Service · Operations intelligence

Waste and quality

Find what actually causes scrap and rework, and reduce it.

Relevant to Manufacturing

The problem

Material and rework eat your margin, and every supervisor has a different theory about why.

QC results, batch records and raw-material lots sit in separate files. Linked together, they usually point at two or three causes.

What it costs today

Typical for process and discrete manufacturing: 2 to 6 percent of material cost in scrap, and rework hours that never appear on any report.

This is for you if
  • Rejection or rework is above 2 percent of material and nobody agrees on why.
  • QC results exist per batch or per shift, even on paper.
  • You are willing to change a setting or a supplier when the data says so.
What we build
  • A ranked list of what drives rejection, by machine, operator, lot, shift and parameter.
  • Batch quality prediction before the batch finishes, so the operator can correct.
  • A weekly quality report that operations managers actually open.
  • Where it pays, a model on in-process readings; where it does not, a clear set of control limits.
How we build itIn this order
  1. 01

    Join QC results to their causes

    Every inspection result gets its batch, machine settings, raw-material lot, operator and shift attached. Most plants have never seen this table.

  2. 02

    Rank before you model

    A ranked list of what drives rejection solves half the problem on its own. We present it to the supervisors first and let them argue with it.

  3. 03

    Predict in process where it pays

    Where in-process readings exist, a model predicts the batch result early enough for the operator to correct. Where they do not, we set control limits and leave it there.

  4. 04

    Make it a weekly habit

    A one-page report, three numbers, the reasons they moved. Reviewed in the production meeting, not in an inbox.

  5. 05

    Measure like for like

    Rejection on the same product, same season, before and after. We do not count a good month as a result.

Data you already haveThis is usually enough to start
  • Production batch records and machine settings
  • QC and inspection results, including the rejected ones
  • Raw-material lot and supplier records
  • Operator and shift mapping
In practice

Illustrative example

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

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.

Read the case study
Typical engagement
Duration2-week assessment, then 6 to 8 week pilot on one product line
TeamEngineering lead, data engineer, quality analyst
You receiveRoot-cause ranking, live quality dashboard, before/after rejection rates
PriceFixed price, quoted after the readiness assessment
What you own at the end
  • The joined quality table and the pipelines that keep it current, in your environment.
  • The dashboard and the weekly report, with the control limits documented.
  • The QC entry form, if one was built, and the training for the people who use it.

Everything runs in your environment and is handed over with documentation. Nothing stops working if we part ways.

Questions we are usually asked

Our QC data is on paper.

We digitise a sample during the assessment and tell you if the signal is there. If it is, a tablet-based QC entry form is usually part of the pilot.

Will this blame operators?

The analysis shows where variation comes from. In most plants it is settings, lots and machine condition more than people. We present it that way.

How is this different from Six Sigma work we have done?

It is the same intent with the full history instead of a sample, and it keeps running after the project ends.

Can it connect to our ERP?

Yes, read-only. We take exports or a database view; we do not modify the ERP.

Can a camera do the inspection?

For surface defects, missing parts and label errors, often yes. That is a separate engagement, visual inspection with cameras, and this analysis usually tells you whether it is worth it.

Often combined with

Further reading

Not sure if your data is ready?

Take the free 20-minute readiness check. You get a short scorecard and a plain answer about what, if anything, is worth doing next.

Start the readiness checkTalk to us