OEE — Overall Equipment Effectiveness — is a useful metric. But as a cost accounting tool for unplanned downtime, it consistently understates the real number. In our work with discrete manufacturers, we've found that the first downtime cost estimate a facility produces is typically 30–60% below the actual figure once you add the costs the OEE formula doesn't capture. Getting the real number matters, because your PdM investment case depends on it.
What OEE Actually Measures (and What It Doesn't)
OEE is the product of three factors: Availability × Performance × Quality. In the downtime context, Availability is the one that matters — it's the ratio of planned production time to actual run time. An asset that was scheduled to run 8 hours and ran 6 hours has 75% Availability.
The problem is that OEE measures the asset's output in isolation. It doesn't capture what happens upstream and downstream when that asset stops. It doesn't account for overtime to recover lost production. It doesn't include the premium paid for emergency parts. And it doesn't touch the long-term cost of running a machine in a degraded state before it finally trips — accelerated wear on adjacent components, increased scrap rate, quality escapes that made it past the sensor before the line went down.
When we help facilities build full downtime cost models, we structure the calculation in four layers. Use this framework against your own CMMS and production data — the inputs are available in systems most facilities already run.
Layer 1 — Direct Downtime Cost
This is the layer OEE captures. Calculate it as:
Downtime cost = Duration (hours) × Throughput value per hour × OEE impact factor
Throughput value per hour is your net revenue per production hour on that line or asset — not just variable margin, but the full value of units that weren't produced. For a stamping press feeding an assembly line at a Tier-1 automotive supplier, this number typically runs $15,000–$45,000 per hour depending on part mix and customer requirements. For a machining cell in a job shop environment, it's lower, maybe $4,000–$12,000 per hour, because the work can often be rescheduled. Know your number before you start the cost model.
The OEE impact factor accounts for cascading effects on the line. If one asset stops and it's in a single-path process, the entire line stops — the factor is 1.0 or higher. If there's buffer or a parallel path, the factor is less than 1.0. Many plants don't calculate this precisely, which leads to underestimation of single-path asset criticality.
Layer 2 — Reactive Maintenance Premium
Emergency maintenance costs more than planned maintenance. The premium shows up in three places.
Parts cost: Emergency procurement of a bearing, seal, or gearbox component outside of normal purchase orders typically costs 35–50% above the planned catalog price. Overnight freight adds $200–$600 per shipment. If the part isn't in stock domestically and needs to be air-freighted from overseas, that number climbs to $2,000–$5,000 in freight alone for a single critical component.
Labor premium: Calling in technicians outside of scheduled shifts — weekend callout, overnight emergency — typically carries a 1.5–2.0× wage multiplier depending on labor agreements. A four-hour repair at $85/hour base rate costs $340 planned and $510–$680 on emergency callout. Multiply by a two-person crew and the shift differential adds up fast.
Contract maintenance escalation: Facilities that use contract maintenance for specialized work — gearbox rebuilders, alignment specialists — pay spot rates during emergencies that are 20–40% above their negotiated contract rates.
Sum all three and add to your Layer 1 figure. For a typical six-hour unplanned failure event at a mid-size discrete manufacturer, Layer 2 adds $3,000–$8,000 to the direct downtime cost.
Layer 3 — Production Recovery Cost
Lost production rarely just disappears. Most facilities have customer commitments that require missed units to be made up — on overtime, on a weekend, or by displacing other scheduled work. Recovery cost is the incremental expense to produce those units outside of normal planned production.
Overtime premium on a two-shift facility running a four-hour recovery shift adds approximately 25–35% to the direct labor cost for that production block. Tooling wear on machines pushed harder during recovery runs can add 5–10% to consumable cost for that period. Expedited outbound freight to meet customer delivery commitments after a production disruption is another frequent hidden cost — $500–$2,500 per shipment for priority logistics.
Some of these costs never get attributed to the original downtime event in the CMMS. They show up in the production budget as "overtime" or in logistics as "expedited freight," disconnected from the equipment failure that triggered them. This is the primary reason facilities undercount downtime cost by 30–60% — the recovery costs are real but invisible in the maintenance reporting system.
Layer 4 — Quality and Scrap Costs
Equipment running in a degraded state before failure produces more defects. A spindle with early bearing wear runs with increased runout, increasing surface roughness variation on machined parts. A hydraulic press with a developing seal leak produces inconsistent forming force, leading to dimensional variation. These quality signals often appear in SPC data before the machine finally trips — but the cost of the increased scrap and rework accumulates in the quality budget, not the maintenance budget.
After an unplanned failure and restart, there's also a warm-up and calibration period during which the first production units should be inspected carefully. Facilities that skip this step — because they're already behind and feel pressure to run — sometimes ship non-conforming parts made during the restart window. The cost of a field quality escape is orders of magnitude above the original downtime cost: warranty claims, customer sorting costs, and potential line-shutdown penalties at the customer's facility.
The full cost of an unplanned failure is a maintenance number, a production number, a logistics number, and a quality number. Most facilities only report the first one.
Putting the Model Together
Here's what the full four-layer model looks like for a representative unplanned failure event — a gearbox failure on a production-critical transfer line at a mid-size US manufacturer:
| Cost Layer | Inputs | Example Range |
|---|---|---|
| Layer 1: Direct downtime | 6 hrs × $25,000/hr throughput value | $150,000 |
| Layer 2: Reactive maintenance premium | Emergency parts + overnight labor callout | $4,500 |
| Layer 3: Production recovery | Weekend overtime + expedited freight | $18,000 |
| Layer 4: Quality / scrap | Increased scrap during restart window | $3,200 |
| Total | $175,700 |
The OEE calculation in this example would capture approximately $150,000. The full model produces $175,700 — 17% higher in this case, though we've seen the gap reach 50–60% in facilities with high overtime costs and customer-facing quality implications.
Using the Model to Prioritize Your PdM Investment
Once you have the four-layer cost model built, it becomes the basis for prioritizing which assets to instrument first. Rank your assets by their expected annual failure cost under the current reactive model, weight by criticality, and instrument the top quintile first.
The data consistently shows that 20% of the asset fleet accounts for 70–80% of total unplanned downtime cost when the full model is applied. Condition monitoring doesn't have to cover everything to deliver most of the return — it has to cover the right assets.
The cost model also gives you the denominator for your ROI calculation. A predictive program that reduces unplanned failures by 60% on a $400,000 annual downtime cost base delivers $240,000 in avoided costs. Program cost comparison against that number is straightforward. Building the model rigorously — all four layers — ensures you're comparing against the real baseline, not a partial one. That's the difference between a business case that gets funded and one that stalls in finance review.
