Take a load off maintenance

Most operating companies don’t adequately measure maintenance staff efficiency — and thus are not in a position to optimize productivity and performance. Our experience indicates that many firms could use a more rigorous approach for measuring utilization and identifying where losses occur. Such a method, called a Labor Activity Analysis (LAA), can assist in understanding causes and pinpointing prospective improvements. So, we will look at how to conduct such an analysis and compile data for use in improvement planning.

Background on the purpose of an LAA, which is a proven industrial engineering technique, was provided in a previous article (CP, October 2005, p. 25) about the calculation of labor productivity. That calculation covers all the direct and indirect activities involved in doing work, including the efficiency and quality of the actual work, as well as time required for travel to the site, retrieving material and preparing the work area. It provides a measure of overall productivity that accounts for all potential losses limiting the amount of work performed. These activities fall into three categories:

1. Direct utilization — the percent of time craft personnel are performing direct work;
2. Direct productivity — the actual units of work completed as a percentage of a standard; and
3. Rework — the percentage of work performed that is flawed and must be redone.

Together these three elements define craft productivity as:

Productivity = (Direct Utilization)/(Target Value) × (Direct Productivity) x (100% - Rework) 

Of these three elements, direct utilization is the factor most people refer to when discussing efficiency. Direct utilization is an easily understood concept and most people can estimate “wrench time” or “tool in hand” factors, because this only requires a visual assessment of work activity during a snapshot of time.

In contrast, direct productivity and rework are more difficult to determine. Direct productivity requires an assessment of actual direct work against a standard time while rework demands a detailed knowledge of tasks performed a second time and why they were redone. Both of these elements call for specific work order analysis over a relatively long period of time.

So, not surprisingly, due to the ease of deriving an approximation of direct utilization, simplistic estimates usually drive the determination of overall productivity as the other factors are often assumed to be at or near 100%.

A better way

An LAA provides a more accurate view of utilization. It involves the random sampling of work activities and their measurement and classification as either direct, delay or support activities. This type of analysis often yields far different results than those from simple estimates, and can identify potential improvement opportunities that otherwise would be missed.

We can use an LAA to quantify where time is lost, loss categories, and to measure the amount of gain expected from specific improvement initiatives.

Conducting an LAA involves four steps:

Preparation — the definition of the scope of the study area and demographics of the craft personnel in that area; definition of sampling routes and calculation of sample sizes; and training observers;

Observation — the recording process and determining types of work activity;

Data compilation — data entry and summary of observations to facilitate grouping by loss category; and

Analysis and reporting — assessment of results for use in subsequent barrier and gap analyses.

Preparation

Prior to performing an LAA, you must review site information to develop a plan for execution. Typical preparation activities include:

• Defining the scope or area of the facility to be studied and identifying applicable technician work areas, e.g., production units, shops, stores and planning areas.
• Determining the number and type of personnel involved in maintenance in that area — if the whole plant is to be assessed, then the entire technician workforce is included.
• Establishing routes to be walked by the observers that allow a scan of the entire study area; walking the routes with all observers to familiarize them; ensuring the combined walking routes will consistently cover what all technicians are doing in the area.
• Setting start times for each route and publishing a schedule.
• Deciding upon the degree of accuracy (sample size, number of rounds and shifts, etc.) needed for the study.
• Selecting the method for identifying the maintenance workforce.
• Scheduling orientation sessions with management, supervisors and the workforce.

After defining the study area on a site layout drawing and the applicable maintenance personnel that support the area, the preparation activity focuses on the route planning and scheduling the number of observation routes required for statistical accuracy.

The plan identifies the walking routes to be used by the observers while the schedule documents the number of tours/routes to be taken during the survey. Both need to be developed early in the preparation period because they drive the accuracy of the study and the length of time required.

Additionally, proper planning will ensure that the initial hours of the first day of the survey result in effective observation sets.

Steps involved in the route planning process include:

• Identifying and tracing the routes on a plot plan and pinpointing start/stop points for each route.
• Establishing walking directions for each route, e.g., clockwise or counterclockwise, to allow it to be randomly walked in either direction.
• Determining a schedule of specific start times to begin the route walks during the day.

Typically each route should be able to be walked in less than 45 minutes. To assure statistical validity of the survey, 85% to 100% of the technician population should be observed. Meeting these objectives may require multiple routes. If so, they should be walked simultaneously and, hence, more than one observer will be necessary.

To achieve this coverage, the routes must be planned with an awareness of the population size and the typical location of the technicians throughout the day. During the observation period, the adequacy of population coverage should be monitored by:

• Comparing the number of shift personnel scheduled to be working the area each day against the actuals observed.
• Calculating the coverage percentage and determining any changes to the observation process that may be required.

The number of required rounds or sets of tours of the facility is determined by dividing the number of observations needed by the size of the workforce. A round should be walked during each hour of the shift to ensure observations are made throughout the day. The accuracy of the survey increases as additional observations are made. Typical accuracy levels would be to 85% to 95% confidence levels at either ±5% or 2.5%, respectively. Assuming that direct activity follows a normal distribution, the formula used to find the minimum number of observations required, n, is:
n = z²p(1 - p)/e² where z is the corresponding value for the desired level of confidence (1.96 for a 95% confidence level), p is the estimated value of the population proportion (i.e., estimated direct activity; we typically use 0.5 or 50%, as it represents a best-of-best target), and e is the maximum likely error acceptable for the confidence interval (i.e., for ±2.5% use 0.025). Therefore, using the aforementioned values, a total of 1,537 observations must be made (for 95% confidence) of the total technician workforce in the area.

For plants or areas with large workforces, identify the maintenance staff clearly to avoid confusion with other personnel. A typical method is to attach color-coded tape or decals to hard hats. This and other issues associated with the conduct of the study should be discussed and outlined during an orientation session with plant personnel to make the objectives of the survey clear and explain the process while orienting management, supervisors, and technicians to the overall purpose of the LAA process.

Observation

Once the preparation activities are complete, the team is ready to begin taking observations in the field. It should classify work observed into three categories:

1.  Direct activity — doing work that directly advances a task’s completion;
2. Support activity — performing elements of a task that indirectly advance its completion; or
3. Delays — periods of lost time that deter direct or support work

Observers should follow a number of rules while they conduct the observation process. Typically, these include:

• Perform each round per the route schedule.
• Before starting a route, record the day and date, start time, route number, round number, start point and direction walked on an Observation Record Form.
• Walk directly to the route start point and begin the sample at the specified time.
• Record only those activities taking place ahead in the direction of the walk; do not record activities to the rear.
• If it becomes necessary to back-track a route, do not record activities observed during the back-tracking.
• Do not take observations during an authorized break, e.g., lunch time.
• Upon completion of a route, record the stop time.

The observers should not unnecessarily interfere with work-related activities of the technicians. They should present themselves in a professional manner, while being prepared to answer questions about the survey and its methodology. Additionally, they should be open to feedback on the process from technician personnel.

Categorizing activities

Because the objective of the observation process is to identify and record activity as direct, support, and delay elements, it is important to be clear about what falls into which categories. So, we’ll now look at examples of direct activity and at how to categorize indirect activity.

Direct activitities include:

• picking up or laying down tools while performing productive work;
• taking or marking measurements immediately prior to performing direct work;
• handling tools, equipment, materials or parts in the execution of a task;
• walking inside the immediate work area (15 ft. of the work assignment);
• inspecting equipment for proper fit or operation;
• cleaning or putting away tools during or after the completion of a task; and
• cleaning the work area during or after the completion of a task.

Observations not counted as direct activities must be recorded as either support or delay. To allow better understanding of the context of such indirect activity and to provide more detailed data for use in follow-on improvement planning, support and delay observations should be broken down into sub-categories.

Typical support activity subcategories include:

• getting equipment or tools from a store or tool room outside the work area (abbreviated GE);
• traveling outside the work area for equipment or tools (ET).
• getting material or parts from a location outside the work area (GM);
• travel for materials or parts outside the work area (MT);
• planning, including receiving, giving, writing or interpreting job instructions (PL); and
• travel outside the work area not related to tools, equipment or materials (TR).

Likewise, typical delay subcategories include:

• equipment delay for access to equipment, tools, or transportation (ED);
• procurement delay for materials or parts at a warehouse or other storage area or searching for material or parts (PD);
• crew delay while another crew member or another crew assigned to the same task completes work (DS);
• supervisory delay in getting instructions to begin, continue, or complete the assigned task (SD);
• allowed breaks and shower and wash up time, and rest and relaxation for field personal (FP); and
• miscellaneous delay (MD).

Data compilation and analysis

Observation data must be entered into a database to compile summary information from the survey. This summary provides a roll-up of the data collected on a real-time basis. It thus allows process checks to be conducted during the observation process:

• determining whether enough observations are being collected for the sample size;
• Identifying if significantly different data are observed between days or shifts; and
• making adjustments to the study process if inconsistent data are received from observers.

Figure 1 shows an example of such a summary of observation data.