Maintenance

Driving MTTR Down: A Maintenance Response Playbook

MTTR decides whether a breakdown costs 45 minutes or half a shift. Here is how to measure it honestly, benchmark it by asset class, and run the daily cadence that drives it under 2 hours.

Mean time to repair decides whether a breakdown costs you 45 minutes or half a shift. Run a line worth $8,000 an hour in contribution margin and the difference between a 1.5 hour MTTR and a 4 hour MTTR is $20,000 per event. A plant averaging 12 breakdowns a month at that rate leaves $240,000 on the table every month it tolerates slow repairs. Availability math makes the same point: with an MTBF of 90 hours, cutting MTTR from 4 hours to 2 lifts availability from 95.7 percent to 97.8 percent, which on a 6,000 hour operating year is 126 extra production hours.

Measure it the simple way first: total repair labor hours divided by completed repairs over the period. If your crew logged 84 repair hours across 42 completed work orders last month, MTTR is 2.0 hours per repair. The MTTR calculator does that division for you; the hard part is defining the clock. Decide whether the clock starts at failure notification or technician arrival, and whether it stops at machine restart or first good part. A plant that measures notification to first good part will report 30 to 50 percent higher MTTR than one measuring wrench time only. Pick one definition, write it down, and never change it mid year.

Benchmarks depend on asset complexity, but useful ranges exist. World-class discrete manufacturers hold MTTR between 1 and 2 hours on production equipment. Typical plants sit at 3 to 5 hours. Anything above 6 hours signals structural problems: parts not stocked, no troubleshooting documentation, or repairs waiting on a single expert. Segment the number before comparing. A CNC spindle replacement legitimately takes 8 hours; a jammed sensor should take 20 minutes. Report MTTR by asset class and by failure code, because a blended 3.2 hour average can hide a conveyor fleet at 45 minutes and a press line at 9 hours.

Break MTTR into its five segments and attack the largest first: detection, notification, diagnosis, repair, and restart. In most plants diagnosis eats 40 to 60 percent of the clock, and repair itself only 20 to 30 percent. Levers that pay: troubleshooting guides for the top 10 failure modes cut diagnosis 30 to 50 percent; a kitted parts program with 95 percent stockroom service level removes 45 to 90 minute parts hunts; radio dispatch instead of walk-and-find saves 15 minutes per event. One food plant cut MTTR from 3.8 to 2.1 hours in six months purely by staging failure-mode kits at the line.

Watch for the ways the number lies. Technicians who close work orders in batches on Friday destroy the timestamps; audit 10 random work orders a week for clock accuracy. Chasing MTTR alone drives quick patches that return as repeat failures: if MTTR drops 20 percent while MTBF drops 30 percent, you lost ground. Excluding long repairs as outliers is another trap; that 14 hour gearbox job is exactly the event to study. And a rising repair count with flat MTTR means reliability is eroding even though the ratio looks stable, so always read MTTR next to failure frequency.

Run it on a cadence. Daily: review every repair over 2 hours from the prior 24 hours in the morning meeting, 5 minutes max, asking which segment ate the time. Weekly: trend MTTR by line and by top 5 failure codes, and assign one countermeasure per week with an owner and a date. Monthly: recalculate MTTR by asset class, compare against the 12 month trend, and reprioritize the kit and documentation backlog. Quarterly: audit clock definitions and CMMS data quality on a 25 work order sample. Plants that run this rhythm typically see 15 to 25 percent MTTR reduction in the first two quarters.

World-class looks like this: MTTR under 2 hours plant wide, under 1 hour on constraint equipment, 90 percent of repairs completed within twice the median, and zero repairs waiting on parts more than 30 minutes. Diagnosis time is under 25 percent of the total clock because the top 20 failure modes have documented procedures with photos. Every repair over 4 hours gets a written 5-why within 48 hours. And the number is public: posted at the line, reviewed daily, owned by the area supervisor rather than buried in a maintenance office. When operators can quote last week's MTTR, you are running it as a management system.

Published 2026-07-02.