Handling Variations in Process Times
You don’t need to know how to do differential equations in this section but what is needed is patience to prevent you from becoming frustrated, basic logic and the ability to do arithmetic.
What usually happens is that managers make no attempt to do Work Measurement because a function’s processes vary in time. The other extreme is that followed by the US government where an elaborate effort is made to set Standards on all processes so that productivity can be measured. I disagree with the US government approach because a valid Standard can only be set on non-variable processes for the measurement of productivity. If a process varies in the length of time to perform it, then the measurement of a process’s productivity becomes invalid and therefore useless. Besides; most employees are smart enough to manipulate their productivity reports to their advantage. I don’t recommend productivity reporting because it is usually invalid and wastes the employee’s time I do recommend a daily log of the activities performed, the type of activity and the number of times it occurs. This will form the basis of the historical data needed for determining the time for the activity and in determining a budget.
The approach recommend by the author
First have a Lean Team study a function’s Process Flow to determine the best way of performing the function.
Second have a OIG Analyst analyze the Lean Team’s method and determine the staffing and expenses required for each process in a function and then document the function’s processes their staffing and costs in spread sheet. Then get out of the employees’ way and let them do their job. Don’t badger employees with recording and reporting their productivity. Staffing does not require measurement to the level of detail that a Standard requires for productivity reporting. The staffing of a function can be an estimate because you cannot staff with less than a full person.
All this effort is to arrive at a satisfactory staffing base with a bottoms-up budget. This simplifies the budgeting process based on real numbers rather than a top-down budget based on last years staffing and expenses.
How you handle variations in process times when staffing a function.
This depends on the characteristics of the process and how it varies. But don’t throw up your hands too soon remember it does not have to be perfect.
For analysis purposes a functions processes can be divided into classifications based on the characteristics of the work content of the processes. The following characteristics may be used:
Technical Capability Involvement
I have found that technical tasks can vary directly with the experience and capability of the employee doing the task. An employee with 3 years experience can do a job in half the time of a new employee. With a new employee I would staff using the pay level of a 3 year employee and expect to absorb the difference in pay grade in overtime expense for the new employee.
An engineer with 10 years experience may be five times more efficient than a new engineering employee. In this case you may have to hire more engineers when the experienced engineer retires but don’t change the budget until he retires.
Routine Activities versus Non-Routine Activities
This is the case where on the job decisions must be made. For example an inspector on a dock inspects incoming pallets of equipment. A time for doing this job can be estimated based on the number of pallets that arrive on the dock based on history. But what varies is when the inspector finds a defective pallet of equipment. This again is best handled by historical data. Remember if the inspector falls behind then due to increased problems then overtime may be necessary. To prevent overtime abuse I have recommended that all employees be a part of a team that works together to get the job done.
Where decisions are the job
All work involves continuous decision making it is just that we are so unaware of the routine decisions that we make that we don’t recognize them as decisions. All decisions can have a preset approach that can lead to more efficiency. Think of an original big decision as a set of known small decisions with something new added. This is generally done by management leaving the position just to be filled.
Then there are the lower level routine decisions which are repeatable within limits. Here again use historical data to determine staffing. I don’t recommend putting a time limit on making important original decisions other pressures will generally determine a time limit.
Estimating Workload by Determining Impact Triggers
One of the necessary jobs of a manager or any planner is determining how long it takes to complete a task. If the task takes too long it brings pressure on a tight schedule and if the schedule is too loose there is lost time waiting. I use the following process to analyze the task for what I call impact triggers. Impact triggers are difficult to explain because they only exist in the particular task at hand and must be identified in each task. They are the key parts to the task that cause it to vary in completion time. When estimating how long it takes to do a task look for those processes in the task that are likely to vary significantly. These are the items most likely to prevent you from meeting your schedule. If the number of these items are only a few then these need to be managed (watched) during the progress of doing the task. If these items are numerous (more than ten) then you should count off their completions. This gives you a lead time warning as to whether or not you will meet the schedule and gives you time to take corrective action by shaving time from other processes or by eliminating a process. You will probably recognize that this as exactly the process used to manage a daily speaking event on the campaign trail.
Now let’s examine this process in another example. A planner in a state highway department needs to estimate how long it will take to complete a road construction project. He does this from drawings, cut and fill estimates and feet of asphalt to be laid. The planner should first estimate all the processes in a strait forward manor using past data. Then he should look for triggers those items that bring significant variability to the project. There are two one is the extent and the hardness of rock lying in the cut areas and the other is the weather both can have a major impact on the schedule. If blasting is required to remove rock it can be very costly because all activities are shutdown during the process. The weather can vary significantly too much rain slows progress too little rain means that tank trucks must add moisture. When managing a project like this both the discovery of unforeseen rock and excessive rain are monitored closely in order to lessen schedule impact.
Assumptions in Work Measurement for Staffing
Staffing levels are infused with many relevant pressures that need to be balanced. Among them are: seasonal work, training, vacations, and special time off (bereavement and pregnancy). All of the above tend to cloud the staffing issue but none more than not knowing how long it actually takes to do the job. This can also be confusing because of a wide range of employee capability. New hires are expected to grow into the job while others my not have the capability or the will to master the details of the job. The key to Work Measurement is in finding an employee to time-study who has mastered all of the details of the job and may be the top performer in the department. It is easier rate this employee at 120% than it is to rate an inexperienced employee 60% and have the results come out at 100%. Where Work Measurement is not used employers tend to rely on history for decision making such as how many employees worked in the department last year.
This example of staffing I encountered at McDonnell Douglas Missile Systems Company demonstrates another issue. When I was making a preliminary assessment to find out how much time should be allowed to process engineering drawings from a new missile into work instructions in the Production Planning department. I encountered an eye-opening discovery. Although the engineering drawings were not finalized the preliminary drawings were forwarded to the planning department. I had been told that the planning was yet to be written and while I was interviewing an exceptionally qualified and experienced planner he simply turned in his chair pointed to the bookshelf full of volumes and said “its already done“. This one employee had somehow written all of the planning for a new missile and was doing it faster than new drawings could come out. The question occurred to me what are all the other fifteen planners doing. With experience and knowledge of the capability of machining processes and having applied all his knowledge from previous missiles this planner could do the impossible. Clearly management did not know what a highly skilled employee could do. A new employee would have taken months to do the same job because every decision made would have had to be checked out in the shop. If I had time-studied this planner even a high rating of 130% would have distorted the nature of the job but I would not have time-studied this planner. I would have picked an experience planner who better represented the group this planner was near retirement and no one could replace him.
My first job experience was as an Industrial Engineer doing time-studies on the final assembly line of the McDonnell Douglas F4 Phantom airplane. While it took only a few months to become proficient at doing time-studies. It took nearly two years for me to become proficient at balancing the assembly line but by the third year I had memorized all of the jobs on the assembly line. So when it came time to balance the assembly line I would balance the first part of the assembly line which was assigned to me and go on to balance the entire assembly line while the five other newer employees watched in amazement. In time the other five Industrial Engineers could be able to do the same. The problem was there were too many Industrial Engineers. The Defense Department had a policy of awarding contracts to the company with the most Engineers and no one was concerned about finding out how many were actually required. This was at a time when work measurement was only used in the shop areas and not in the office.
The conclusion is that Work Measurement should be the basis for all staffing. It can be done using the above time study or Lean team study data to determine the process times. The other issues: seasonal work, training, vacations, and special time off, should be added back in on a percentage basis to balance the work force. Balancing the workforce where there is seasonal work means trying to schedule training, vacations and time off in off peak periods.
A significant difference occurs when reducing staff during budget cuts. The present method tends to lay-off those at bottom usually the most recently employed. The government reform method using Lean Teams keeps in tact those employees that are actually doing the work and eliminates management staff.
Setting Standards and developing a Staffing Base
In industry standards are set for manufacturing widgets which are repeated day after day. The main purpose of the standard is to be able to plan in advance the work to be done including staffing and to determine the efficiency of the completed task using Performance Measurement. In government widgets sometimes are made but mostly you encounter repeatable processes to do a task which amounts to the same thing as manufacturing widgets. This may cover as much as 60% of government tasks. But for those who don’t make widgets the process of establishing a base for staffing is complicated and in non-repeatable tasks will require some time in the collection of data. The data can be used to support staffing requirements but because the base can vary it should not be called a standard nor should it be used for Performance Measurement.
I have indicated two main ways to set standards through Work Measurement they are time-study and Lean Teams. These two methods cover repeatable tasks very easily. Now for those areas of non-repeatable tasks where the processes vary in time I suggest this method for setting a staffing base. Select several knowledgeable employees who actually do the tasks. Have them collect data for a few months by recording each activity they do and the actual time to do the activity. From this data you can create the staffing base. This process is sometimes complicated but not impossible to do. Since this is a non-repeatable base you will have to rely on history for estimating the number of occurrences for budgeting.
In the most difficult instances the base can be created using a matrix of tasks and simply picking off those that apply to come up with a calculated base. A calculated base is most useful for planning the time length of the task for scheduling purposes and less for Budgeting. This is a one time calculation to aid in setting the staffing budget and for planning and scheduling but because of the day to day variations it should not be used for Performance Measurement.
Another way of approaching this problem is after collection of the work data look for repeated sets of processes of the same time length these become the base for some of the tasks or they may be combined into single processes and used for calculating the base for staffing. You may also look for key impact items (triggers) which make a set of processes unique with their own base. This approach is used also in estimating for contacts.
Decision Making relates to repeated occurrences of the same type of decision process. I don’t recommend setting standards for making decisions. The results of the decision is more important than measuring the time to arrive at a decision. The rule here is that important tasks are always more important than their productivity measurement.
A general rule for setting standards in a new area is that the task should be repeated at least several times before a standard is set. The first time a task is done can take as much as ten times the second time the process is repeated. Once standards are set they can be used anywhere in the state where the same task is used, that is where the task does not vary. They can also be elevated to Best Practices with recognition of those who developed the process. Why are Standards important? They can reduce budgets by more than ten percent.
The collection of data is the key to proper Work Measurement. And the recording of the number of times a function occurs is needed to establish a bottoms-up budget. I recommend that this data be collected by the Team Leader of each Lean team in a Daily Log.
Functional Restructuring Rather than Departmental Restructuring
Because Functions can precisely describe a task, breaking down a budget into Functions simplifies the budgeting process. This allows Functions to be easily excluded or included in the budget as a funded entity. Lean Teams study data for each Function is costed-out using Work Measurement as a bottom-up budgeting strategy. In contrast departmental level funding using the Top-down budgeting strategy uses gross estimates and blurs the issue of just where and how the funds are being spent leaving open the debate on the amount of departmental funding needed.
When several departments are to be consolidated breaking the departments into functions makes for a smooth consolidation allowing independent functions and their budgets to be easily moved to the new desired locations.
Reorganizing using Vertical and Horizontal Functions
The greatest danger with reorganization is simply throwing different departmental organizations together to reduce personnel by combining duplicate or nearly duplicate processes. This can lead to internal turf battles over petty issues like, how big my office is, or how far away am I from the boss’s office etc. Productivity gains can be lost in a quagmire of endless bickering. It is better to leave an organization in place and restructure its Functions.
What should be done is to organize a high level Lean Team make sure that Functions are clearly identified within a department and budgeted separately. Then if reorganization is necessary the Function can be more easily moved to where it makes the most sense taking its budgeted amount with it. Each department or organization contains one or more Functions and these are usually vertical Functions (stand alone functions). A Function must satisfy a specific “need” (service to the public) and contains a set of processes to accomplish that.
An example of a vertical Function is “elementary education” within the Department of Elementary and Secondary Education. An example of a horizontal Function is “child welfare”, which crosses such vertical Functions as ”foster parents”, “elementary education”, “secondary education”, “juvenile justice”, “public health” etc. Because of the fact that a particular state organization or department my have both vertical and horizontal Functional obligations well meaning persons may try to switch a vertical structured organization into a horizontally structured organization. It cannot be structured both ways and there is little to gained in this type of reorganization.
A high level horizontal Function is managed by a Steering Committee or a Cabinet (as some states call this organization) where most of the members come from the various vertical Functions that have obligations to the horizontal Function. This is a high level policy making group of stake holders who make the rules both practical and legal (with legislative approval) for the County level horizontal Function Lean Teams. County level Lean Teams are the best and safest way to provide management and over site for horizontal Functions such as “child welfare“. One of the obligations of the State level Cabinet is to review the Value Stream Maps developed by each County level Lean Team and to recommend the implementation of the best Value Stream Maps. The Cabinet is also responsible for the development of the budget for the horizontal Function. The Lean team’s Value Stream Maps provide the basis for calculating the number of caseworkers needed at the county level. Caseworkers may have to be moved to counties where caseloads have increased.
About: Needs, Functions, Funding Formulas, and Budgeting
A “Needs List” (mostly called by other names) is a list of the state’s obligations to its citizens in various areas: highways, foster children, health care for low-income seniors etc. Each need is satisfied by one or more Functions. Functions have processes that are depicted on a Value Stream Map in detail by each Lean Team organized and self-managed by the employees in the Function. The man hours required to do a Function is derived directly from the flowchart. Other cost inputs such as purchased items, mileage etc are tied directly to each process so if a process is eliminated all of the costs involved are also eliminated. The Value Stream Map becomes the input for the development of a funding formula for the Function and it in turn becomes a direct input to the Budgeting system. Note that Functions are budgeted not departments or organizations which means that the cost of managing a Function is not included in the functions working budget. However management costs may be added by prorating the cost of management and overhead to each Function.
What we want to do is to cost out the Function for example, in highways per “highway mile” and in foster children per ‘foster child served” creating a Funding Formula for each Function. We now have the basis for forming a budget based on actual costs we simply multiply the number of highway miles needed or expected foster children by the function’s funding formula to arrive at the new budget. Using the funding formula method simplifies the whole budgeting process. The actual expenditure from the year before is then used in calculating next year’s budget. Note that for the first time the cost of managing the function can be calculated by subtracting the working budget from a previous budget which included management and overhead costs.
The development of a valid Funding Formula is difficult. Here is an example of the approach I would use for the function, “Foster children“. After the Value Stream Map has been developed by the Lean Team and mileage and other costs are added to each process on the flow chart. I would request that the Lean Team submit a proposal for the number (range) of foster children that can be safely managed by a caseworker. I would do this to obtain a consensus among all stake holders of when to add extra caseworkers. Having done this, note that there is a relationship between the mileage traveled and the number of foster children that each caseworker can manage. Rural areas have fewer cases but must travel further than urban areas. I would develop two charts one for rural and one for urban showing how many foster children can be safely managed in each and when this number or range is exceeded then an extra case worker should be added to the budget. The VSM should cover 80% of the foster child cases leaving 20% of the cases requiring additional expenditures and unique decisions to be made by the Lean Team. Thus 20% of the cases may have extra expenses that need to be budgeted for. Because these are not predictable they should be budgeted outside the funding formula. What I have done is to try and account for all of the variables that can occur in the funding formula thus making the final budget for the Function dependent on projections of case loads for each county either rural or urban. Each county budget is rolled up into the state budget. Where the Function’s budget is presented to the legislative budget committee along with last year’s performance numbers. The unique expenditures budgeted for 20% of foster children not a part of the Funding Formula are not included in the next year’s budget. I caution you that the above example was done to show how difficulties in getting a valid funding formula can be overcome but the actual process in most instances may be different.
There is also a “Wants List” those things that are good for the State but are discretionary and not a direct obligation of the state. An example of the items in this list is Tourist promotion, business promotion etc. Each want has one or more functions and as in the above Need’s List each has a funding formula for budgeting. The Functions in both the Needs list and the Wants list are combined and prioritized and then approved by the legislature. Functions at the bottom of the combined list may not be funded and those not approved are not funded.
Establishing a Funding Formula for Budgeting
The Funding Formula is the key to simplifying the budgeting process. While current state budgeting systems provide for Departmental level budget expenditures the actual expenditures at the functional process level are not known and are not measured. I call this budgeting method “Top-Down Budgeting”. Because legislatures meet only for a few months each year oversight of Departmental budgets is not possible leaving bureaucratic Departments with control of their own budgets. The budgeting method I recommend I call “Bottoms-up Budgeting”. The method involves measuring each Function’s processes and building up the actual expenditures for staffing and expenses for each Function to obtain an actual budget for doing the work of the Department. Management costs are budgeted separately.
Each function is broken down into the “one best set“ of processes determined by the self-managed Lean Team which does the Function. An OIG Analyst determines the man-hour cost (staffing) for each process from the Lean Team’s data or an Industrial Engineer Time Studies each process to obtain the same information. Additional expenditures are added to each process such as mileage and purchased supplies. All of the above is represented on a detailed flow chart with man hour costs, mileage and expenses being keyed to each process. If a process were to be eliminated all of the costs associated with it would also be eliminated. The funding formula is created from this data providing a solid base for the bottoms-up budgeting method.
Establishing a Balanced work Load
Most managers fail to understand the importance of having a balanced work load. This means simply having the right person, with the right tools, at the right time. This prevents workers from having to wait on other people to do their jobs. The best example of this is with the Boeing assembly of aircraft. After time study of the necessary jobs is completed the work is laid out with on a scale showing what jobs are to be in the correct order such that no worker has to wait on another. If this is not done workers will continually have to wait on others causing built in inefficiency. This is called Short Interval Scheduling if a supervisor monitors work progress at short intervals.
This also applies to jobs that repeated and are time studied but it can also help when planning and scheduling large operations.
No comments:
Post a Comment