Waste in manufacturing isn’t one problem. It’s a dozen small ones distributed across procurement, production, and distribution, each individually tolerable and collectively expensive. What analysis contributes is mostly identification — finding where waste occurs and what causes it, which is harder than it sounds when it’s spread thin.

Material Waste at Source

Over-ordering is waste before anything is made. Material bought against an inflated forecast, obsoleted by a design change, or expired before use is a complete loss of everything embodied in it. Better forecasting and stock policy prevent this more effectively than any downstream recycling.

Cutting and nesting optimisation reduces offcuts on sheet and bar stock, sometimes substantially. This is a computational problem where software genuinely outperforms manual layout.

Process Waste

Scrap and rework are the visible category. Predictive quality analysis — identifying the process conditions that precede defects — lets you intervene before material is spoiled rather than after.

Yield variation is less visible and often larger in aggregate. Two shifts running the same product with different yields is a signal worth investigating, and the gap between best and average performance is usually recoverable without any capital investment.

Changeover waste — material consumed during setup and startup before output meets specification — is rarely measured separately and adds up quickly on short runs. Sequencing to reduce changeover frequency addresses it directly.

Energy

Energy is waste that leaves no residue, which is why it goes unexamined. Equipment idling between jobs, compressed air leaks, heating and cooling running against each other, and processes operating at higher settings than the product requires all consume continuously.

Pattern analysis against production schedules identifies consumption that doesn’t correspond to output. The finding is frequently that a meaningful share of energy is drawn when nothing is being produced.

Distribution Waste

Part-empty vehicles, inefficient routing, packaging heavier or bulkier than the product needs, and damage in transit are all recoverable through the same optimisation that reduces logistics cost. This is the clearest area where cost and environmental objectives point the same direction.

Making it Stick

Two conditions determine whether waste reduction sustains.

Measurement at a useful granularity. Total scrap as a percentage tells you nothing actionable. Scrap by product, stage, cause, and shift tells you where to go. Most improvement stalls because the data isn’t specific enough to point anywhere.

Feedback to the people who influence it. Teams shown their own waste data, regularly and without blame attached, reduce it. Teams whose data disappears into a monthly report don’t.

Where to Start

Quantify current waste in money by category for one quarter — material scrapped, obsolete written off, energy, rework labour, damaged goods. The ranking is usually different from what people expect, and it points at where the return is rather than where the attention has been.

Do the exercise with finance rather than alone. Waste sits across several accounts and some of it is embedded in variances rather than named as waste, so pulling a defensible total generally requires someone who knows how the accounts are structured.

ticktick.ai tracks waste by category, stage, and cause against production and material data, so the largest sources surface from routine transactions.