PlantOps BI

KPI Library / Production

MTTR (Mean Time To Repair)

Formula: Total repair time ÷ number of repairs
Typical range: Under 1 hour is strong for straightforward mechanical repairs; complex electrical or controls failures often run several hours even in a well-run shop

MTTR measures the average time it takes to restore a failed asset to production, from the moment it goes down to the moment it’s back to making good parts. Paired with MTBF, it separates “how often things break” from “how fast we fix them,” which are two different problems with two different owners.

What good looks like

MTTR depends heavily on failure complexity and spare parts availability, so there’s no single target that applies across asset types. A well-run maintenance shop with good spares on hand and clear repair procedures should see straightforward mechanical repairs resolved in under an hour. Complex electrical, controls, or custom-fabrication repairs legitimately take longer, and that’s not a maintenance failure, it’s the nature of the repair.

The clock-stopping definition is where MTTR gets gamed. If a technician logs the repair as “complete” the moment the machine restarts, but it takes another 40 minutes to run scrap before the first good part comes off, the reported MTTR looks better than the actual disruption to the schedule. Define “repaired” as first good part, not “power on,” and MTTR will stop flattering itself.

MTTR in Power BI (DAX)

With a maintenance events fact carrying repair start and end timestamps by asset:

Total Repair Minutes =
SUM ( fact_maintenance_events[repair_minutes] )

Repair Count =
CALCULATE (
    COUNTROWS ( fact_maintenance_events ),
    fact_maintenance_events[event_type] = "Unplanned Failure"
)

MTTR Hours =
DIVIDE ( [Total Repair Minutes] / 60, [Repair Count] )

Break MTTR Hours out by failure category (mechanical, electrical, controls, material). A rising plant-wide average is much more actionable once you know which category is actually driving it.

Common mistakes

  1. Stopping the clock at “power on” instead of “first good part.” This is the single most common way MTTR gets understated.
  2. Excluding travel and diagnostic time from the repair window. If a technician spends 20 minutes finding the right part before starting the fix, that’s part of the repair, not a separate bucket that disappears from the metric.
  3. Reporting a single plant-wide MTTR. It hides whether the problem is diagnosis time, parts availability, or hands-on repair skill, three very different fixes.