Improving & Right-sizing a Maintenance Organization
You can analyse your actual schedule compliance for weekly and daily work. This helps the maintenance department to reach the right capacity in the future.
– How do I know that our maintenance organization is the right size, is one of the most common questions I am asked, especially from those who are new to maintenance.
Many have tried to find a simple answer to this question. I, myself, have collected large amounts of data to find this simple answer. For example, number of motor drives, pumps or other common equipment within respective industries, in relation to the size of the maintenance workforce.
Some believe that maintenance size should be based on estimated replacement value. It is commonly said that maintenance costs should be less than 2 percent of estimated replacement value and the cost of employees is then a percentage of that cost, often 30 to 60 percent depending on where in the world the facility is located.
This is, of course, very difficult to determine because it is difficult to assess the replacement value of older facilities. Additionally, maintenance costs vary from facility to facility, even within the same company. Many of these costs depend on how rules are applied; in some companies for example lubricants are seen as an operating cost and in others as a maintenance cost.
The same thing goes for the replacement of machine clothing, tools, refiner plates and many others. One of the biggest variations includes the application of rules for what can be defined as a maintenance cost versus capital cost. A few years ago, I worked with two competing companies.
One company had maintenance costs of about US$ 55 per ton, the other had maintenance costs of US$ 120 per ton. The cost of manufacturing was about the same for both companies. The company with the lower maintenance costs later purchased the other company.
Subsequently, the acquired company adjusted the rules of capitalization and their maintenance costs ended up at the same level. Maybe the acquired company capitalised as much as possible because they knew they could be bought?
Another common methodology divides the number of hours in your backlog with available hours and then follows this trend. Unless the maintenance organization has an accurate time and resource estimate of all work in the planned backlog it is difficult to use this method because many companies do not have accurate time estimates in their backlogs.
The Reactive Work Environment
It does not make sense to measure anything if your work environment is reactive!
In a reactive work environment the workload is out of control because it varies too much and the maintenance organization will always require more resources. The very basic maintenance management processes must be in place and executed well before it makes sense to measure anything. These processes are described in Figure 1.
These basics are defined as:
• Prevent includes lubrication, alignment, balancing, operating practices and other activities.
• Inspect includes basic look, listen, feel and smell-type inspections with smart methods, and predictive technologies such as vibration, wear particle analyses and IR to mention but a few.
• Plan answers the questions: What? How? How long to do the work?
• Schedule answers the questions: When? Who does the work?
• Execute is to do the work according to the set schedule.
How do you find out if your maintenance work is too reactive? The easiest way to find out is to analyse your actual schedule compliance for weekly and daily work and, if applicable, shut down work.
Table 1 shows an example of a non-reactive organization and what best performers achieve.
• Daily schedule compliance is based on schedule closing time 20 hours before the start of execution of scheduled work.
• Weekly schedule compliance is based on schedule closing time 3 days before the start of execution of scheduled work.
• Shut down schedule compliance is based on schedule closing time two weeks before the start of execution of scheduled work.
Table 2 shows the typical schedule compliance in many organizations.
Daily schedule compliance is based on schedule closing time 20 hours before start of execution of scheduled work.
If the wasted time associated with unplanned and unscheduled work is 50 percent and the percentage of break-in work is reduced from 60 to 40, a 25 percent improvement in time performance will be achieved. Wasted time includes time to find out what to do, find people, find spare parts and so on.
The use of this method, rather that outdated methods such as “wrench time”, measures the efficiency of the process. In comparison, wrench time studies is a negative approach to people, and only measures the symptoms of the inefficiencies of the process that people work in. It is not true that busy people are productive, unless they are busy doing the right things.
Basic inspections and condition monitoring must be in place to feed the work management process with the early detection of failures.
Assuming that daily schedule compliance in table 2 improves to 90 percent (10 percent break-in work in schedules) as in this table (table 3), and wasted time associated with unplanned and unscheduled work remains the same, then multiplying the break-in work percent with wasted time percent will reduce Total Wasted Time by 25 percent.
Mark Twain famously insisted that there are three types of lies: lies, damned lies, and statistics. Even so, we have often successfully applied statistics in order to estimate the correct sizing of a maintenance organization.
The chart below shows an example from a maintenance organization with 30 hourly employees. In 2010, they cut down the maintenance work force from 30 to 26 people. Since nothing was done to improve the execution of the basic maintenance processes, the obvious result was that extra hours in overtime and contractor hours increased from 2010 to 2014.
There was still a strong product demand in 2012, 2013, and 2014 and at the same time production reliability went down. I see similar examples over and over again. Short-term savings precedes long-term common sense.
Total hours have gone up from 70,900 to 89,196. The key here is to not analyse the number of employed people but rather look at all the maintenance hours and how they are distributed within total paid hours.
The data in Table 4 was used to estimate the total number of hours needed after improvement initiatives were implemented. The estimated attrition rate, mostly due to retirements, was also included.
In this case, a realistic efficiency improvement of 15 percent was calculated through the improvement and execution of the basic maintenance processes. This leads to 61,200 hours needed in the future. The implementation and results are expected to increase gradually within a period of four years. The attrition rate during these years is 5 own skilled workers.
Overtime and purchased services will return to a normal level of 10 percent each. This means that the right amount of hours for the maintenance department is calculated to 55,200 hours or 27 employees. This will lead to four new people being needed, recruited and trained in four years.
By implementing and performing better and better maintenance processes, the maintenance department will reach the right capacity in the future.
Contact us if you would like me to visit your organization.
Direct Telephone: +1 919 723 2683.
A growing number of companies want to use big data analytics in their predictive maintenance and are also investing in the resources needed for this. Of the companies already using this technology, no less than 95 percent say that they have already achieved concrete results. This is the conclusion of a follow-up study conducted by PwC and Mainnovation in recent months, among 268 companies in the Netherlands, Germany and Belgium.