Austin Zhai | Plan · Kit · Deliver 中文 · ← All cases
Case · One-piece flow

One line took WIP
from 2,874 pieces to 173

A textile instrument components plant needed daily output to rise from 350 to 400 while overtime came down. But the line was broken: material waiting, equipment stopping, changeovers slow, semi-finished goods piled between operations. Seven steps later, work in progress was down 94% and floor space down 59%.

Industry Textile instrument components Model High-mix, low-volume Target Daily output 350 → 400, less overtime Method Seven-step shop-floor improvement + one-piece flow
352 → 400Daily output (+13.6%)
+39.9%Productivity
2,874 → 173Work in progress (−94%)
−59.3%Floor space used

High WIP is not a sign that operators are not trying. It is a sign that the line is broken. When one operation finishes, the material is moved away, waits, is found again, and only then goes to the next operation. One-piece flow removes that waiting, moving and searching so material flows piece by piece.

The problem

  • Output was stuck. The target was 350 → 400 a day, and overtime could no longer make up the difference.
  • WIP piled up between operations. 2,874 pieces on the line, much of it "finished, waiting for the next operation", occupying space and cash and hiding problems.
  • Time went into waiting and changing over. 28.5% of operator time spent waiting, 10.67 hours of downtime, 5 hours of changeover — the time actually spent working was fragmented.
  • No space. Semi-finished goods took more than half the floor; there was nowhere to add a line or equipment.
  • Nobody owned the bottleneck. The line was held back by one or two slow stations; other stations running faster made no difference, yet everyone worked to their own rhythm.

The seven steps

  • 1 · Training. Make sure the team understands what is being done and why; otherwise the first action reverts.
  • 2 · Define the task. Which line, to what level, and who owns it.
  • 3 · Analysis. Work-element analysis and line balancing: break every station's motions down to seconds and work out who is busy, who is idle and where the blockage is.
  • 4 · Solution. Re-sequence stations, balance the load, design workstations, set PFEP (where each part is stored, in what quantity, and how it is replenished).
  • 5 · Improvement. Change the layout, build fixtures, change material handling — put the solution on the floor.
  • 6 · Check. Watch results on an hourly output board and find the cause on the spot when output drops.
  • 7 · Sustain. Turn the new way of working into standard work and hold it.

The tools we used

  • Before/after layout: from "workshops by process" to "lines by product family".
  • Work-element analysis and line balancing: find the bottleneck station and level the load.
  • Workstation design: redesign height, bin position and tool placement around ergonomics and takt.
  • PFEP: usage, packaging, storage point and replenishment frequency for every part, fixed on one sheet.
  • Hourly output board: the floor can see how much was made this hour and how far short it is.
  • Bottleneck fixture: a dedicated fixture for the slowest operation to cut its time.

Results

Work in progress
2,874 → 173 pcs
Floor space
−59.3%
Waiting share
28.5% → 7.0%
Downtime
10.67 → 0.67 h
Changeover
5 → 0.55 h

Bars are illustrative, drawn to the before/after ratio; figures come from the project's before/after records.

+13.6%Daily output (352 → 400)
+39.9%Productivity
−94%Work in progress (2,874 → 173 pcs)
−59.3%Floor space used

Why WIP could fall so far

WIP was not managed down; it was flowed down. Stations connected, material at the station, takt aligned — when one operation finishes the next picks up immediately, and nothing needs to pile up. Those 2,700-odd pieces were paying for the break and the wait.

This is also how to judge whether a line can run one-piece flow: look at where WIP piles. Between two operations means the line is broken; in front of one station means that station is the bottleneck.

Who this is for

  • High-mix, low-volume assembly plants with lines split by process and high WIP.
  • Plants that want more output but have reached the limit of overtime.
  • Plants short of space for a new line, where semi-finished goods are quietly taking up the floor.
  • Bottlenecks covered by effort and overtime, with no dedicated fixtures or workstation design.
  • Not for: very large single pieces (heavy equipment), or products so unstable that the routing changes every day.

Answer one question first: where does your WIP pile up?

A two-day on-site diagnostic: walk the line, see where semi-finished goods pile up, work out where the bottleneck is and who is using the space, and leave a 90-day one-piece-flow roadmap. No system first, no large programme first.

© Austin Zhai · Lean digital consultant Case 06 / One-piece flow 沪ICP备17034718号-7