Pack-to-melt: the container glass KPI that actually pays

Most hot-end superintendents I've sat with over the years quote OEE first. It's the number on the daily report, the number the VP asks about on the plant walk, the number that goes into the monthly deck. The problem is that OEE can improve while you're haemorrhaging glass. Pack-to-melt doesn't let that happen. That distinction matters more than most plants are willing to act on.

OEE can look fine while your melt cost bleeds

The mechanics are well understood. OEE measures availability, performance and quality against the nameplate rate of the IS machine. It does not see the furnace. It does not care what happens between the forehearth exit and the cold-end palletiser. You can run sections flat-out, post 88% OEE, and still report a pack-to-melt ratio that would make your finance team wince.

I audited a two-furnace, six-line container glass operation in Southeast Asia last year. Privately owned, respectable volume, and the team was proud of their efficiency numbers. They were running 87.3% OEE across the forming lines. Their pack-to-melt ratio was sitting at 82.1%. For the ware mix they were running, that number should have been 88% or better. The gap was roughly six percentage points of glass going nowhere: pulled from the furnace, processed through the forehearth, formed into bottles, and then lost to cullet, seeds, leaners and a cold-end rejection rate that nobody had traced back to forming root cause.

Six points on pack-to-melt at that plant's melt tonnage translated to several hundred thousand dollars of recoverable EBITDA per year. OEE told them they were doing fine.

Where the ratio actually bleeds

Pack-to-melt losses sit in three places on most plants I've assessed:

• Forming losses: gob weight variance, section timing drift, and plunger drift driving choked necks and heavy heels.
• Furnace and forehearth losses: seeds and blisters from fining problems, and temperature non-uniformity in the working zone that compounds gob weight instability.
• Job change waste: every hour the line runs a cold mould set with baffle marks and settle wave on the shoulder is melt pulled at zero pack value.

Forming is the obvious category, but it rarely accounts for the majority of the gap on a reasonably tuned line. The furnace and forehearth contribution is underestimated on almost every plant I've walked.

Seeds and blisters in finished ware trace back to fining problems: glass temperature, residence time, sulphate chemistry. Every seed-rejected pallet is melt energy consumed and then scrapped. The forehearth compounds this. A working-zone temperature profile running ±8°C instead of the target ±2°C across five zones will push gob weight CV above 0.6%. The accepted spec on most lines is ≤0.4%. That variance shows up at the packer, not on the IS machine efficiency report, which is exactly why OEE misses it.

Job changes are the third source, and the most controllable. The hot-end superintendent owns the job change execution clock. If that clock isn't being tracked with the same rigour as OEE uptime, you're flying blind on a significant chunk of your ratio losses.

Making the ratio visible to the shift team

The pack-to-melt ratio calculation is straightforward: good packs divided by tonnes pulled, normalised to standard weight. The hard part is surfacing it to the shift team in real time and tracing variances back to root cause fast enough to act on them.

Start with the data plumbing. Most container glass plants can calculate pack-to-melt from existing SCADA and packer data. It just isn't being surfaced at shift level. The 0600 handover misses night-shift cullet data roughly 70% of the time on lines I've audited. That gap means the incoming superintendent is making forming decisions without knowing whether the ratio degraded at 2am or held steady through the night.

Fix the handover first. Then build a shift-level pack-to-melt board. Not a weekly management report. A live board showing the target range, the actual, and the variance by section if you have the resolution. If a section's contribution to cullet jumps, the team should see it before the hour is out, not at the next morning meeting.

OEE tells you how well you ran the machine. Pack-to-melt tells you how well you ran the plant.

A vendor-neutral hot end audit is the fastest way to identify where the ratio is leaking. That kind of review covers the full chain: furnace fining, forehearth profile, gob conditioning, IS timing, inspection sensitivity and cold-end handling. An OEM-affiliated review will find issues in the OEM's scope of supply. A generic Lean/Six Sigma boutique will map your process against a template it applies to any industry. A full-chain, independent audit finds where the glass actually goes. The pack-to-melt glossary entry covers the calculation method in detail, and it's worth checking your current approach against it before you build the shift board.

The ownership question nobody likes to answer

The cultural shift is harder than the technical one. OEE is a machine metric. Pack-to-melt is a plant metric. Asking a hot-end team to own a number that includes furnace losses, forehearth chemistry, job change efficiency and cold-end rejection rates is a bigger ask than it sounds.

The way I've seen it done well is to decompose the ratio into its controllable components and assign ownership explicitly. The melt superintendent owns the furnace and forehearth contribution. The hot-end superintendent owns forming losses and job change execution. The cold-end team owns inspection sensitivity and handling damage. Each reports to the shift ratio. The shift feeds the weekly number. The weekly feeds the monthly EBITDA picture.

When the ratio is decomposed that way, a -2 percent-pack event in the early hours of a Tuesday doesn't stay mysterious. Someone can point to the forehearth temperature excursion logged at 01:45, or the section six plunger drift the previous shift noted and didn't correct before handover. Pack-to-melt improvement becomes a management discipline, not a project you commission once and forget.

If your plant is running 4-8 OEE points below nameplate, forming efficiency still matters and both metrics earn their place in the daily report. But if OEE is respectable and margin is still thin, the ratio gap is almost always where the answer sits. A structured hot end audit will tell you in days what another year of OEE reporting won't.