A water spider, or mizusumashi in Japanese, is a dedicated material handler who runs a fixed, timed route delivering parts, removing finished work, and handling small tasks so operators never leave their stations to fetch anything. The role keeps builders building and turns material replenishment into a predictable rhythm instead of a scramble.

The name comes from the whirligig beetle that skims across a pond in quick, precise loops. On a plant floor the water spider does the same: constant motion on a set path, touching every station on a cycle. This guide covers what the role is, why it lifts output, and how to design a route that actually holds. It is a core piece of any lean material system and the mechanism that makes a work cell run without operators wandering off to hunt for parts.

What is a water spider (mizusumashi)?

A water spider is a trained material handler assigned to a repeating route that services a group of stations or a cell on a fixed time interval. Rather than reacting to shortages, the water spider works to a schedule: every so many minutes the same loop is run, dropping off what each station will need next and picking up completed work and empty containers. The route, the timing, and the tasks are all standardized, which is what separates the role from a runner who simply responds to whoever shouts loudest.

The idea comes straight from the Toyota Production System, where the mizusumashi was the moving link that let operators stay in place and keep a steady pace. Toyota treated the route itself as standard work: a defined circuit, a defined cycle time, and a defined set of actions at each stop. That discipline is the whole point. An undisciplined "gofer" adds motion; a disciplined water spider removes it.

Why does a water spider make operators more productive?

A water spider makes operators more productive by removing every reason for them to leave the station. When an operator walks to a rack for parts, waits for a forklift, or hunts for an empty tote, that time is pure waiting and motion waste and the station stops producing while they are gone. Centralizing all of that fetching into one person on a route means the builders build continuously.

The economics are simple. If five operators each lose eight minutes an hour to fetching and staging, that is 40 minutes of lost build time every hour, spread invisibly across the cell. Consolidate it into one water spider whose entire job is that fetching, and you convert five partly-productive builders into five fully-productive builders plus one handler. The math usually favors the trade the moment a cell has more than a few stations, because the fetching does not disappear when you ignore it, it just gets done badly by the wrong people.

Before and after: operators fetching versus a water spider routeBEFORE - operators fetchAFTER - water spider runs a loopparts rackfour stations, four interruptions, tangled paths one loop, timed, operators never leave the station
Before, each operator interrupts their own build to fetch. After, one water spider on a fixed loop feeds every station, and the builders stay put and productive.

What does a water spider actually do on the route?

A water spider does a defined set of tasks at each stop, always the same, so the route stays predictable. The exact list depends on the cell, but the core duties recur across every plant:

Notice what is not on the list: building product. The moment a water spider gets pulled into assembly to "help out," the route slips, stations start starving, and the role collapses back into the chaos it was meant to replace. The route is the job.

How do you design a water spider route?

You design a water spider route by fixing the cycle time first, then loading the loop with exactly the work that fits inside it. The route is standard work, so it gets built with the same rigor as any operator's job.

  1. Define the loop and its stops. Walk the cell and list every station the route must serve and every drop-off and pickup point. Draw the physical path; keep it a loop, not an out-and-back, so the water spider is always moving forward.
  2. Set the route cycle time. Pick an interval, often pegged to a multiple of takt time at which the full loop repeats, for example every 20 minutes. This becomes the drumbeat the whole cell is fed on.
  3. Size the delivery quantities. Each station gets enough material to last exactly one route cycle, plus a small safety amount. Deliver less and it starves; deliver more and you are back to storing inventory at the station.
  4. Time every task on the loop. Measure each pickup, each delivery, and the walking or driving between stops. The sum has to fit inside the route cycle time with margin, or the route will run late and stations will wait.
  5. Standardize presentation at each stop. Fix where parts land, which way containers face, and where empties go. Standard presentation is what lets the water spider hit each stop in seconds instead of improvising.
  6. Write it up as standard work. Document the path, the timing, and the task at each stop as posted standard work so any trained handler can run the route identically across shifts.
  7. Run it, time it, and rebalance. Watch the first week. If the loop runs long, offload a task or add capacity; if it runs short, the cycle time or the route scope can tighten. Treat the route as a living balance, not a one-time design.
A timed water spider route as a fixed loopA 20-minute loop, timed and repeatingSUPERMARKETS14 minS25 minS34 minS44 mindeliver parts,pull finished work,carry the kanban17 min of tasks + walking, inside a 20-min cycle - 3 min of margin
The route is a fixed loop with a time budget at each stop. Keeping total task time under the cycle time, with margin, is what stops the route from slipping and stations from starving.

By the numbers. The mizusumashi is a defined element of the Toyota Production System, catalogued in the Lean Enterprise Institute's lean lexicon under material handling as a fixed-route, timed conveyance system that supports continuous flow (Lean Enterprise Institute, Material Handling). The labor case for the role is straightforward: with U.S. manufacturing carrying hundreds of thousands of open production jobs and rising labor costs (BLS, Manufacturing: NAICS 31-33), keeping skilled builders building instead of walking to racks is one of the cheapest capacity gains a plant can make.

Route time budget: tasks must fit inside the cycleThe route only holds if tasks fit the cycleHEALTHY - 17 min of work in a 20-min cycledeliverpickuptravelmargin20 minOVERLOADED - tasks spill past the cycle, route runs latespillover
Every task and every step of travel has to fit inside the route cycle with margin to spare. Overload the loop and it runs late, which starves the stations it was built to feed.

How is a water spider different from a forklift driver?

A water spider runs a timed, standardized loop for a defined cell; a forklift driver typically responds to demand across the whole plant. The difference is rhythm and scope, and it changes everything about how reliable the material supply is.

DimensionWater spider (mizusumashi)Traditional material handler
TriggerFixed time interval (the route cycle)Reactive: a shortage or a call
PathStandardized loop, same every cyclePoint-to-point, varies each trip
ScopeOne cell or line, small quantities, oftenWhole plant, large loads
QuantityOne route-cycle worth per stopFull pallets, whatever was requested
Effect on operatorsThey never leave the stationThey wait for delivery or fetch it themselves
A water spider trades big infrequent deliveries for small frequent ones on a rhythm. That rhythm is what keeps stations fed without piling inventory at the point of use.

This does not make forklifts obsolete; it changes their job. Forklifts and tuggers often replenish the supermarket the water spider draws from, handling the bulk moves while the water spider handles the last leg to the station in small doses. The two work as a system: bulk to the supermarket, sips to the line.

What are the common water spider mistakes?

The role fails in predictable ways. The first is letting the water spider build product; the route slips the moment they get absorbed into assembly. The second is an unstandardized route, where the path and quantities change every cycle and the "route" is really just reactive running with a nicer title. The third is overloading the loop so total task time exceeds the cycle time, guaranteeing the route runs late and stations starve. The fourth is skipping the standard-work documentation, so the route lives only in one person's head and dies the day they call in sick.

Each mistake has the same root: treating the water spider as a person rather than a designed system. The route is the deliverable. When it is timed, standardized, and documented, a trained handler can pick it up cold and the cell keeps flowing. Plants that instrument their floor with live station-level visibility can see when a station is about to starve before it does, which turns route design from guesswork into a tuned loop. See how one plant made floor material flow visible in our CLS case study. No rip-and-replace; the water spider just always arrives before the station runs dry.