What is Mean Downtime (MDT) in Manufacturing

Most plants track Mean Time To Repair (MTTR). It measures from "technician starts work" to "machine running again." Typical EU packaging line MTTR is 35-90 minutes.

That number sounds fine. The problem is what comes before "technician starts work." The wait. The latency tax.

Mean Downtime (MDT) covers the full event:

• Time from failure to alarm reaching maintenance: 8-25 minutes (manual reporting), 0-2 minutes (automated capture)
• Wait for technician to arrive at the asset: 15-45 minutes (depends on shift coverage and plant size)
• Wait for spare parts: 10 minutes (on-shelf) to several days (special order)
• Actual repair time: this is what MTTR captures, 35-90 minutes typical
• Wait for approval to restart, QA sign-off, line warm-up: 10-30 minutes

Add it up. MDT on a typical EU packaging line is 2-5 hours per failure. MTTR is 35-90 minutes. The other 60-80% is the latency tax that never appears on dashboards.

See the OEE + MTBF + MTTR + MDT stack for how all four metrics fit together.

Quick answer: Mean Downtime (MDT) is the average total time a system is unavailable per failure, including both repair time (MTTR) and waiting time (parts, technician, approval). MDT = MTTR + logistics delay. World-class plants target MDT under 2 hours; reactive plants commonly sit at 6-12 hours because waiting time dominates.

Related deep-dives: the MTBF / MTTR paradox · increase MTBF using native OEE · MTTR + MTBF analytics · closing the OEE-CMMS loop.

If your maintenance team is proud of their MTTR, that is good. But if MDT is 3-5x higher, they are running a fast pit stop after a 4-hour walk to the pit lane.

Switch your weekly maintenance review to MDT. Same machine, same failures, but the number now includes everything between failure and restart. The latency immediately becomes visible. The conversation shifts from "we fixed it fast" to "why did the fix take so long to start?"

A modern OEE solution with native CMMS calculates MDT automatically per asset. You see the latency tax broken down by category every day, not after a quarterly audit. That is the difference between Fabrico and an MTTR-only scorecard.

The formula is simple:

MDT = Total downtime hours ÷ Number of failures

Where "total downtime hours" includes EVERY minute from failure detection to ready-to-run, not just repair. Three real numbers a typical EU packaging plant tracks for one quarter:

• Total unplanned downtime: 184 hours
• Number of unplanned failures: 62
• MDT = 184 ÷ 62 = 2.97 hours per failure

That same plant's MTTR for the same period was 52 minutes. The gap is 2.97 hours - 0.87 hours = 2.1 hours of pure latency per failure. Across 62 failures that is 130 hours, the equivalent of 16 full shifts lost to waiting.

EU benchmarks for MDT on packaging lines:

• Top 10%: under 1.5 hours per failure
• Median: 2.5-4.0 hours per failure
• Bottom 25%: over 5.5 hours per failure

See how MDT and MTTR feed into OEE.

If you only track MTTR, you optimize the wrong thing. The 35-90 minute repair time is rarely the lever. The 2-3 hours of waiting around it is.

Three places the latency hides and the fix for each:

Failure-to-alarm gap. Operators discover the failure, walk to the terminal, log the event, supervisor calls maintenance. Easy 20 minutes lost.

• Fix: automated capture via PLC or Computer Vision. Stop is detected and routed to the right technician in under 90 seconds.

Spare-parts wait. The technician shows up, looks at the machine, walks to stores, finds the part missing, orders it.

• Fix: failure-mode-driven spare parts policy tied to MTBF data. Top 10% of EU plants have 95%+ on-shelf availability for critical-path parts.

Approval-to-restart gap. Repair is done. QA needs to check. Supervisor needs to sign. Line needs warm-up. 20-30 minutes more.

• Fix: pre-defined restart playbooks with QA criteria built into the CMMS. The technician hits a button and the approval flow runs in parallel with warm-up.

See the pit stop strategy for how preparation eliminates wait time.