OEE loss waterfall with denominator discipline

OEE loss waterfall with denominator discipline

Declare the planned production time before calculating equipment effectiveness. In this exercise, 360 run minutes from 450 planned minutes gives 80% availability. Six hundred units at an ideal half-minute each represent 300 ideal minutes, giving 83.33% performance during run time. Of those units, 570 are good first time, giving 95% quality. Multiplying the three factors yields 63.33% OEE, equivalent to 285 good ideal minutes out of 450. Investigate the underlying loss rather than chasing the score. Reworked output is not first-pass good, and making unwanted inventory cannot establish better customer flow.

Follow the method

1. Planned time
2. Run time
3. Ideal total time
4. Good ideal time

Reconcile OEE to good ideal time

Fictional machine-shift record: 450 minutes were planned for production, the machine ran 360 minutes and made 600 units, of which 570 were first-pass good. The approved ideal cycle for this exercise is 0.5 minute per unit. Analyst Tessa must explain the loss picture before recommending improvement work.

Follow the method

1. Availability
2. Performance
3. Quality
4. OEE
5. Reconcile

One percentage, three different loss questions

Reconcile OEE factors and equivalent-time losses, retain first-pass quality accounting and choose a loss investigation without mistaking the percentage for a cause.

Fictional machine-shift record: 450 minutes were planned for production, the machine ran 360 minutes and made 600 units, of which 570 were first-pass good. The approved ideal cycle for this exercise is 0.5 minute per unit. Analyst Tessa must explain the loss picture before recommending improvement work.

Role: Production analyst with operator, quality and maintenance representatives.

Normal condition

Time scope, count definitions and ideal cycle are explicit and consistent; the factors reconcile to good ideal time divided by planned time.

The gap

The dashboard shows a low result but the team is treating all lost time as downtime and counting reworked units as first-pass good.

• Planned-production time and ideal cycle are supplied teaching inputs; real use requires a controlled definition.
• Quality and performance losses below are ideal-time equivalents, not necessarily separately observed clock intervals.

1. Calculate availability

   Divide 360 run minutes by 450 planned minutes to obtain 80%. The difference is 90 minutes of availability loss under the supplied time definition.

   Why: The denominator must remain the agreed planned scope; removing a loss from the denominator would improve the number without improving the work.

   Evidence: 450 = 360 + 90 minutes.

2. Calculate performance

   600 total units at 0.5 ideal minute require 300 ideal minutes. Divide 300 by 360 run minutes to obtain 83.33% performance.

   Why: Performance compares ideal output time with run time; it is not another downtime subtraction.

   Evidence: 360 minus 300 = 60 minutes performance loss equivalent.

3. Keep first-pass quality intact

   Divide 570 first-pass good units by 600 total units to obtain 95% quality. Good ideal time is 570 × 0.5 = 285 minutes.

   Why: Later rework may recover product but does not undo the original failure to be good first pass.

   Evidence: 300 minus 285 = 15 minutes quality loss at ideal time.

4. Reconcile and route learning

   Multiply 0.80 × 0.833333… × 0.95 = 0.633333…, or 63.33%. Check 285 / 450 gives the same result. Review loss events with the responsible teams before selecting an investigation.

   Why: OEE locates loss categories but does not identify a failed component, prove a cause or justify maximizing every machine’s output.

   Evidence: 90 + 60 + 15 + 285 = 450 minutes; actual causes require event evidence.

Twenty rejected units are recovered later

The 30 non-first-pass units include 20 subsequently reworked and accepted. A colleague proposes changing first-pass good from 570 to 590.

Keep first-pass quality at 570 / 600 = 95% for this defined original pass and record rework/recovery separately. Clarify how any rework resource/time enters the relevant reporting scope.

Recovery changes disposition, not first-pass history. Mixing scopes can count output or time twice.

Retain the original counts, rework identities, subsequent disposition and approved time-accounting rule.

Run a practice session

Materials

• Native 450 → 360 → 300 → 285 waterfall
• Blank calculation sheet
• Calculator; first-pass/rework tokens

1. Define denominators · 5 minutes

   Which time is planned, run and ideal?

2. Reconcile the worked shift · 8 minutes

   Where do the 15 quality-loss minutes come from?

3. Calculate changed shift · 10 minutes

   Can both OEE paths agree?

4. Debrief recovery and action · 5 minutes

   What event evidence is needed before choosing a repair?

Debrief

• Check scope before correcting a formula.
• Make the learner label equivalent time versus observed clock time.

Calculate factors and direct good-ideal-time ratio independently, then use their agreement as a reconciliation check.

Transfer into the work

Owner: Performance-data owner with production, quality and maintenance

Record: Controlled shift loss record linked to time and unit events

Review: At loss review and after a targeted countermeasure trial

Evidence: Consistent definitions, reconciled counts/time and observed loss mechanisms

Repair data scope or investigate event records before changing targets or attributing causes.