You will need to use a mix of lagging indicators and leading indicators to provide information on the performance of your maintenance efforts.
I work as electronics & instrumentation supervisor in the production engineering department in a pharmaceutical manufacturing company. Our department is responsible for maintenance of the mechanical & electrical/electronics of production machines, plus electronic/control maintenance of other sections like water treatment plant, HVAC and boilers. We had a request from our management to set-up maintenance KPI’s for all maintenance activities. Do you have any advice where and how to start?
There are many purposes in using performance indicators (PIs). One is to show the latest level of achievement in meeting an objective. A second is to observe the rate of progress being made toward the objective. Another is to highlight problems and issues. They can be used to show us if the changes we make or actions we take are actually helping to resolve an issue.
This implies that the objective we want to achieve, for which we use KPIs to measure achievement, has a target that defines an acceptable level of performance. Results on the good side of the target, or rapid progress toward the goal, make people happy and content, whereas results on the bad side make them sad and worried. Key performance indicators (KPIs) are those performance indicators that measure the critical success factors which affect a desired outcome.
Maintenance performance indicators reflect achievement and progress in meeting an agreed maintenance benchmark. In measuring maintenance performance we are concerned not only with doing good maintenance work, we are also concerned that the maintenance work we do successfully removes risk of failure from our plant and equipment. It requires us to use a mix of lagging indicators and leading indicators to give us a clear understanding of what is happening to the risk and performance of our operational assets through of our maintenance efforts.
Lagging indicators use historic data to build a performance trend line. The trend shows our progress and can be projected forward a little to forecast likely progress (but not too far ahead, since the future is unknowable, maybe a few weeks, provided all things remain the same). Leading indicators use historic data to monitor if we are doing those activities that are known to produce good results. The reasoning is that if we do the right things rightly then the right outcomes will result.
An example of a lagging indicator related to machine health is Mean Time Between Failures (I define failure as an unplanned stoppage for any reason). MTBF is the number of historic failures for an item of plant in a time period, usually a year. We plot each time period and develop a graphic trend. An example of a leading indicator in maintenance is the percentage of condition inspection work orders performed when they fall due. If we are not observing the health of our machines for warning signs of ill-health we will have many unwanted failures. Hence a low percentage of inspection WOs completed is a warning that we will suffer a rising rate of breakdowns and unwanted stoppages.
You want to have as few performance indicators as possible (because they take resources and effort to gather, analyse, store, report and read), yet sufficient to provide good awareness of what is happening.
Good maintenance KPI selection gives us an overview of the situation in our plant—the past is clear and so is the future. The range of KPIs and PIs to use to monitor and focus maintenance on what is important to do include checking:
- the delivery and quality of maintenance work,
- the reduction in risk to the operating equipment, and
- the effectiveness of the money and resources expended to achieve those results.
You categorise your maintenance effort and then look at a series of PIs that reflect the category and trend them to give an overall impression of the situation within a category. Possible categories and their measures include:
- Proportion of WOs Performed when First Scheduled (this monitors how well scheduling is being done in the operation)
- Proportion of Planned WOs in All WOs (tells me if the Planning effort expended is sufficient to meet the goal of all work orders properly and thoroughly planned. Planned work is the least expensive way to do maintenance work and it is a controlled and intentional way to ensure reliability is delivered to the plant and equipment by doing planned quality work)
- Backlog Size by Work Order Priority and by Equipment Criticality (this shows if we have an ‘urgent work gets the attention’ culture)
- Backlog Size by Work Order Type i.e. Preventive, Predictive, Corrective, Proactive (this helps to see if we are doing the important things first)
Maintenance Work Quality:
- Number of Rework WOs (I define a rework job as any early life failure unless it was due to induced failure, like brinelling, manufacturing errors or commissioning errors)
- Training courses and refresher courses completed per Maintainer (only the people with the answers can solve the problems)
- Proportion of WOs with error proof procedures (or at least having Inspection and Test Plans stating pass/reject quality criteria)
- Asset MTBF (the time between unplanned outages of an asset needs to be increasing) This identifies ‘bad actor’ assets with performance that needs to be addressed. Strictly, MTBF only applies to random failures and not early-life failures or end-of-life failures.
- Production MTBF (the time between unplanned Production stoppages needs to be increasing) To prevent production disruption during a forced outage many plants are designed with standby equipment, or large buffer stock between production steps. In such a situation when an asset is failed we trick ourselves into believing it does not need to be reliable because Production is not affected. What we easily miss seeing is the amount of maintenance resources and money needlessly lost fixing repetitive failures. In this case we want to record all production unplanned failures, regardless if they are early-life, random or end-of-life events.
Operational Risk Reduction:
- Number of Equipment Improvement WOs completed (unless we improve equipment reliability we cannot get a more reliable operation)
- Number of Creative Disassembly root cause investigations undertaken by Trade Type (maintainers need to be problem solvers and through their work make the plant and equipment more reliable) In this case you would need to set up a new Creative Disassembly report that Maintainers completed, with the report back contents uploaded into an electronic database.
- Percent Emergency Work Orders (if this value needs to be decreasing rapidly, if instead it is rising or stable then our maintenance strategies are not working or they are the wrong ones to be using)
Maintenance Resource Usage:
- Work Orders Completed by Trade Type (checks to see how efficiently the workforce is being managed) This should be increasing over time as we get more expert, efficient and innovative.
- Proportion of WOs Started at the Time Scheduled to Start (is your Maintainer’s tool time being maximised?)
- Proportion of WOs completed to within 10% of Planned Job Duration (is the work being conducted efficiently?)
- Maintenance Cost Component of Unit Cost of Production (i.e. how much maintenance expenditure is in each unit of product made) Should be decreasing overtime as we get more expert, efficient and innovative.
- Ratio of Actual Maintenance Cost vs. Budgeted Maintenance Cost (you could the break that KPI into individual PIs that look at the Trade Labour, Trade Parts and Materials, Stores Management, Department Overheads, etc.)
- Maintenance Cost per Maintenance Strategy (separate cost of Preventive, Predictive, Breakdown, Proactive Maintenance Strategies)
- Maintenance Cost Spent per Asset (trend the maintenance expenditure for each asset to see what parts of the operation use up Maintenance resources)
- Proportion of Contracted Labour and Services in Maintenance Expenditure (are we using too much specialist expertise that we should instead develop in house)
What still needs to be done for each KPI is to set the benchmark value that designates passable performance. We need the goal to strive for and that goal will be a measurable value against which we compare our actual results. You can research the value yourself and set a figure that is sensible for the circumstances. If you are not yet a world class operation then setting world class benchmarks will not be an appropriate choice for your company at the moment. As you move toward world class results the KPIs change and/or the benchmarks become more demanding as each value is achieved.
The maintenance measures that you have been asked to develop by your management implies using a mix of KPIs to manage your processes and PIs to monitor performance. In the list above I have mixed them together. You should sit down with the managers who want the information and find out what they specifically want to monitor. Is it costs they are concerned about? Is the time take to get things done? Is it the amount of contracted work sent off-site, etc? You can then select KPIs to measure and PIs to monitor the company managers’ concerns.
Once you select the KPIs to use you also need to have a way to collect the necessary information for the KPI. You may need to introduce new cost codes, new data collection processes, and do training to teach people the new ways needed to gather data.
A most important factor in choosing KPIs is that they are meaningful and they are used by the people that need them. Collecting data, analysing inference and writing reports that no one uses is senseless. Meaningful KPIs are measures that have value to people—Users look forward to getting the KPIs as evidence of performance. Meaningful KPIs help to change practices and behaviour for the better.
DuPont Chemicals uses the Maintenance KPIs listed below which they standardised on some years ago. They are a process chemical manufacturer and the measures relate to their business. But the management focus they take, the categories they monitor and the maintenance performance measures they use are informative in how KPIs are applied in measuring maintenance effort. You may need to alter or add KPIs to reflect your production need for high quality pharmaceutical products.
I hope that you can get some useful ideas from the above. If you have further questions please feel free to ask me.
My best regards to you,
Lifetime Reliability Solutions HQ
PS. If you require advice on industrial asset management, industrial equipment maintenance strategy, defect elimination and failure prevention or plant and equipment maintenance and reliability, please feel free to contact me by email at