Why Sampling for OEE?
OEE (Overall Equipment Effectiveness) is traditionally measured with sensors and SCADA systems. However, work sampling is anaffordable and fast alternative for:
- Plants without automated measurement infrastructure
- Initial diagnosis before investing in systems
- Validate/audit data from existing systems
- Identify the causes of low efficiency (not just the number)
The 3 Factors of OEE
Availability
% of time the machine is available to produce
Typical target: >90%
Performance
% of real speed vs nominal speed
Typical target: >95%
Quality
% of good parts vs total produced
Typical target: >99%
OEE Formula
OEE = Availability × Performance × Quality
Example: 85% × 90% × 98% = 75% OEE
What Sampling Can Measure
Work sampling is especially useful for measuring Availabilityand the causes of loss:
| Observed State | Category | OEE Impact |
|---|---|---|
| Producing | Productive | ✅ Available |
| Changeover | Planned delay | Availability loss |
| Breakdown | Unplanned delay | Availability loss |
| Waiting for operator | Management delay | Availability loss |
| Micro-stop | Frequent delay | Performance loss |
| Missing material | Logistics delay | Availability loss |
Case Study: Capacity Diagnosis
📋 Scenario
A plastic injection plant considers buying a new machine ($50,000) because it "can't keep up" with current demand.
Study performed:
- • 2 weeks of sampling
- • 400 observations
- • 95% confidence level
Sampling Results
Analysis
The 30% "Waiting for Operator" was the key revelation:
- The machine finishes its cycle but the operator is attending another machine
- A single operator handles 3 injection machines
- It is not a lack of machine capacity, it is a lack of staff
❌ Without the study
Buy new machine
Cost: $50,000 CAPEX
+ The problem would persist (still missing an operator)
✅ With the study
Hire an auxiliary operator
Cost: $15,000/year OPEX
+ Increases utilisation of the 3 existing machines
How to Configure the Study
In our app, create a study with these categories:
📂 PRODUCTIVE
├── Producing (normal cycle)
└── Producing (slow cycle)
📂 PLANNED DELAY
├── Changeover/mould change
├── Preventive maintenance
└── Scheduled cleaning
📂 UNPLANNED DELAY
├── Mechanical breakdown
├── Electrical breakdown
├── Waiting for operator
├── Waiting for material
├── Waiting for quality
└── Micro-stop (jam, adjustment)
Study ROI
A 2-week sampling study costs approximately:
- 10-15 hours of engineer for configuration and analysis
- 5 minutes per observation × 400 observations = ~33 hours of recording
- Total cost: ~$2,000-3,000 (at $50/hour)
If it avoids an unnecessary investment of $50,000, the ROI is 15:1.