For maintenance and reliability leadership
Reliability evidence is made in the last ten minutes of a job.
You are accountable for uptime, for the safety of the people doing the work, and for the cost of everything that arrived unplanned. All three are evidenced at close-out, on a shift already behind.
Written for: Maintenance Director · Head of Maintenance · Maintenance & Reliability Manager · Engineering Director
What you are measured on
- Unplanned work
- The share of the week that arrives rather than gets planned, priced in overtime, expedited parts and lost output.
- Schedule compliance
- Whether planned work happened in the window it was planned for, and what got dropped when it did not.
- Safety and statutory completion
- Inspections, permits and safety-critical maintenance closed on time, with evidence that holds six months later.
- A budget defended annually
- Finance sees spend. You show what the spend bought, from data produced by people whose priority that day was the asset.
One Tuesday, five minutes of it
A worked shift, read at the asset and read in the record
The pattern maintenance leaders describe once an estate has grown faster than the process around it. The work is done well in every one of these minutes.
At the asset
In IBM Maximo
06:05
The supervisor reorders the day at the whiteboard: a pump on the critical list tripped overnight, so two planned jobs move.
Yesterday’s schedule stands. The reason two jobs moved is in a photograph of a whiteboard and a messaging group.
07:40
A technician takes the corrective job, collects the isolation, and finds the seal has failed the way it failed in March.
The work order goes to INPRG. The March job is in history, closed with a free-text comment and no failure code.
11:20
Failure, cause and remedy are captured at the asset while the seal is still in hand, in the order a technician thinks in.
Coded failure, cause and remedy land against the child asset, with parts issued and the actual duration.
13:15
The second planned job is dropped: a four-hour round against three hours of available technician.
The deferral is recorded against criticality, so the compliance figure carries the reason it moved.
16:00
Handover names one asset for the reliability engineer rather than five.
Two failures on that asset share a cause code, and so do three interventions across the asset class.
Two words, one backlog line
Planned work and scheduled work are different problems
Which one you are short of decides whether you need planners, a scheduling discipline agreed with operations, or both.
Planned
The job is understood: a method, a job plan, the parts, the access requirement, the trade, and a duration somebody believes.
Scheduled
The job has a date, a named crew, and a window operations has agreed to release the asset for.
What it needs
Job plans corrected from real completions, a bill of materials, and durations measured from recorded work.
An agreed window, an available crew, the parts on site, and the permit or isolation cleared in advance.
What happens when only this one exists
A well specified job that never gets a window and ages into a backlog line nobody rereads.
A crew at an asset without the right part or permit, which becomes a returned job and a compliance miss.
Which measure it moves
Wrench time, rework, and the accuracy of your duration estimates.
Schedule compliance and the age profile of the backlog.
Who owns it
The planner, working several weeks ahead of the shift.
The scheduler and operations, working this week and next.
Close-out
The last ten minutes, in five beats
The 07:40 corrective job, closed at the asset by the technician who did it.
-
Beat 1
Failure is selected from a short list scoped to the asset class, in place of free text.
Written A failure code against the child asset, so history rolls up to the class.
-
Beat 2
Cause and remedy are chosen in the order a technician thinks in: what was found, then what was done.
Written Cause and remedy codes, with a photograph attached to the work order.
-
Beat 3
Parts used are confirmed against the reservation, including the two taken from van stock.
Written Issued items and actual quantities, so the job cost is the job cost.
-
Beat 4
Labour and travel hours are confirmed once, on the device, before the technician leaves site.
Written Actual duration against the job plan task, which corrects the next estimate.
-
Beat 5
The supervisor accepts the close-out at the end of the shift rather than retyping it next morning.
Written Completion status with an audit trail naming who closed it and when.
None of the five beats adds a step to the job. On rounds and calibration work, where both hands are already busy, MaxIron Nova captures the same fields hands-free.
Scope
Four constraints that stay exactly as hard
Stated now rather than in month four.
The evidence base improves from the day capture improves
A year of work orders closed with free text carries no reliable failure data. MaxIron AI Smart Data proposes structure on the long tail for a person to confirm, and coded history accumulates from the day close-out changes.
Staffing arithmetic stays what it is
Where a round takes four hours and three hours of technician are available, the trade-off becomes visible and criticality-led, so the work dropped is work you chose to drop. Where the operating model is the constraint rather than the tooling, that is business process optimisation work.
Releasing an asset stays an operations decision
Access, permits and production priority sit inside your operating model. We make the cost of deferral visible and evidenced, and the negotiation with operations stays yours.
Parts and shutdown windows keep their own clock
Lead times and a stores function sized for a smaller estate are unchanged, and a shutdown window lasts as long as it lasts. Better evidence changes what you put in the window.
Run before
Two estates that started at capture rather than at reporting
-
UK gas distribution network
Gas distribution
Situation
Reliability reporting assembled by people each period rather than produced from the record.
What changed
An embedded data lead worked on capture quality instead of a reporting layer bolted over the top.
Read the case study → -
UK oil and gas major
Oil and gas
Situation
Failure history that had to survive audit and roll up into the reliability templates.
What changed
Failure coding to a recognised standard, so day-to-day capture feeds the templates directly.
Read the case study →
The same sequence runs across utilities and oil and gas, where regulated returns and safety-critical maintenance set the evidence standard. MaxIron Sentinel makes the path from work request to close-out visible while a period is still open.
Where to start
Four routes, taken by what is costing you the week
Each is a different first phase, and only one of them starts with a screen.
The work is being done around the system
Maximo Mobile rollouts →You are on it if
Technicians carry paper or a phone camera because the app does not fit lone working, permits, materials or a substation with no signal.
The wrong move here
Retraining the crew on the existing flow. The flow is what has to change.
The reliability return is assembled by people
MaxIron AI Smart Data →You are on it if
The compliance figure exists in three versions, and the period return is rebuilt by hand every time.
The wrong move here
Buying a reporting layer over completions that carry no failure code.
The calendar is running the asset
IBM Maximo Health →You are on it if
The interval on a critical asset was set years ago by somebody who has left, then copied across when the asset was replaced.
The wrong move here
Adding inspections to an asset whose failure history nobody has read.
The problem is broader than one process
Maximo Health Check and Heal →You are on it if
Three functions each describe a different root cause, and no two of them agree on the numbers.
The wrong move here
Fixing the process that complains loudest before the estate has been read.
Maintenance leadership, frequently asked
- Is this a mobile project or a data project?
- Both, in that order. Reliability data is created at close-out, so how the work is captured is where a data programme starts. Fixing reporting first produces a better view of the same unreliable data, which is the pattern set out in asset data governance that actually works.
- We already have Maximo Mobile and it did not land. What now?
- That conversation is nearly always about the shift rather than the software: the start and end of shift, lone working, permits, materials, and what happens where there is no signal. We restart from one real shift and rebuild the flow around it. The rollout guide sets out what we look for.
- How long before the reliability reporting is worth reading?
- Capture quality changes within weeks of the close-out flow changing, because it is a behaviour change with a system behind it. Trend analysis needs enough coded completions to be meaningful, so the first defensible failure analysis lands a few months later, on your highest-volume asset classes first.
- Do we need MAS Monitor or Predict for this?
- Not to start. Monitor and Predict earn their place once Manage is in good shape and the asset class has failure history worth learning from, and condition insight turns those signals into an intervention a planner can act on. Prediction layered over inconsistent completion data gives confident answers on a weak base.
- What do you need from my team?
- A supervisor and a planner for the first two weeks, one technician who is candid about what the current flow costs them, and twelve months of work order history. No requirements document, and we would rather you did not write one before we have watched a shift.
Bring your ten worst assets.
Tell us which assets cost you most, show us one round and one close-out at shift start, and we will say which problems are capture, which are planning, and which are neither.
Bring this to the first conversation
- Twelve months of work order history on the ten assets that cost you most
- Last period compliance figure, and how it was produced
- One round and one corrective job we can watch at shift start
- The report you least enjoy assembling each month