When doing more maintenance makes things worse

If preventive work is being completed on time and equipment is still failing, more preventive work will not help. Something else is wrong, and it is usually one of three things: the checks are looking at the wrong thing, the interval is not related to how the equipment fails, or the maintenance is causing the faults.

The third one surprises people

Opening a machine introduces risk. Seals get disturbed, fasteners get over-torqued, contamination gets in, a connector goes back on loosely. For equipment that fails randomly rather than with wear, a routine strip-down is a scheduled opportunity to break something that was working.

The tell is failures clustering in the days after a service. Sort your failure history by "days since last maintenance" and look at the left-hand end. If there is a spike there, you have found something.

Checks that look at the wrong thing

A monthly visual inspection of a pump that fails from bearing wear will pass every month and then the bearing will go. The check is being done properly and it cannot see the failure. Adding a second monthly inspection changes nothing.

Fix this by asking, for each step: what failure is this step there to catch? Any step with no answer is a habit rather than a control.

Signs you are over-maintaining

• •Checks that have passed forty times in a row without a finding
• •Preventive jobs closed in under two minutes, consistently
• •Parts changed on a schedule that still look new when they come off
• •A schedule that has never had anything removed from it

What to do instead

Move the effort from frequency to quality. One well-designed check on the right failure mode, done properly with a reading recorded, beats four visual walk-pasts. And where a machine fails randomly rather than with wear, stop scheduling intrusive work and put the money into detecting the failure early instead.

Getting permission to do less

This is the hard part, because reducing maintenance sounds like a cut. Frame it as a swap: the hours removed go into the equipment that is actually failing, and you will name that equipment. Nobody argues with moving effort towards the problem.