What Italian Container Glass Plants Get Right About Job Changes

It is 02:40 on a Tuesday at a two-furnace plant north of Pesaro. The machine setter is already kneeling in front of section 4, a pre-heated mould set at his side. Sections 1, 2, and 3 are still running. He won't stop the machine.

That is not what most operators outside Italy expect the first time they see it. An Italian container glass manufacturer running a section-by-section changeover on a live IS machine, swapping from a 75 cl Burgundy into a 50 cl Bordeaux, mid-shift, while the downstream lehr stays loaded and cold-end packing continues. The change is invisible to the palletiser.

Italy ranks second in EU container glass production at approximately 3.8 million tonnes per year across 14 active furnace sites, per FEVE's 2024 Annual Report. That volume comes from a high-mix SKU environment of wine, spirits, olive oil, and condiment glass, where changeover frequency isn't a quarterly event. It's a daily operational discipline. What Italian operators have built out of necessity, most plants elsewhere haven't systemised at all.

Section-by-section changeover is a culture, not a scheduling trick

The section-by-section model works like this: replacement mould sets swap one section at a time while adjacent sections continue producing. Total machine dead-time drops from the industry-typical 4–6 hours to 90–150 minutes. Each 30-minute section settling window targets below 2% cavity-level scrap before the next section rolls. In a plant running 12–15 SKU changes per week across a high-mix wine and spirits portfolio, that difference computes directly into furnace-days of recovered production.

But SBS is not something you install. The plant north of Pesaro was running a BDF Industries 10-section line with modern electronic timing controls. The hardware wasn't the constraint. The machine setter, the end-of-arm tooling operator, and the hot-end superintendent all have to be present simultaneously for each section swap, working from the same documented recipe. In 2019 I audited a Spanish wine bottle plant that had the right equipment and had read about SBS. They ran three sections cleanly, got a timing conflict on section 4, and the shift leader called a full stop to sort it out. They lost the advantage entirely because the handover protocol wasn't built to support an SBS cadence. The methodology works. The culture that sustains it is harder to copy.

Look, every plant that has tried to import SBS without rebuilding the shift-handover framework around it has reverted to full-machine-stop within a few months. Generic Lean/Six Sigma boutiques applying standard SMED to container glass changeovers routinely misclassify the irreducible 15–25 minute mould temperature stabilisation window as internal changeover time amenable to reduction. It isn't. Setting unachievable targets on that basis is one of the fastest ways to get your SOPs quietly abandoned after the consultant leaves.

The two non-negotiables before gobs touch the new mould

Every Italian plant running a competent operation pre-heats replacement mould sets to 400–450°C in a dedicated oven for a minimum of 45 minutes before installation. Four hundred and fifty degrees. Non-negotiable. Cold-mould insertion is the primary root cause of check defects (micro-cracks at the parison-to-mould interface) in the first 15–20 minutes after a swap. On a lightweight wine bottle running NNPB, those checks appear at the cold-end polariscope and the AQL hits before you've had time to stabilise gob weight. Italian SOPs treat the preheat hold as a gated checkpoint, not a guideline.

The second non-negotiable is gob weight control. Italian best-practice mandates a ±0.3 g tolerance on target gob weight within the first 10 cuts after a mould swap, verified by inline weight-check every five minutes. The wider industry norm at changeover runs ±0.8–1.2 g, and that variance propagates wall-thickness variation through the annealing lehr. On a 100 g wine bottle, a 0.9 g gob overshoot doesn't register as a quality issue until you see the wavy-finish reject rate three hours later at the polariscope line (and by then it's already been palletised by someone who wasn't told there was a change).

For NNPB production, plunger contact time must be re-set to ±1 ms after every job change. Drift beyond ±3 ms produces wall-thickness coefficient of variation above 8% and elevated wavy-finish rejects. Operators who have only run blow-and-blow don't feel that sensitivity until they're chasing the defect on a 187 ml spirits miniature and can't isolate the source because nobody logged the plunger re-set at the changeover.

The mould preheat and the gob weight tolerance aren't Italian secrets. They're written in the physics of the process. The difference is that Italian operators treat them as gates, not guidelines, and the SOP enforces that distinction on every change, every crew, every shift.

The 40% of post-change scrap that doesn't show up in your data

In 2017 I was auditing a mid-sized plant in southern Italy. Four IS lines, roughly 800 t/day across flint and green campaigns for domestic wine bottlers and northern European export accounts. The plant had acceptable mean changeover times. What it had as a problem was a scrap spike that nobody could cleanly attribute to the change event itself. The data looked like normal process variation.

Not a furnace problem. A handover problem.

The pattern only emerged when I overlaid defect timestamps with the shift-change log. Every spike fell within 90 minutes of a crew handover that had coincided with an open changeover. Italian plants with the lowest post-change defect rates mandate a 30-minute overlapping shift whenever a changeover straddles a crew change. The outgoing machine setter documents the current IS timing card and the forehearth temperature log. The incoming setter countersigns before assuming control. That 30-minute overlap eliminates what field data consistently show is 35–45% of excess post-change scrap in plants without this discipline.

And this is exactly why an OEM-affiliated consultancy reviewing job-change KPIs from a control-room dashboard won't find it. The data doesn't signal a changeover event. It signals random variation. Until you annotate the shifts and sit with the operators through a handover, the root cause stays invisible. There is no substitute for that work.

Why EU regulation is making Italian changeover discipline the competitive baseline

Italy's container glass sector emits approximately 2.1 million tonnes of CO2 per year at a specific intensity of around 0.38 tCO2 per tonne of glass, tight against the EU ETS product benchmark of 0.385 tCO2 per tonne. Under EU ETS Phase IV, the linear reduction factor rose to 4.3% per annum from January 2024. Every tonne of glass lost to job-change scrap is CO2 spend with no product revenue behind it. The Packaging and Packaging Waste Regulation's mandate for 70% recycled cullet content in glass packaging by 2030 tightens the picture further: at cullet ratios above 60%, furnace pull temperatures drop by up to 50°C and the forming viscosity window narrows, amplifying the quality sensitivity of every changeover.

CBAM full implementation from 1 January 2026 applies the prevailing EU ETS carbon price, forecast at €65–80 per tonne CO2, to container glass imported from non-EU suppliers. That changes the cost structure for every Italian producer defending domestic shelf space. Verallia's FY 2023 results attributed part of its Southern and Western Europe segment's leading operating leverage to shorter job-change cycles enabling higher SKU throughput per furnace-day. The advantage is measurable and it compounds. It only accrues to plants that have actually systemised the changeover, not just reduced the average time on paper.

If your plant is running full-machine-stop changeovers on a high-mix line and needs a decarbonisation story that holds at board level, the job-change cycle is where you start. A vendor-neutral forming audit will show you exactly where you sit against Italian SBS benchmarks. The Job Change Tool gives you the systemised execution layer to close that gap, with nine-stage lifecycle management, section-level KPI tracking, and shift-handover discipline built in. As an independent container glass consultant, I have run this process across plants in Europe, North America, and the GCC. If you want to know where your operation stands, reach out through our Italian container glass advisory page.