Case study · Illustrative example

Reduced rejection on a flexible-packaging line from 4.8% to 2.9%

Illustrative example. This shows how an engagement of this kind typically runs. It is not a report of a specific client's project, and the figures are typical outcomes, not measured client results. Your numbers depend on your data; the readiness assessment tells you what to expect.

The typical situation

Rejection on the main line had drifted up over two years and every supervisor had a different explanation: the new ink supplier, the night shift, the humidity, a particular operator.

QC results were recorded on paper, then typed into a monthly summary that lost the batch-level detail.

What the data looked like

Eighteen months of batch records from the ERP, paper QC sheets digitised for a six-month sample, raw-material lot records and the machine settings log. Ambient humidity from the plant's own logger.

What we built

  • A ranked list of rejection drivers: two raw-material lots and one machine-setting combination explained most of the variance; shift and operator explained almost none.
  • A tablet-based QC entry form so batch-level results stopped disappearing into monthly summaries.
  • Control limits on the two settings that mattered, shown on the operator screen.
Your systemsClean data layerModel or rulesAlerts and reports

Typical outcome

MeasureBeforeAfterPeriod
Rejection rate4.8%2.9%3 months before vs 3 months after
Rework hours~160 / month~95 / monthsame period

What usually trips these projects up

The machine-learning model we built in week six added little over the control limits from week four. We kept the simpler system in production and said so to the client.

Have a similar problem?

Tell us what is going wrong and what data you keep. The first conversation is free and usually ends with a straight answer about whether this is worth doing.

Start the readiness checkTalk to us