Reliability

Raising Maintenance Availability Against the Calendar

Of the 8,760 hours the calendar gives an asset, how many does maintenance hand back? A playbook for the calendar-basis number: PM optimization, startup recovery, and the cadence that recovers 2 to 4 points.

Maintenance availability answers a harder question than uptime: of all 8,760 hours the calendar gives an asset in a year, how many did maintenance actually deliver back to production? A 24/7 plant that loses 400 hours to breakdowns, 300 to preventive work, and 120 to slow startups after maintenance is running 90.6 percent, and the 820 lost hours are the entire capital case for reliability spending. At $6,000 an hour of contribution on a constraint asset, each availability point on the calendar basis is worth 87.6 hours, about $525,000 a year. That is why finance should see this number, not just engineering.

Compute it against calendar time, not planned time, so nothing hides. Take 720 hours in a month, subtract 28 breakdown hours, 20 preventive maintenance hours, and 6 hours of post-maintenance startup recovery: 666 over 720 is 92.5 percent. The Maintenance Availability calculator handles the stack, including PM execution and startup recovery, which most plants forget to count. Startup recovery matters more than it looks: if every PM ends with 45 minutes of warmup, purge, and first-article checks, 4 PMs a week is 3 lost hours nobody logged. Track breakdown, PM, and recovery as three separate buckets, because each has a different owner and a different fix.

Calibrate against calendar-basis ranges: 95 percent plus is excellent for continuous operations, 90 to 95 is solid, and under 85 means maintenance is consuming the plant rather than protecting it. The mix matters as much as the total. A healthy plant runs a planned-to-unplanned ratio of 80/20 or better; a plant losing 60 percent of its maintenance hours to breakdowns will have a low number and no way to schedule improvement. PM programs should consume 3 to 6 percent of calendar time on most equipment; if yours takes 9 percent, either the PMs are bloated or they are compensating for design problems that should be engineered out.

Three levers move it. First, PM optimization: audit every PM against failure history, and expect to delete or extend 20 to 30 percent of tasks; plants that run PM reviews annually typically recover 1 to 2 availability points without adding risk. Second, schedule PMs into demand gaps and changeovers; converting half your PM hours to windows the schedule already lost is a free point. Third, kill startup recovery with documented restart procedures and pre-staged first-article checks; taking recovery from 45 to 15 minutes per event across 200 events a year returns 100 hours. Breakdown reduction still pays the most, but these three are faster.

Common failure modes: counting PM time as if it were free because it was planned, which teaches planners to schedule 6 hour PMs that need 3; letting deferred PMs pile up until compliance drops under 80 percent, at which point breakdown hours climb 6 to 12 months later; and booking startup scrap and slow ramp as a production problem so maintenance never sees the cost of a sloppy handback. Watch also for availability gained by deferral: skipping PMs lifts this month's number and mortgages next quarter. Pair the metric with PM compliance and MTBF so rising availability with falling compliance triggers alarm, not celebration.

Run the cadence. Weekly: schedule attainment review, PM compliance versus a 95 percent target, and every startup recovery over 30 minutes gets a reason. Monthly: recalculate calendar-basis availability by asset, split into breakdown, PM, and recovery buckets, compare against the 12 month trend, and assign one bucket-owner action per constraint asset. Quarterly: PM content review on the worst 10 assets, deleting tasks with no failure-history justification. Annually: rebuild the availability budget asset by asset and tie the reliability capital plan to the hours it buys back. Expect 2 to 4 points of improvement in year one if you start below 90.

World-class maintenance availability looks like 95 to 97 percent on a calendar basis, PM compliance above 95 percent, a planned-to-unplanned ratio of 85/15, and startup recovery under 15 minutes on 90 percent of handbacks. Every maintenance window has a countdown plan the way a pit crew does, and handback includes a first-good-part confirmation before the technician leaves. The number appears in the plant P&L review, translated into hours and dollars, and the maintenance manager can name the three worst assets and the recovery plan for each without opening a laptop. That is maintenance run as a production function, not a cost center.

Published 2026-07-02.