Maintenance Personnel Numbers

Let us assume that you have worked out that you need a minimum staffing of 20 operators per shift. There is a correlation between operator numbers and the size of the maintenance workforce for each specific industry. Hence it is possible to derive maintenance numbers reasonably accurately. As an example, for the petroleum refining industry, 20 operators per shift would correlate to a total maintenance workforce size of 150-200, depending on the efficiency of the operation. Table 9.2 shows how this total might typically be broken down for specific activities in a refinery.

Staffing of Instrument Group

There is a reasonably good correlation between the number of instrument control loops (or valves) and workload, hence, staffing levels. This can be used as a check on the numbers derived above.

Staffing of Electrical Group

In a similar way, the electrical workload correlates with the number of electric motors. This ratio again can provide a check.


  • 1. You can always manage with fewer people than you think, but few people will agree with you and they will have many excuses.
  • 2. No plant start-up managers ever get fired for having too many people so they will have too many. Make sure this surplus does not become a problem for the permanent organization.
  • 3. Many consultants produce high-level algorithms with which to derive staffing levels. No one at the work level will believe these so you need to derive and explain your numbers in a more practical way.
  • 4. The resource numbers situation should be kept under continuous review. This should cover not only labor, but also supervisory and managerial grades. It is much easier to make gradual reductions rather than step changes.
  • 5. Adding more incompetent people to cover competence shortages will not make your life easier. Nor will it bring business benefits.


  • 1. Have only five layers from facility manager to hands-on tool workforce.
  • 2. Work out what you will need for an efficient operation and set your staffing levels at 90% of this.
  • 3. Keep reducing numbers at all levels until there is clear evidence that you have too few.
  • 4. Make your algorithm saleable to the staff at site.
< Prev   CONTENTS   Source   Next >