Almost all of preventive maintenance planning is one decision made repeatedly: how often should this be done. Everything else, the checklists, the scheduling, the software, follows from the intervals you choose, and a plan built on intervals nobody can justify is a plan that gets argued about every time it is inconvenient. Three inputs settle an interval, they carry different weights for different equipment, and being explicit about which one drove each choice is what makes a plan defensible later.
Input one: the manufacturer, adjusted for duty
The manual is the starting point and it assumes a duty cycle that may not be yours. A unit running continuously in a dusty environment is not the unit the manual was written for. Adjusting the interval for actual duty is normal engineering judgement, and it belongs to whoever in your organisation is competent to make it, recorded with the reason.
Input two: your own failure history
The strongest evidence available is what has actually failed at your site, and it is the input most sites cannot use because they never kept the record. This is the practical argument for recording completed visits and findings from day one: two years in, it lets you shorten the intervals that were too long and lengthen the ones that were wasting visits.
Input three: what a failure would cost
Two identical pumps deserve different intervals if one stops production and the other does not. Consequence belongs in the interval decision explicitly, not as an instinct. It is also the honest answer to why the plan is not uniform, which is a question somebody always asks when they see the schedule. Write the reason beside every interval, in one sentence, at the moment you choose it. It costs seconds then and it is the difference between a successor who can adjust the plan and one who inherits a set of numbers nobody can account for and therefore never changes any of them.
Questions people ask about preventive maintenance planning
Is more frequent maintenance always safer?
No. Every intervention carries a chance of introducing a fault, and over-servicing consumes hours that other assets need. The aim is the interval that catches the failure mode, not the shortest one you can staff.
How do we know an interval is wrong?
Two signals: failures happening between visits, which means too long, and visits that consistently find nothing at all over a long period, which may mean too short. Both need the record to be visible.
Who signs off an interval change?
Whoever is competent for that equipment at your site, and the decision belongs written down with its reason. Software should hold the interval, never propose it.