sticky notes.

Ref: Rother & Shook, Learning to See · Little's Law · AIAG

II — The Map

Current state and future state, authored as structured data rather than shapes on a canvas. Because the map is data, every number below it is derived rather than typed, and a change to a step propagates to the timeline, the simulation and the sizing at once.

ElementWhat it carriesLive
Process steps Cycle time, changeover, uptime, shifts, operators, scrap rate, units per cycle, MTBF/MTTR
Inventory Triangles placed between steps, quantity and days-of-supply, carried into lead time
Transport links Movement between steps, with distance derived from the floor when one is drawn
Kaizen bursts Improvement opportunities pinned to the step they belong to
Timeline Two-level Rother/Shook ladder — non-value-added on the top line, process time notched beneath
Derived metrics Lead time, process time, process cycle efficiency, WIP, bottleneck
Future state & compare Build a future state, compare it to current, keep the history; export CSV, Excel or JSON

The map validates itself: silent defaults, demand-rate consistency and Little's Law reconciliation are checked and reported rather than assumed.

III — Simulation

The map runs as a discrete-event simulation. This is the part most mapping tools do not have and most simulation tools price as capital equipment: the queues, failures and changeovers that decide real lead time, executed against the map you already drew, with no second model to maintain.

BehaviorModeled asLive
Queues and blocking Work waits where capacity is short; buffers fill and starve downstream
Breakdowns MTBF and MTTR per station, sampled during the run
Changeover Setup time between runs, charged to the station that incurs it
Batch equipment Genuine batch processing for ovens and autoclaves, distinct from a continuous station
Quality loss Defect rate and its effect on effective throughput
Labour Operators per station as a finite resource
Replications Repeat runs with confidence intervals — a single run of a stochastic model is an anecdote
Outputs Average lead time, average WIP, throughput, and the bottleneck the run actually found

Replications with confidence intervals are the reason the number is worth quoting: the simulator reports a range, not a single lucky run.

IV — The Floor

Place the same steps on a scaled floor, in meters, and the model gains geometry. Travel stops being a guess and becomes a distance.

CapabilityDetailLive
Scaled layout Floor bounds in meters; equipment library of curated machine envelopes, each still editable
Clearance checking Minimum aisle and operator clearance enforced as a rule, reported as violations
Routed lanes Paths routed around equipment footprints rather than straight through them
Replay Watch jobs walk the floor from the simulation's own event log — play, pause, scrub, and a playback rate kept separate from transport speed

Replay interpolates only between timestamps the engine actually reported, and holds position across an instantaneous hop — it never draws a journey the run did not make.

V — Pull Sizing

The arithmetic that turns a future-state idea into a quantity somebody can put on a card, a lane or a shelf.

CalculationAnswersLive
Lot recommendation Economic run quantity for a step, from its setup and holding cost
EPEI Every-part-every-interval — how often the mix can be cycled at a given lot size
SMED impact What a setup-time reduction is actually worth in lot size and interval
Family lot sizing Sizing across a product family sharing a work center
Supermarket sizing Cycle, buffer and safety stock for a pull supermarket
FIFO lane sizing Lane capacity where a supermarket does not fit

VI — From Map to Project

A map that ends in a PDF changes nothing. The point of the future state is the work it implies, so the improvements leave the map as reviewable items and arrive in the vandrebok carrying where they came from.

StepWhat happensLive
Waste register Waste analyzed per step against the map's real demand and takt, listed rather than narrated
Proposals Generated from the future state with an expected saving, then approved or rejected
Promotion An approved proposal mints an entry in the vandrebok, carrying its charter
Provenance The entry remembers which map and which burst produced it; promoting twice returns the same entry
Targets Measures attached to the value stream, with readings over time

Approval is required before promotion, and it is enforced in one place — there is no route that mints a project from an unreviewed idea.

VII — Boundaries

What this does not do yet. We would rather you find this here than after you have paid for it.

Information and scheduling flow is not authorable. The map models material flow. The production-control box, EDI arrows and schedule drops that sit across the top of a classic Rother/Shook map cannot be drawn yet, so a SVEND map shows how material moves, not how the schedule reaches the floor.
Only process steps occupy the floor. Racking, dock doors and staging lanes take up real space and constrain real clearance, and today the floor holds process steps only — so a clearance verdict is optimistic against a plant with storage in it.
Comparison pairs steps by name. Renaming a step between current and future state drops it from the step-by-step comparison, though the headline lead time, WIP and efficiency figures still move. The comparison reports how many steps it paired, so you can see when this bites.

VIII — Where It Sits

Value stream mapping tools generally stop at the drawing. Discrete-event simulators generally start at a separate model, licensed per seat as capital expenditure, and are handed to a specialist. Statistical packages do neither. The gap that costs improvement teams the most is the one between the map and the model — two artefacts of the same process, maintained separately, disagreeing quietly.

SVEND keeps them as one artefact: the map is the model, the simulation runs on it, the floor is the same steps in meters, and the improvements become tracked projects. It is included in Pro at $19 per month, billed monthly, with no implementation engagement attached.