What is MTTR and which of the four MTTR variants do you actually want?

Last verified: 16 May 2026. Mean Time To Repair (MTTR) is a maintainability metric defined in MIL-HDBK-338B (Electronic Reliability Design Handbook, U.S. Department of Defense, 1 October 1998, Section 3) as the average time required to restore a failed item to operational condition. The Society for Maintenance and Reliability Professionals (SMRP) Best Practices 5th edition (2017) provides the practitioner-oriented definition aligned to plant maintenance operations.

“Mean Time To Repair (MTTR): The total corrective maintenance time divided by the total number of corrective maintenance actions during a given period of time.” — MIL-HDBK-338B, Section 3, U.S. Department of Defense, 1998.

The most common source of confusion when reporting MTTR is that the acronym is used for four different metrics in engineering practice: Mean Time To Repair, Mean Time To Recovery, Mean Time To Resolve, and Mean Time To Respond. Each measures a different time interval and supports different operational decisions. This entry distinguishes them against the MIL-HDBK-338B and SMRP definitions.

The four MTTRs and what each measures

• Mean Time To Repair (MIL-HDBK-338B): the active maintenance time from start of repair to completion of repair. Excludes wait time, parts logistics, and administrative delay.
• Mean Time To Recovery (SMRP 5.5.2.1): the total time from failure occurrence to return to operational status. Includes detection, response, repair, verification.
• Mean Time To Resolve: used in IT and software contexts. Time from incident creation to ticket closure. Not directly relevant to industrial manufacturing.
• Mean Time To Respond: the time from failure detection to the start of repair activity. Often tracked as a separate KPI for maintenance crew dispatch performance.

For OEE Availability calculations under ISO 22400-2:2014, the relevant interval is the full downtime duration — equivalent to Mean Time To Recovery. For maintenance crew performance reviews, the relevant metric is the active repair time alone — Mean Time To Repair in the strict MIL-HDBK-338B sense. Mixing the two is a frequent source of measurement error.

The MTTR formula

MTTR = Total Maintenance Time ÷ Number of Maintenance Actions

For a packaging line experiencing 12 failure events over 90 days with total repair time of 47.5 hours, the MIL-HDBK-338B MTTR is 47.5 ÷ 12 = 3.96 hours per repair. If the failure-to-recovery time (including the 18 hours of cumulative wait time for parts) is 65.5 hours total, the SMRP Mean Time To Recovery is 65.5 ÷ 12 = 5.46 hours per failure.

MTTR and the Availability identity

MIL-HDBK-338B §5.8 gives the canonical steady-state availability identity for repairable systems:

Availability = MTBF ÷ (MTBF + MTTR)

This identity assumes both MTBF and MTTR refer to the same time accounting; conventionally, both are measured from the equipment state stream and include the full downtime duration. Using the strict MIL-HDBK-338B MTTR (repair-time-only) in this identity overstates Availability because it excludes the wait time that the equipment is actually unavailable.

The lognormal MTTR distribution

Unlike failure times under the constant-failure-rate assumption (which follow an exponential distribution), repair times typically follow a lognormal distribution per Ebeling 1997 §10.4. This has two implications:

1. The mean MTTR overstates the typical repair time. The median repair time is often 30-50% shorter than the mean. Reporting both mean and median is recommended.
2. The tail of the MTTR distribution dominates Availability impact. A handful of long repairs (parts unavailable, specialist required) drives the average; reducing the tail through spare parts strategy and crew training has greater Availability impact than incremental improvement on short repairs.

“Repair times tend to be lognormally distributed, with the right tail dominating average behavior. Maintenance planning should target the tail.” — Ebeling, C.E., An Introduction to Reliability and Maintainability Engineering, McGraw-Hill, 1997, §10.4.

Worked example: comparing MTTR variants on a real line

An automotive Tier-1 paint line experiences 8 failure events in 30 days. The event log captures:

• Time from failure occurrence to alarm acknowledged by maintenance: 0.4 hours average (Mean Time To Respond)
• Time from alarm acknowledged to start of physical repair: 1.2 hours average (typically wait for parts or technician)
• Time from start of repair to repair complete: 2.1 hours average (Mean Time To Repair, MIL-HDBK-338B)
• Time from repair complete to line returning to spec output: 0.6 hours average (warm-up and quality verification)
• Total: 4.3 hours per failure (Mean Time To Recovery, SMRP)

The MIL-HDBK-338B MTTR is 2.1 hours; the Mean Time To Recovery (the value that drives Availability) is 4.3 hours. Reporting only the MIL-HDBK-338B figure to operations management would understate the operational impact by a factor of two.

MTTR improvement levers

Smith and Hinchcliffe 2003 in RCM-Gateway to World Class Maintenance identify four primary MTTR improvement levers, ordered by typical impact:

1. Spare parts strategy — kit parts at the equipment, not at central stores. This single change typically reduces tail MTTR events by 40-60%.
2. Diagnostic instrumentation — sensors that pinpoint the failure mode reduce diagnostic time, often the largest single MTTR component. JEMBA AI’s root-cause analysis layer targets this directly.
3. Standard work procedures — documented repair sequences reduce variability in repair time and accelerate technician training.
4. Crew skill matching — assigning the right specialist to the right failure mode (electrical to electrical, hydraulic to hydraulic). Misassignment is a major source of tail MTTR events.

Frequently asked questions

What is the difference between MTTR and MTTF?

MTTR is a maintainability metric measuring repair duration. MTTF is a reliability metric measuring the lifetime of non-repairable items before failure. The two are not comparable.

Why does my MTTR figure differ from my maintenance manager’s MTTR figure?

The most common reason is the two of you are tracking different MTTR variants — Mean Time To Repair (active repair time) versus Mean Time To Recovery (failure-to-return-to-service). Reconcile the time accounting before comparing numbers.

What is the world-class MTTR benchmark?

No single benchmark exists because MTTR is highly context-dependent. SMRP Best Practices 5th edition provides industry-specific benchmark ranges; for discrete manufacturing the typical world-class range is 30 minutes to 2 hours for the MIL-HDBK-338B definition.

How does MTTR relate to Availability?

Under MIL-HDBK-338B §5.8: Availability = MTBF ÷ (MTBF + MTTR). Higher MTTR depresses Availability.

Should MTTR be reported as mean or median?

Both. The mean is driven by the lognormal right tail; the median represents the typical experience. Reporting only the mean obscures the operational signal.

Can MTTR be improved through preventive maintenance?

Preventive maintenance primarily improves MTBF (frequency of failures), not MTTR (duration of repair). MTTR improvement requires spare parts strategy, diagnostic instrumentation, and procedure standardization.

How does TeepTrak measure MTTR?

TeepTrak measures Mean Time To Recovery by default — the full downtime duration captured in the equipment state stream. Mean Time To Repair (excluding wait time) is reported as a secondary metric when the maintenance CMMS integration is configured.

What is a reasonable MTTR target for a new line?

Targets should be derived from the failure mode distribution and the supporting maintenance organization capacity. Setting an MTTR target before the failure modes are characterized is the most common error in new line commissioning.

Is MTTR the same in IT incident management as in manufacturing?

No. In IT, MTTR commonly refers to Mean Time To Resolve (ticket closure), which can include administrative steps not present in physical equipment repair. The two are not interchangeable.

How does MTTR interact with the Six Big Losses framework?

MTTR is the time component of the Downtime Loss category (Big Loss #1 in Nakajima 1988). High MTTR concentrates downtime impact regardless of failure frequency.

References

1. MIL-HDBK-338B (1998). Electronic Reliability Design Handbook. U.S. Department of Defense, 1 October 1998.
2. SMRP (2017). Society for Maintenance & Reliability Professionals Best Practices, 5th edition.
3. IEC 60050-192:2015. International Electrotechnical Vocabulary – Part 192: Dependability. International Electrotechnical Commission, Geneva.
4. Ebeling, C.E. (1997). An Introduction to Reliability and Maintainability Engineering. McGraw-Hill. ISBN 0-07-018852-1.
5. Smith, A.M. and Hinchcliffe, G.R. (2003). RCM-Gateway to World Class Maintenance, 2nd edition. Butterworth-Heinemann. ISBN 0-7506-7461-X.

Author: Bastien Affeltranger, CTO, TeepTrak. Cross-references: MTBF, OEE, Six Big Losses. Last verified 16 May 2026 against MIL-HDBK-338B and SMRP Best Practices 5th edition.