Preventive maintenance schedules that survive staff turnover

Every building has a preventive maintenance schedule. Most of them live in one of three places: a binder from the year the building opened, a spreadsheet maintained by an engineer who has been there eleven years, or that engineer's head. All three work perfectly until the day the engineer takes another job.

That is the honest case for preventive maintenance software for commercial buildings. Not efficiency. Continuity. The schedule has to survive the person who knows it.

What gets lost at turnover

Not the obvious items. Nobody forgets the elevator inspection or the fire panel test, because those come with a regulator and a date.

What gets lost is the local knowledge wrapped around each task. That the north cooling tower needs its strainer checked more often than the south one because of the tree. That the emergency generator has to be exercised before eight, because a tenant on two complains about the noise. That the roof drains fill with seed pods for three weeks every spring.

None of that is in the manufacturer's manual. All of it is why the building runs. A schedule that captures the task but not the note is half a handover.

Frequency is a judgement call

Manufacturer intervals are a starting point written by someone who has never seen your building. Run time, load, water quality, filter loading and how hard the equipment is pushed in August all shift the real interval.

The teams that get this right treat frequency as something they revise. Twice a year, they look at what was found on each visit. If a quarterly task finds nothing four times running, it becomes semi-annual. If a semi-annual task keeps finding something, it moves the other way. That review is fifteen minutes per system, and it is the difference between a schedule that reflects the building and a schedule that reflects a PDF.

Procedures beat task names

"Quarterly AHU service" means one thing to the engineer who wrote it and something else to the contractor covering a shift. A procedure with steps, torque values, part numbers and the two things that usually go wrong means the same thing to both.

Procedures are also what makes vendor work auditable. When a contractor marks a task complete, you want to know which steps were performed, not that a box was ticked at 14:03.

What good preventive maintenance looks like

• Tasks belong to the asset. Equipment carries its own history, so when the chiller is replaced its schedule and its record go with it instead of being rebuilt.
• Schedules generate work automatically. The task appears in the queue on its own. It does not depend on somebody remembering to raise it.
• Every task carries its procedure. Steps, parts, and the local note about the tree, the noise complaint or the seed pods.
• Findings become work. An inspection that finds a worn belt should produce a repair with one action instead of an email that somebody re-types on Monday.
• Completion produces a record. Who, when, what was found. Insurers, buyers and regulators all eventually ask, and the answer should be an export.

The deferral problem

Preventive work loses to reactive work every single time, because a leak is in front of you and a filter change is not. In a busy quarter the planned queue keeps growing and nobody makes a decision about it.

Deferral is legitimate. Silent deferral is not. The practical fix is to make skipping a task an action that leaves a trace: a reason, a new date, and visibility for whoever is accountable. A schedule with three documented deferrals is a managed building. A schedule with forty untouched overdue tasks is a building that has stopped using its schedule, and the second one is far more common.

Inspections are the cheap half

Routes matter as much as tasks. A porter walking a set floor route with a phone, logging what they see, generates more useful early signal than any sensor you can buy this year, and it costs nothing you are not already spending.

The requirement is that findings turn into work orders without leaving the phone. A clipboard route that ends in a photo nobody looks at is worse than no route, because it creates the impression of coverage.

In-house, vendor, or both

Who performs the work changes what the system has to do, and it is worth deciding deliberately instead of inheriting it.

In-house engineers are fast, know the building and cost the same whether the queue is full or empty. Their weakness is the record. Work done by someone who has been there for years tends to get logged thinly, because the person doing it already knows what happened.

Vendors bring specialist skill and produce paperwork by default, since paperwork is how they get paid. Their weakness is context. A contractor on a rotation has never met the tree, the noise complaint or the seed pods, so whatever is not written down does not happen.

Most buildings run both, which means the system has to serve two audiences at once: quick completion for the person who knows the building, and structured procedure for the person who does not. A platform that only does one of those will be worked around by half your team within a month.

The cleanest split is usually by asset criticality. Life safety, lifts and anything with a compliance date go to the specialist with the certificate and the paper trail. Filters, belts, drains and rounds stay in-house where speed matters more than documentation, with a procedure attached anyway so the record survives.

What this is worth

Avoid the temptation to promise a specific percentage. The defensible case is narrower and stronger: equipment that is serviced on a schedule fails less often at inconvenient times, warranty claims survive scrutiny when there is a service record, and a building that can produce a complete maintenance history is worth more in diligence than one that cannot.

The other return is the one nobody puts in the model. When the engineer with eleven years of local knowledge leaves, the building keeps running the same way in month two.

One caution about starting. The instinct is to load every asset in the building before going live, which turns a four-week project into a nine-month one and usually stalls. Start with the equipment whose failure would cost the most: chillers, boilers, lifts, the fire system, anything with a compliance date. Get those on schedule with real procedures. Add the long tail as it comes up, because the second year of data is worth more than a complete first-year inventory that nobody trusts.

Where BuildingHub fits

Preventive maintenance in Operations Hub hangs schedules off the asset, generates the work when it is due, carries a procedure with each task and turns inspection findings into work orders without re-keying. Reactive requests run in the same system but their own lane, so a leak never competes with a filter change, and the completion record is exportable when an insurer or a buyer asks. Related reading: the anatomy of a work order that closes on time.