Skip to content

Value Stream Redesign

Reduced idle time by 70% and cycle time by 15.5% on a precision component production line — through value stream analysis, bottleneck resolution, and a lean target state with Kanban, Heijunka, and continuous flow.

Standalone project.

Lean Six Sigma VSM

The Problem

The production line for precision components had no production planning system and no demand forecasting beyond blanket orders. The primary processing step — the bottleneck — exceeded available working hours, forcing overtime and temporary labor while overloading the machines. An analog ordering process for production materials used hand-filled paper forms with a 20% rejection rate (illegible, incorrectly completed), causing late deliveries and idle time on the shop floor. Intermediate stock sat in chaotic mixed storage areas between process steps, with no FIFO enforcement and no visibility into machine availability. Pull control by individual workers led to high WIP, inconsistent cycle times, and irregular deliveries to the end customer. Total lead time was 1,507 minutes with 10 days of idle time across the value stream.

The Approach

Current-state analysis

Value stream mapping
To establish a shared factual baseline before proposing changes, the entire production flow was mapped ramp-to-ramp — from goods receipt of raw materials to outbound delivery note — using stopwatch time studies, ERP work-order data (full year), and on-site Gemba walks with production staff and management. Idle times were calculated as the average delta between completion and start dates across all work orders for the period.
Product family selection
The highest-revenue product group was identified as the A/X category — roughly 50% of revenue with stable demand from blanket orders. The highest-volume variant was selected for the value stream, representing a product family where all variants share the same manufacturing process and material flow despite mechanical diversity.
Bottleneck identification
The primary processing step was identified as the constraint: its cycle time exceeded available working time, requiring overtime or temporary staff and preventing maintenance windows. To focus effort where it would actually move throughput, a cost–benefit matrix of all identified issues guided prioritization — high-cost, low-benefit items like implementing a full production planning system were excluded from the target state in favor of measures that could be implemented and stabilized incrementally.

Target-state design

Bottleneck relief
To create a capacity buffer without adding labor, a third machine was specified to increase throughput at the constraint — eliminating overtime and machine overload within available working hours. A specialized welding machine was procured to cut welding cycle time by 50%.
Continuous flow with FIFO lanes
Chaotic intermediate storage areas were replaced with FIFO lanes between process steps. Where takt alignment was possible, continuous flow was established gated by signal Kanban at the bottleneck rate. The welding and straightening steps were parallelized.
Supermarket pull system
Because components cure overnight — a natural decoupling point in the flow — a supermarket pull system was introduced at the curing stage. Production Kanban cards control work order release into the second segment, allowing priority changes on short notice without disrupting flow.
Production levelling (Heijunka)
To smooth the volume fluctuations caused by mixing make-to-stock and make-to-order, the production mix was levelled across all product families. Customer deliveries were restructured to weekly fixed quantities, aligning production takt to customer takt.
Material supply via Kanban
To eliminate the 20% form-rejection rate and the delays it caused, the analog paper-based ordering process was replaced with a combination of ERP-based MIN/MAX replenishment and Kanban. Specialized materials are commissioned per work order, while sub-materials and packaging are commissioned once per week and staged directly in the production warehouse without an intermediate stop in the main warehouse. Lot sizes were revised to match optimal production batch sizes and reduce lead times.

Implementation planning

Phased rollout
To avoid disrupting a running production line with simultaneous changes, an 11-step action plan followed a continuous improvement (CIP) approach — each measure implemented and stabilized before the next begins. Responsibilities were assigned across sales, purchasing, production, logistics, and engineering, with a kick-off to detailed planning cycle of roughly three months.

Architecture

SUPPLIER 01raw · monthlySUPPLIER 02parts · irregularSUPPLIER 03consumables · bi-monthlyPLANNING · PURCHASING · SALESmanual · spreadsheets · verbal handoffsCUSTOMERtakt · variablepaper ordersorders · phone · email · faxdaily paper work orderMAIN WAREHOUSEmixed · no min/maxPROD. STOREmixed SKUsSTEP 1·TT30C/O3.5CT1.5OPS1UPTIME?no OEE dataSTEP 2·TT17C/O1CT0.9OPS1STEP 3·TT527C/O0CT26.4OPS1MACH2 idleparallel unusedSTEP 4·TT727C/O<15CT36.4OPS1STEP 5·TT22C/O0CT1.1OPS212% reworkSTEP 6·TT31C/O4CT1.6OPS1failures → step 4STEP 7·TT29C/O0CT1.4OPS1STEP 8·TT45AV/OR<15CT2.3OPS1ops fetch materials142231269118rework loops1d1d3d3d1d1d1d2d30 m17 m527 m727 m22 m31 m29 m45 mLEAD~13 dVA · CT1428 m12345678
SUPPLIERconsolidated · weeklyPLANNING · PURCHASING · SALESERP · weekly heijunka · one signalCUSTOMERtakt 110 / wkEDI · weeklyEDI orders · dailySUPERMARKET2-bin · kanbanpacemaker scheduleCELL A · PREPTT65 mC/O<1CT3.3 mOPS2OEE85%merges steps 1–2CELL B · CORETT395 mC/O<5CT19.7 mOPS2MACH2 ∥parallel · merges steps 3–5CELL C · FINISHTT125 mC/O<3CT6.3 mOPS2OEE88%single test · merges steps 6–11PACK ◆PACEMAKERTT45 mC/O<2CT2.3 mOPS1packaging deliveredFG MARKET2-day coverSHIPPINGfixed daily windowFIFOFIFOFIFOFIFOwithdrawal kanban1 d1 d1 d1 d2 d65 m395 m125 m45 mLEAD~6 dVA · CT1290 m

The Result

Idle time reduction

70% — from 10 days to 3 days across the value stream

Lead time reduction

15.5% — from 1,507 to 1,274 minutes per lot

Cycle time reduction

15.5% — from 75.4 to 63.7 minutes per piece

Bottleneck capacity

Third machine at the constraint creates a capacity buffer, eliminating overtime and machine overload

Process consolidation

9 discrete steps consolidated into 6 workstations through parallelization and merging