Maintenance planning: turning a work request into a job the crew can execute

Planning and scheduling are two different jobs

The planner answers what and how; the scheduler answers when and by whom. Fusing the roles into one person almost always means urgent work swallows prepared work — the day goes on chasing callouts and nothing ever gets properly readied. A planner working ahead of the crew should be looking at jobs that run weeks out, never at jobs running this shift. There is a blunt test for this: if the planner is on the radio all morning, they are dispatching, not planning. Separating the two functions is often the single change that lifts hands-on tool time without adding a technician to the payroll.

A job package is the planner's actual product

What gets handed over is a package, not a line of text. It carries the task sequence, an estimate of trade hours by discipline, a parts list showing bin locations and reservation status, the drawing or procedure extract that matters, isolation and permit requirements, and any lifting or access equipment. Estimates carry more weight than people expect: without a credible hour figure nobody can build a weekly schedule that fills crew capacity. Those estimates only improve if completed jobs feed back — actual time taken, parts substituted, steps that turned out to be wrong — so the technician has to close the loop in writing.

Backlog is an inventory that needs grooming

Backlog is not a shame list; it is the pool next week's schedule is drawn from, and it has to be sized and sorted. Ready backlog — planned, parts on site, nothing outstanding — should cover several weeks of crew capacity, or the schedule fills with whatever happens to be at hand. Work waiting on a part, a shutdown window or an engineering decision belongs in separate buckets so nobody schedules it by accident. Pruning matters too. Requests overtaken by a modification, duplicated by a second reporter, or made moot by a rebuild should be closed with a reason rather than carried indefinitely.

Schedule compliance tells you whether the front end works

The share of scheduled jobs finished in the week they were scheduled is the honest read on planning quality. Track it beside how much of the week's available hours were committed in advance. Weak compliance with heavy commitment points at poor estimates or constant interruption; light commitment points at a thin ready backlog. Both readings need production's own account of what stopped the work sitting next to them. None of it means anything unless the weekly schedule was agreed with the production manager before issue — a plan the plant never intended to honour measures nothing but wishful thinking.

Break-in work and the rule that governs it

Every plant generates work that will not wait, and the perennial fight is over how much of the schedule that work may eat. Agree the rule before the argument: who may authorise pulling a technician off a prepared job, what evidence justifies it, and where the displaced job lands. Leaving a declared allowance for reactive work in the weekly plan is more honest than pretending the crew is fully committed and then watching the plan collapse by Wednesday. Log every break-in against a reason code. The pattern almost always points at a small set of assets that deserve an engineering fix rather than repeated rescue.

Frequently asked questions

Does the planner have to walk the job before planning it?

For anything unfamiliar, yes. Photographs and drawings hide the things that wreck an estimate: a valve that cannot be reached without scaffold, a motor that will not come out without lifting a guard, a cable route nobody documented. A scoping visit takes minutes and prevents a crew arriving with the wrong kit. For repeat jobs a good written package removes the need, which is why capturing the first execution properly pays back on every repeat afterwards.

What do we do when production will not release equipment for planned work?

Escalate it as a scheduling conflict, not a maintenance complaint. Bring the specific asset, the consequence of deferral, and an alternative window — a changeover, a shift break, a planned product trial. If refusals cluster on one line, the real issue is that the plant has no agreed maintenance window and is running on borrowed reliability. That decision belongs to the site manager, in writing, with the deferred jobs listed so the risk being accepted is visible rather than assumed.

How do we start planning when we have no job history to draw on?

Begin with the assets that hurt most and build packages as work arises, capturing each execution as you go. Trade supervisors can estimate hours well enough to start; accuracy comes from correction, not from a first guess. Equipment manuals give task steps and part numbers for the initial set. Within a few months of disciplined capture the plant has its own library, which is worth considerably more than a generic task list bought in from outside.