How to Build an Equipment Maintenance Plan (Step by Step)

The first time I asked a maintenance manager to show me his equipment maintenance plan, he opened a binder. The binder was three years old. Half the assets in it had been replaced, two of them had been sold, and the one machine that broke down every quarter wasn’t in there at all. Everyone on his team still kept the place running, but they did it from memory and habit, not from the plan.

That’s the thing about equipment maintenance plans. Most plants have one. Very few have one that’s actually being used. It ends up as a document you wrote for an auditor rather than a system your team runs on. The cost of that gap is real: Aberdeen research estimates unplanned downtime costs U.S. manufacturers $50 billion a year, and the top cause - aging equipment and deferred maintenance - accounts for roughly a third of those failures.

I’m Yuval, CEO of Glitter AI, and I spend a lot of time with maintenance and operations teams who want their plan to survive a vacation, a turnover, and a bad week. So here’s how to build an equipment maintenance plan from scratch, in the order that actually works: asset list, strategy, schedule, resources.

Turn your maintenance plan into guides any technician can follow

Teach your co-workers or customers how to get stuff done – in seconds.

Start for Free

Replay video

0:00 / 0:00

What an Equipment Maintenance Plan Actually Is

An equipment maintenance plan is the document and system that defines, for every asset you own, what gets maintained, which strategy each asset gets, when the work happens, and who and what it takes to do it. It’s the level above a single preventive maintenance program: the PM program handles the time-based and condition-based work, while the maintenance plan covers the full picture, including the assets you’ve deliberately decided to run to failure.

Here’s the distinction I’d put on the wall. A maintenance plan answers four questions, and it isn’t finished until all four are written down:

1. What do we maintain? (the asset list)
2. How do we maintain each one? (the strategy)
3. When does each task happen? (the schedule)
4. With what do we get it done? (the resources)

Skip any one of those and you don’t have a plan. You have a wish.

Step 1: Build the Asset List

You can’t plan maintenance for equipment you haven’t named. The asset list is the foundation, and it’s the step most teams rush.

Walk the floor. Don’t pull last year’s spreadsheet and assume it’s right. Physically go to every line, room, and utility closet, and write down every asset that can fail in a way that costs you money, safety, or compliance.

For each asset, capture the minimum useful record:

• Asset ID and name (a unique tag, not “the big compressor”)
• Location (line, area, room)
• Make, model, serial number
• Function (what it does and what stops if it goes down)
• Install date or age
• Criticality (we’ll rank this next)

Then rank criticality. You can’t maintain everything with the same intensity, and you shouldn’t try. A simple 1-3 scale works: a critical asset (1) stops production, creates a safety hazard, or triggers a compliance issue if it fails; a moderate asset (2) slows things down or has a backup; a low asset (3) is annoying but not urgent. This ranking is what makes the rest of the plan affordable, because it tells you where to spend your hours.

If you want a clean structure to record all of this consistently, our equipment maintenance log guide covers exactly what fields to track and why.

Turn your maintenance plan into guides any technician can follow

Teach your co-workers or customers how to get stuff done – in seconds.

Start for Free

Replay video

0:00 / 0:00

Step 2: Choose a Maintenance Strategy per Asset

This is where a real plan separates itself from a calendar full of reminders. Not every asset deserves the same strategy, and assigning one is a deliberate decision, not a default.

There are four strategies, and your plan should explicitly tag each asset with one:

• Reactive (run to failure). You intentionally let it run until it breaks, then fix or replace it. This is the right call for cheap, non-critical, easily replaced assets where prevention costs more than the failure. The keyword is intentionally. Run-to-failure is a strategy when you chose it and a problem when it chose you.
• Preventive. Time-based or usage-based tasks performed on a fixed interval, regardless of condition. This is the workhorse for most critical and moderate assets. The structure for this lives in your preventive maintenance program. The recurring task lists come from a preventive maintenance checklist.
• Predictive. You monitor condition (vibration, temperature, oil analysis, runtime) and act only when the data says failure is coming. Reserve this for high-criticality assets where unplanned downtime is expensive enough to justify the sensors and analysis.
• Reliability-centered. For your most critical assets, you analyze likely failure modes and design the maintenance specifically to prevent the failures that actually matter. This pairs well with root cause analysis so that every repeat failure feeds back into the plan.

Here’s the simple decision rule I give teams. Start with criticality from Step 1. Criticality 1 assets get preventive at minimum, and predictive or reliability-centered if the downtime cost justifies it. Criticality 2 assets get preventive. Criticality 3 assets are candidates for deliberate run-to-failure. Write the chosen strategy next to every asset on the list. That column is the brain of the whole plan.

One caution on preventive: ARC Advisory Group research found that roughly 50% of preventive maintenance tasks add no value as written, and about 30% are done more frequently than necessary. That’s not an argument against PM - it’s an argument for using real failure history to set intervals rather than treating manufacturer defaults as gospel.

If you’re standardizing how this care happens day to day across the whole team, this is the operational expression of total productive maintenance.

Step 3: Build the Schedule

Now you turn strategy into dated, assigned work. A strategy that never lands on a calendar is just an opinion.

For every asset with a preventive or predictive strategy, define the actual tasks and their triggers, ideally as a reusable preventive maintenance checklist:

• The task (what specifically gets done, e.g., “replace belt, check tension, lubricate bearings”)
• The trigger (calendar-based like “every 90 days,” usage-based like “every 500 run hours,” or condition-based like “when vibration exceeds threshold”)
• The owner (a role, ideally a named person)
• Estimated time (so you can see if the plan is even physically possible)

Then stagger it. The single most common mistake I see is a schedule where everything is due the first Monday of the month, which guarantees nothing gets done well. Spread tasks across the calendar so weekly load is roughly even and your team isn’t choosing between fire and fire.

Be honest with the math here. Add up the estimated hours of every scheduled task in a typical month, then compare that to the labor hours you actually have. If the plan needs 400 hours and you have 250, the plan is fiction. Your options: add resources (Step 4), downgrade some assets to a lighter strategy, or accept more run-to-failure on low-criticality equipment, on purpose, in writing.

Each scheduled task also needs a procedure a technician can follow without you standing next to them. A line that says “service pump 4” isn’t a task; it’s a guess. The task needs the steps. This is exactly the gap I built Glitter AI to close, because the difference between a plan that works and a plan that doesn’t is usually whether the person doing the work knows how, not whether it’s on the calendar.

Step 4: Line Up the Resources

A plan is only as real as the resources behind it. This is the step that gets cut first and breaks the plan quietly.

Walk through each category and write down what the plan really requires:

• People. Headcount, skills, and certifications. Map which technicians can perform which tasks. The single point of failure here is the one senior tech who knows the tricky equipment, because when he’s out, that part of the plan stops.
• Spare parts. The critical spares that need to be on the shelf so a planned task doesn’t turn into a three-week wait for a part. Tie this to your criticality ranking, not to whatever you ran out of last.
• Tools and access. Specialized tools, lockout/tagout equipment, and the production windows when an asset can actually be taken down for service.
• Documentation. The step-by-step procedures for every recurring task. This is the resource that decides whether your plan survives turnover.
• A system to run it on. A whiteboard works for ten assets. Past that, you need a CMMS to generate work orders, track completion, and tell you whether the plan is real or theater. When you’re ready, our CMMS implementation guide walks through doing that without the project stalling.

The documentation resource is the one I’d push you hardest on. A maintenance plan that depends on tribal knowledge isn’t a plan, it’s a person, and people leave. Every recurring task in your schedule should have a written, visual procedure that a new technician could follow on day one. That’s the difference between a plan that survives a retirement and one that retires with the retiree.

Turn your maintenance plan into guides any technician can follow

Teach your co-workers or customers how to get stuff done – in seconds.

Start for Free

Replay video

0:00 / 0:00

How to Tell If the Plan Is Actually Working

Once it’s running, judge the plan by a few numbers, not by whether the binder looks nice:

• PM compliance - what percentage of scheduled tasks actually got done on time. Below ~85% and your plan is mostly fiction.
• Planned vs. reactive ratio - how much of your labor is scheduled work versus firefighting. A maturing plan moves this toward planned over time. Industry maintenance benchmarks show world-class operations achieve 85-95% planned work; most facilities sit closer to 55-65%, meaning a large share of labor hours still go to reactive firefighting.
• Downtime on critical assets - the number the plan exists to protect.
• Repeat failures - the same asset failing the same way means the plan’s strategy for that asset is wrong, and it should feed straight back into Step 2.

Review the whole plan on a fixed cadence, quarterly at least. The asset list changes, criticality changes, and the strategy column should change with it. A maintenance plan is a living document or it’s a dead one.

A Note on Keeping the Plan Followed, Not Just Written

I’ll be honest about the part that’s hardest. Writing the plan is the easy 20%. Getting it consistently followed when the floor is on fire is the other 80%, and it almost always comes down to whether the procedures are clear enough to follow under pressure - a tight preventive maintenance procedure for each critical asset is what makes that possible.

That’s the specific problem Glitter AI solves. You record a task once, the way you’d actually do it, and it becomes a clean step-by-step guide with screenshots that any technician can follow without asking you. Your maintenance plan stops being a binder and becomes something the whole team can actually run, right down to the maintenance work order process each task flows through and the machine maintenance checklist and maintenance checklist template behind it. Treat each procedure like a proper standard operating procedure, and the plan stops depending on whoever happens to remember.

That’s the whole game: a plan isn’t real until the person doing the work can do it without you. Build the asset list, assign the strategy, schedule the work, line up the resources, and write down the how. Do that, and breakdowns get boring. Boring is the goal.