Scrap down 52% for a plastics manufacturer on ERPNext

The Machines Were Fine. The Data Was the Problem.

Production throughput was reasonable. The rejection rate was not. Between 8 and 12 percent of output was being scrapped every month — a number the team had come to accept as normal.

The root cause was not mechanical. Machine downtime logs, mould maintenance records, and material batch tracking were all maintained separately — in different files, by different supervisors, updated at the end of each shift from memory.

When a rejection spike hit a particular SKU, no one could reliably trace it to a specific mould condition, material lot, or shift parameter. Every investigation started from scratch. Corrective action was based on experience, not data.

Mould maintenance was scheduled on a calendar, not on actual shot counts — so moulds were sometimes serviced too early and sometimes ran well past their optimal window. The sales team was quoting lead times based on assumed capacity, not actual machine availability.

• Rejection data recorded manually at shift end with no link to machine, mould, or material batch — making root cause analysis nearly impossible
• Mould maintenance scheduled by calendar date, not actual shot count — causing both premature servicing and overrun
• No real-time visibility into machine availability; sales was quoting lead times based on assumed capacity
• Material consumption recorded in aggregate, not by production order — making per-SKU cost tracking inaccurate
• Rework and scrap quantities entered manually in a summary sheet with no traceability back to the specific run

One Connected System Across Every Function

• Manufacturing: Production orders linked to specific machines and moulds with real-time capacity tracking across both shifts
• Maintenance Management: Mould maintenance triggered by shot count thresholds — not calendar — with full service history per mould
• Inventory Management: Material consumption recorded per production order and batch, enabling accurate per-SKU cost tracking
• Quality Inspection: Rejection and rework data linked to machine, mould, material lot, and shift — enabling real root cause analysis
• Sales Order Management: Sales team quotes lead times against live machine capacity, not assumptions

Live in 15 weeks. Both shifts trained. Zero unplanned downtime.

Training and cutover were sequenced so neither shift was taken off the floor at a critical production moment.

Process and Data Audit (Weeks 1–3)

We spent time on the floor across both shifts mapping how rejection data was actually captured, how mould maintenance decisions were made, and where machine capacity information was being lost. All twelve machines and their associated moulds were catalogued.

ERPNext Configuration and Data Migration (Weeks 4–9)

Item master, BOM, mould records with historical shot counts, and machine workstation profiles were migrated and validated. Quality inspection templates, maintenance triggers, and capacity routing were configured to match the actual production floor — including shift-wise scheduling logic.

Parallel Run with Live Orders (Weeks 10–13)

Real production orders ran through ERPNext in parallel with existing logs. Rejection entries and maintenance alerts were cross-checked against manual records until data confidence was confirmed for all twelve machines.

Full Cutover and Shift-Wise Training (Weeks 14–15)

All operations moved to ERPNext as the single system of record. Training was delivered separately to day and night shift supervisors using their own machine and mould data. A 30-day hyper-care period followed with dedicated on-call support.

We used to argue about why a particular mould was rejecting. Now we pull up the shot count, the last service date, and the material lot in one screen. The argument is over in thirty seconds. That is what changed everything. Plant Manager, Plastics Manufacturer