First pass yield, rework, start-up losses, giveaway and fill weight — the complete guide to the Quality component of OEE for food and FMCG manufacturing.
Quality is the third and final component of OEE. It measures the proportion of total units produced that meet specification first time, without any rework, reprocessing or rejection. If Availability tells you how much of your scheduled time the line was running, and Performance tells you how fast it was running during that time, Quality tells you how much of what it produced was actually good.
Quality is the only OEE component where the loss is visible as a physical output — rejected product, reworked trays, hold stock awaiting QA release. This makes it the most tangible of the three losses, and often the easiest to trace back to a root cause. It is also, in most food manufacturing environments, the component with the most financial complexity — because quality losses appear not just as scrapped product, but as giveaway, rework labour, hold costs, and potential customer complaints.
Quality losses come from two specific loss types in the Six Big Losses framework: defects and rejects (units failing to meet specification during steady-state production) and start-up and yield losses (units produced during line start-up or changeover that fail to meet specification while the process stabilises). Both reduce Good Units relative to Total Units Produced.
In this example, 186 reworked pouches that were eventually sold are still counted as Quality losses. The start-up rejects (320 pouches) represent the largest single loss category — which is typical for filling lines where product temperature, cap torque and fill weight all need time to stabilise after a start or changeover.
Like Performance, Quality is damaged by exactly two loss types. Understanding the difference matters because they occur at different points in the production cycle and require different preventive approaches.
Units that fail to meet specification during normal production — after the line has stabilised and is running at steady state. These are the losses captured by in-line quality checks, end-of-line inspection and checkweighers.
Units produced at the beginning of a run or immediately after a changeover that fail to meet specification while the process stabilises. This period of instability is predictable but in many sites is not systematically minimised.
The most common steady-state quality loss in food manufacturing. Fill weight drift occurs when filler nozzles wear, product viscosity changes with temperature, or pump calibration drifts over a shift. A checkweigher running in reject mode rather than feedback mode catches the failure — it does not prevent it.
Caused by contaminated sealing jaws, product in the seal area, sealing temperature drift, or film tension issues. Seal failures are high-risk in food manufacturing because a failed seal is a food safety risk, not just a quality defect. Any seal failure should trigger a hold and investigation of the preceding product window.
Metal detected by in-line metal detectors, or non-metal foreign bodies identified by X-ray. Any foreign body reject triggers an investigation window — typically the preceding 30 minutes of production is placed on hold pending investigation. The OEE Quality loss includes both the confirmed reject and the hold product that is subsequently condemned.
Accurate Quality measurement requires capturing every non-conforming unit at every point of inspection — not just end-of-line rejects. In many sites, in-process rejects at checkweighers or metal detectors are captured separately from end-of-line rejects, and rework volumes are tracked on a paper-based system that is never consolidated into the OEE calculation. The result is a Quality figure that only reflects part of the true loss.
Rank all defect types by total units lost over 4 weeks. In most food sites the top 2 failure modes account for 70%+ of all quality losses. Knowing which defect type is dominant tells you where to focus engineering resource.
A checkweigher that rejects underweight packs is detection — it stops the defect reaching the customer but does not reduce the loss. A filler with closed-loop feedback control that adjusts fill volume automatically is prevention — it stops the defect being produced. Wherever possible, shift quality control from detection to prevention.
Statistical Process Control (SPC) on fill weight, seal temperature and other critical parameters allows operators to see drift before it becomes a defect. If fill weight is trending toward the lower control limit, the filler is adjusted before product goes out of tolerance — not after it has been rejected.
Foreign body contamination should never be treated as a random event. Every detection should trigger a structured investigation: source identification, entry point, why the detection system caught it (or didn't), and what physical change prevents recurrence. A site that has repeat foreign body events from the same source has a systemic failure, not bad luck.
Start-up losses are the units produced during the period between line start or changeover completion and the point at which all critical parameters are stable and within specification. They are classified as Quality losses because the units produced during this window do not meet specification first time.
What distinguishes start-up losses from steady-state defects is that they are predictable. Every line start and every changeover will produce some start-up loss. The question is not whether it happens, but how many units are lost before stability is reached — and whether that number is being actively minimised.
Sealing jaws, filling heads, pasteurisers and heat tunnels all need time to reach operating temperature after a cold start or changeover. Product running through the system before temperature is stable produces sealing failures, fill weight errors and product temperature deviations. Pre-heating protocols — starting the heating elements before the line is ready to run — directly reduce this loss window.
Fillers require product flowing through the system to stabilise at the target fill weight. The first 20–50 fills after a start are typically variable until product flow is consistent and nozzle temperature is stable. Automated tare weight zeroing and pre-run calibration checks before production release reduce the number of units produced before stability is confirmed.
During product changeovers, residual product from the previous SKU contaminates the first units of the new SKU. The carry-over window depends on the line design, the products involved and the changeover procedure. Allergen changeovers require the longest verification window. Units produced before the changeover verification is complete must be held and assessed — they cannot be assumed to be good.
Lines released into production without completing all pre-start checks produce quality losses that are entirely preventable. A sealing jaw not torqued to specification, a checkweigher not calibrated, a date code not verified — each represents a defect window that only exists because the pre-start procedure was incomplete or bypassed under time pressure.
Count the units from line start to first-off confirmation separately from steady-state production. Track the number of units lost and the time taken to reach stability for every start and every changeover. Without this data, the loss is invisible and unmanageable.
A documented, auditable pre-start checklist — with sign-off before the line releases into production — eliminates the majority of missing pre-start check losses. The checklist should be product-specific, not generic. A salad line checklist and a sauce filling checklist are not interchangeable.
Where line design permits, start heating sealing jaws, pasteurisers and heat tunnels during the changeover rather than after it. If the sealing jaw is at operating temperature by the time the changeover is complete, the thermal stabilisation window is eliminated entirely.
The line should not enter steady-state production until defined first-off criteria are confirmed — fill weight within tolerance, seal integrity passed, date code verified, label position confirmed, checkweigher calibrated and running. These criteria should be product-specific and documented. The release decision should require a sign-off, not be left to operator discretion.
Once start-up loss is measured, set a target. Express it as units lost per changeover or minutes from start to first-off confirmation. Review it weekly. Sites that track and review start-up loss consistently achieve 30–50% reductions within 3 months through procedural improvements alone, before any capital investment.
Giveaway is the difference between the actual average fill weight of a product and its target or nominal fill weight. Because UK food law (the Weights and Measures Act and the Packaged Goods Regulations) requires that products meet their declared weight, production lines are set to overfill by a controlled margin — the Target Mean Level (TML) — to ensure the vast majority of packs meet the declared weight.
The problem is that giveaway is a direct cost. Every gram of product given away above the TML is product that was manufactured and given to the customer for free. On a high-volume food line, a 1g reduction in average giveaway can be worth tens of thousands of pounds per year.
| Parameter | Scenario A — current | Scenario B — improved |
|---|---|---|
| Declared weight | 500g | 500g |
| Fill weight std deviation | 4.0g | 2.5g |
| Target Mean Level (TML) | 508g | 505g |
| Giveaway per pack | 8g | 5g |
| Annual production volume | 10,000,000 packs | 10,000,000 packs |
| Total giveaway per year | 80,000 kg | 50,000 kg |
| Product cost per kg | £2.50 | £2.50 |
| Annual giveaway cost saving | — | £75,000/year |
Reducing fill weight standard deviation from 4g to 2.5g — a 37% improvement in process capability — saves £75,000 per year on this example line. The investment required is typically filler maintenance, nozzle replacement, and SPC implementation — all well within the payback period justified by the saving.
| Sector | Typical Quality range | Dominant loss type | Key improvement lever |
|---|---|---|---|
| Ready meals / chilled food | 96–99% | Start-up losses at sealing; giveaway on portioning | Pre-heat sealing jaws; SPC on fill weight; reduce changeover count |
| Baby food (pouches / jars) | 97–99.5% | Start-up losses on filling and capping; seal integrity | Thermal pre-conditioning; capper torque engineering; first-off criteria |
| Bakery (bread, morning goods) | 95–98% | Weight variation on depositing; start-up on oven temperatures | Dough weight SPC; oven pre-heat protocol; depositor maintenance |
| Beverage (carbonated / still) | 98–99.8% | Fill level variation; cap application failures | Filler head maintenance; cap torque monitoring; carbonation SPC |
| Fresh produce (salad, veg) | 94–98% | Weight variation on multihead weighers; bag seal failures | Weigher calibration frequency; seal jaw cleaning schedule |
| Confectionery / snacks | 97–99.5% | Weight variation; seal and wrap failures; start-up scrap | Weigher maintenance; wrapper tension control; oven temperature SPC |
| Sauces and condiments | 97–99% | Fill weight variation; cap torque; label application | Pump calibration SPC; cap torque monitoring; label register checks |
| Quality score | Classification | What it typically means |
|---|---|---|
| 99.9%+ | World-class | Near-zero defects; SPC in place; start-up losses minimised |
| 99–99.9% | Good | Low defect rate; start-up losses present but controlled |
| 95–99% | Typical food/FMCG | Regular defect losses; start-up scrap not fully minimised |
| 90–94% | Below average | Significant Quality losses — structured investigation required |
| Below 90% | Poor | Major Quality issue — root cause analysis and engineering review needed urgently |
Enter your shift data to get Availability, Performance and Quality with full breakdown.