An MTTR reduction plan has four parts: measure where restore time goes, rank failures by hours lost, apply one fix per failure mode, then check the gain in availability and OEE. Waiting, diagnosis and parts are the first places to look, and your data decides. The 12-week plan below is a template to adapt.
The plan on one page
Time every repair by phase: notice, call, wait, diagnose, parts, repair, test, restart.
List failure modes by hours lost, which is count x mean restore time.
Apply one fix for each of the top three failure modes, with one named owner each.
Agree who gets called, and after how many minutes.
Compare with the baseline, then write the fix into standard work.
Use one definition from day one.
Run the clock from the failure until the machine works again, in line with the IEC term mean time to restoration. Some plants stop it at the first good product instead, so write your choice down.
Hands-on time alone would hide most of what this plan goes after. The calculation itself is on how to calculate MTTR.
Weeks 1-2: measure the phases
Put a one-line sheet at the machine, or read stop times from a signal. Ask the technician to note the arrival time, the time the fault was found, and the time the part arrived.
Do not chase accuracy to the second. You need to know whether diagnosis takes 5 minutes or 50.
The result is the baseline. If you have no target yet, what is a good MTTR shows how to derive one from these same phases.
Weeks 3-4: rank failures by hours, not by count
Example: a palletizing cell over one quarter with 1,000 planned hours. It logged 33 failures, three modes in all.
| Failure mode | Failures | Mean restore | Hours lost | Where the time goes |
|---|---|---|---|---|
| Drive trips | 18 | 40 min | 12.0 | Waiting for someone to reset |
| Gripper sensor | 12 | 65 min | 13.0 | Diagnosis, then finding a sensor |
| Gearbox | 3 | 480 min | 24.0 | Part not in stock |
| Total | 33 | 89 min | 49.0 |
By count, the drive trips look like the priority. By hours, the rare gearbox takes half the total, and each mode needs a different fix.
A downtime Pareto chart by hours makes this ranking visible. Keep the count beside it, since each failure also interrupts flow.
Weeks 5-8: match each fix to the phase that hurts
| Phase | Typical fix | Test that it worked |
|---|---|---|
| Notice | Stop signal, stack light, andon | Minutes from stop to first call |
| Call and wait | On-call rule, and rights for operators to do safe resets | Minutes from call to arrival |
| Diagnose | Fault cards at the machine, repair history, fault code lists | Minutes from arrival to fault found |
| Parts | Common spares near the line, exchange units for slow repairs | Minutes from fault found to part in hand |
| Repair | Tool kit per failure mode, written steps | Hands-on minutes |
| Restart | Short checklist, first-off check | Scrap in the first minutes after restart |
In the example, each failure mode gets one fix:
- Gearbox: a spare or exchange unit on the shelf cuts 480 minutes to 240. See spare parts management.
- Gripper sensor: a fault card and a spare sensor at the cell cut 65 minutes to 35.
- Drive trips: a code list and a trained, safe operator reset cut 40 minutes to 25. Manufacturers' codes are in the troubleshooting directory.
The total falls from 49 to 26.5 hours with the same number of failures. MTTR falls from 89 to 48 minutes, since 26.5 / 33 = 0.80 hours.
Spares carry a cost, so test each stocking choice before buying. The spare parts inventory calculator weighs the carrying cost against a stockout.
Weeks 9-10: set the clock for escalation
Long stops are often long because nobody decided when to call for help. Write down who is called after how many minutes, by name or role, for each shift.
A clock also protects the operator, who no longer has to judge when to ask. See the ladder in how to reduce unplanned downtime, and the routine in first response to a line stop.
Weeks 11-12: check the gain in OEE
Example: 49 hours lost in 1,000 planned hours is availability of 95.1% from failures. At 26.5 hours lost, availability is 97.4%.
At 86% performance and 97% quality, OEE moves from 79.3% to 81.2%, or 1.9 points. That is the number to show the plant, and the downtime cost calculator turns those hours into money.
Then write each fix into standard work, or the gain will fade. Fault cards go out of date, and spares drift out of stock unless someone owns them.
Guard metrics: how a plan like this goes wrong
- MTBF falls. Quick resets can leave the cause in place. Track MTBF and repeat failures within a day of the repair.
- Restart scrap rises. A rushed restart moves the loss from availability to quality.
- The numbers improve on paper. Reclassifying long repairs as planned work lowers MTTR and saves nothing, so check against total hours lost.
- Safety steps get skipped. Lockout and safe reset rules stay as they are, and they are never where you look for minutes.
Short stops under your threshold do not belong in this plan. They are a different problem, covered in how to reduce micro-stops.
This plan starts by knowing what happened during each stop. Where cameras are fitted, Fabrico can stream video of a stop on request next to its machine data, and operators tag the cause at the line. See how Fabrico uses cameras for OEE.
What people ask next
What is the fastest way to reduce MTTR?
Find the phase with the most waiting and fix that first. It is often response, diagnosis or parts, though your own timing decides. Start with first response to a line stop.
How do you measure MTTR reduction?
Compare the same failure modes before and after, using the same clock. Also track total hours lost, because MTTR alone can improve when only the failure mix changes. The MTBF and MTTR calculator helps with the sums.
How long does an MTTR reduction plan take?
This template runs for 12 weeks on one line, which is enough for a few fixes and a first check. Rare failures need longer before you can judge them, since a mode that happens three times a quarter gives little data.
Should I focus on MTTR or MTBF first?
Measure both and compare the hours each could win. Response fixes often come sooner, because they work on every failure, but prevention removes the loss for good. See how to improve MTBF.
Who should own the MTTR plan?
One person per failure mode, with production and technicians both at the table. Response time depends on both groups. Review it in the daily production meeting.
Fabrico is a light MES for OEE. It takes stops, speed and counts from PLCs or IoT sensors, with add-on AI cameras counting output where a machine has no usable signal, and operators tag the cause at the line.