Where cost disappears in plastics manufacturing
Cost is usually lost not in the material price but in cycles that are not measured per machine and mould. When the same part is produced on a different machine at a different cycle time, unit cost changes; unless that difference is allocated per work centre, product profitability does not reflect reality.
Critical processes
- Mould management: mould inventory, cavity count, maintenance interval, shot counter and a mould-to-machine compatibility matrix.
- Cycle time and OEE: theoretical versus actual cycle per machine, downtime reasons and setup times.
- Material blends: virgin/regrind ratio, masterbatch and additive tracking, with the blend ratio written to the batch record.
- Work centre costing: allocating energy, depreciation, labour and overhead to machine hours.
- Quality: dimensional control, shrinkage allowance, scrap and regrind return.
- Mould change planning: grouping orders that share a mould to reduce setup time.
Which solution maps to what
Bills of material, routings, work centre costing and production planning fall under [ERP consulting and development](/en/solutions/erp-consulting). Machine and line data capture from existing PLC/SCADA systems or counters is done through [API and system integration](/en/solutions/api-integration). Order channels for dealers and OEM customers use the [B2B dealer portal](/en/products/b2b-dealer-portal), and shift and overtime calculation uses [time and attendance](/en/products/time-attendance).
Mandatory integrations
Injection and extrusion machine counters, PLC or SCADA data capture, weighbridges, barcode and label printers, e-invoicing and e-dispatch, bank statements, accounting and quality measurement devices.
Where we start
The starting picture at plastics manufacturers is usually this: a production form kept at the machine, a separate mould logbook, and total material consumption on the accounting side. Because cost is not calculated per machine, product profitability is estimated from overall turnover.
The first exercise produces a work centre definition: which machine belongs to which work centre, which cost elements make up its hourly rate, and whether each element is fed by an energy meter or an allocation key. Without that definition, collected machine data produces charts rather than costs.
What the data model must make explicit
- Mould record: cavity count, shot counter, maintenance threshold, compatible machine list and ownership (ours or the customer's).
- Cycle definition: theoretical cycle time bound to the mould-machine pair; a single cycle time per product is misleading.
- Blend ratio: virgin, regrind, masterbatch and additive ratios written to the batch record.
- Downtime reason: defined as a closed list; free-text downtime cannot be analysed.
- Setup separation: setup time and run time held in separate fields.
- Scrap-to-regrind link: which regrind lot came from which scrap batch.
Roll-out sequence
1. Master data: bills of material, mould inventory, work centres and hourly cost elements.
2. Production recording: work orders, shot counts, downtime reasons and material consumption — manual entry first, automatic capture second.
3. Data capture: connecting counters, PLC or OPC-UA and matching them to work orders.
4. Costing: allocating work centre hourly cost and reporting planned-versus-actual variance.
The order of phases two and three matters: the manual period produces the reference against which automatically captured data is validated. Going straight to automation leaves counter faults undetected for a long time.
What we measure
- Variance between theoretical and actual cycle time per mould-machine pair
- Downtime distribution by reason, and setup time as a share of the total
- Mould maintenance overdue against the shot counter
- Regrind ratio and scrap ratio per batch
- Contribution of work centre hourly cost to product unit cost
Common mistakes
- Binding cycle time to the product. The same part runs at different speeds on different machines; cost must follow the mould-machine pair.
- Leaving downtime reasons as free text. A reason field without a closed list cannot be reported.
- Treating regrind as off-book stock. Regrind with zeroed cost makes the scrap ratio look better than it is.
- Not matching counter data to work orders. Counter data without an order says the machine ran; it is not a cost.
Frequently asked questions
What infrastructure is needed to collect machine data?
It depends on the interface the machine exposes: a digital output/counter, OPC-UA or a vendor-specific protocol. The common requirement in all cases is matching the counter to a work order; unmatched counter data cannot be used in costing.
Does using regrind break traceability?
Not if the blend ratio is written to the batch record. If the scrap batch the regrind came from is not recorded, the backward traceability chain breaks.
Should mould maintenance be in the system?
When planned maintenance is not triggered from the shot counter, maintenance slips and scrap rates rise. Updating the shot counter automatically from production records is markedly more reliable than tracking it by hand.
How is energy cost allocated per machine?
Two routes exist: fitting a meter per machine, or using an allocation key based on connected load and running time. Meters are more accurate; allocation keys are faster to set up. Either way, the method must be stated explicitly in the report, otherwise a cost difference between two periods cannot be interpreted.
If the customer owns the mould, how should it appear in stock?
The mould belongs in the inventory even when the customer owns it, but stays outside depreciation and valuation. Without an ownership field, the return or disposal of long-unused moulds cannot be tracked.
What data is essential for an OEE calculation?
Three components: planned running time, actual output quantity and quality (accepted quantity). OEE can be calculated without downtime records, but the reason behind a low figure stays invisible, which is why downtime reasons are mandatory in practice.