Ask the Expert: Lean
The meaning of lean
A. I am not aware whether IIE has defined lean in its own terms. However, in the sessions that I have conducted for IIE, the following definitions of lean have been used:
1) “A systematic approach to identifying and eliminating waste (non-value-added activities) through continuous improvement by flowing the product at the pull of the customer in pursuit of perfection.” – MEP Lean Network
2) “A business system for organizing and managing product development, operations, suppliers, and customer relations that requires less human effort, less space, less capital, and less time to make products, with fewer defects to precise customer desires, compared with previous system of mass production.” – Lean Enterprise Institute
Merwan Mehta, Ph.D.
The right count
A. Inventory accuracy is a measure of how close the inventory records match the physical inventory. This is typically measured in dollars or by count. Applying lean techniques can reduce the need for inventory and reduce transaction volume. This will greatly increase the inventory accuracy but may take some time to reach this point.
There are several ways to improve inventory accuracy. The first is to perform a physical inventory that counts every item, compares the count to inventory and corrects the records as necessary. All normal operations cease for this activity. The next is to perform cycle counting. Cycle counting is a physical count of a small number of items on a set interval. Again, with this technique the physical count is compared to the records, and the records are corrected as necessary. The third way is to use process improvemnt methods on the transaction process. Finally, the most effective single way to improve inventory accuracy is to reduce the number of transactions. Reducing the number of transactions reduces the opportunity for error.
Ideally, you would want to combine all four methods for the most effective method. Using lean thinking, you would want to improve the process so that you could increase the inventory turns.
Elizabeth Cudney, Ph.D.
Batch vs. flow
A: One-piece flow is more efficient than batch production provided you cut the wasted motion out of the process. I agree that you have to fill the order. The assumption that the customer sales order equals batch size is not always correct, however. In some cases, several customer orders may be grouped together to make up a production batch. In other cases, planning factors in the ERP system dictate a certain batch size. If this amount is greater than the customer order, the balance of the production batch goes into inventory, tying up cash and space. These mismatches can distort true customer demand, leading to longer lead-times and inventory.
Even if you assume customer order equals production batch, cellular flow with a goal of one-piece flow is superior. Using your example of a 100-piece batch, let's assume each piece goes through three operations that take one minute each. In a one-piece flow line, unit one of production goes through all three operations in three minutes, resulting in the first saleable product in three minutes' lead-time. Unit two starts in the second minute, unit three in the third minute, and so on. After 102 minutes (about 1.75 hours), the entire batch is complete. In the case of a 100-piece batch going through three operations, the optimal lead-time would be 3 X 100 = 300 minutes (five hours) for the entire batch or for the first saleable product.
The reality of production is usually more of a contrast between batch and flow. Typically, the larger component of lead-time is queue time, which is the time delay caused by multiple batches lined up for the same resource. There are other delays with a variety of causes that further extend lead-time such as move time, wait time, equipment downtime, and setup. Much of this is removed in a cellular flow.
Additionally, in setting up a flow line, you study the process and remove waste. In a batch operation, parts may be put into containers for transportation to the next operation, where they have to be unpacked and re-oriented. Containers may be placed on pallets for transportation. Specialized equipment such as a forklift is required to move pallets. All of this packing, moving, and unpacking is eliminated in a cell.
You also study to balance the line and remove wasted motion such as unnecessary reaches and search time. You may discover that one operation is slower than another and can then redistribute work to allow better productivity and throughput. We typically videotape operations and conduct standard work analysis to remove waste. We uncover rework and non-standard practices, as well.
In your letter-handling example, many of the same issues apply. There is wasted motion to tray and untray batches of letters. There is a lot of extra movement in a batch flow. If we can design equipment or a manual handling process to handle letters in a continuous, one-piece flow, it will cut out a lot of waste. The trouble is that few businesses think that broadly.
A: Regarding maximum acceptable resource utilization, a lot depends on how accurately you have modeled the various tasks the resource will undertake. If you have accounted for every possible use of time, then theoretically you can push a resource to 100 percent. However, it is virtually impossible to model all eventualities. A rule of thumb is 80 percent to 85 percent utilization, assuming that major factors such as process times, breaks, downtimes, and other work tasks have been modeled. Another way to look at this is to plot the utilization over time and observe how often the maximum value is approached. You can also check the sensitivity of the system as high utilization is attained. What associated metrics are affected? Is it a queue buildup at a certain step? If so, how long does the queue remain after the utilization comes down? By varying parameters such as process times and rates, you can isolate the critical step to see if a certain high utilization level is really a detriment to the overall process performance.
Keep lean fair
A: Mass updates based on hypothetical assumptions should be avoided. There is a general tendency to pad time standards and not to report improvements to time standards in such a system.
Updating time standards after improvements are made can be done immediately after implementing a new method if operators agree that there is no learning curve. The important point is the concurrence of operators. When there is a learning curve, allow operators to become fully trained before setting a new standard.
Lean efforts fail when operators believe that management is pursuing continuous improvement simply to make them work harder.
Create working conditions in which trust exists between managers and the work force. This will set the stage for operators to want to produce more while continuously improving processes so that the company can earn more and share the added benefits equitably among the work force. Teams, communication, and management commitment to fairness — which are the tenets of lean — are keys to achieving this.
Merwan Mehta, Ph.D.
A: The changing role of the IE in a lean environment will require the following:
IEs will have to play the role of a catalyst for change. Working under champions such as the plant manager or the vice president, IEs will be able to do this. Traditionally, IEs were on management’s side — doing time studies to improve productivity — and sometimes attempting to help employees by making their jobs easier through ergonomics. In lean, IEs will have to be the true intermediaries, bridging the gap between management’s expectations and the shop floor’s realities to reduce waste in all forms and maximize the ratio of value-added activities to non-value-added activities.
IEs will have to be value stream managers, leading the charge to optimize the entire value stream through participation and involvement of shop employees.
IEs will be called upon to create value stream maps, identifying kaizens needed for achieving the future state in the value stream maps and leading the kaizen activities to achieve the future state.
Many times, it is absolutely essential to simulate the envisioned future state on a computer before implementation, and IEs will be called upon to simulate value streams.
Finally, IEs will have to have knowledge in the implementation of other lean initiatives such as process improvement (Six Sigma), total preventive maintenance, setup reduction, pokayoke (mistake-proofing), cellular flow and plant redesign, and incorporating design for manufacturing and assembly concepts into products. Only through knowing these tools well will they be able to use them well.
Merwan Mehta, Ph.D.
A: Lean manufacturing is the holistic consideration of all issues related to manufacturing to eliminate waste.
In terms of trade-offs, after considering the entire value stream, you should decide which options to accept. Hence, if attempting to eliminate waste in a process increases energy costs, the entire process needs to be evaluated holistically to the more acceptable alternative. Usually, the benefits realized through the implementation of lean principles far outweigh the additional cost in energy that might have to be borne when increased power is needed.
Merwan Mehta, Ph.D.
A: In the first cut of value stream mapping, you should attempt to optimize around 70 percent of your flow. After that, hand over the process to process owners or value stream managers on the shop floor to further improve it. To achieve this, you need to divide the entire product line into product families and then tackle similar products that make up a family as one unit to come up with the value stream map.
When you have a resource that is used by multiple product families, take into account the percentage of time the resource gets consumed by the product families. This will help you optimize the flow. The shared resource can then be placed in a layout so that it optimizes the flow of the product family that uses the resource the most.
In a custom setting, it is difficult to pinpoint on a day-to-day basis exactly what part of the resource will be used by which part family. Therefore, in custom shops it makes sense to use a generic representative product to help create a value stream map and optimize flow.
For example, in a custom machine shop, the general flow of products is cutting, turning, milling, heat treatment, grinding, and assembly. Some products might not need some of the steps, but you can create a generic representative product to allow you to follow the general sequence and optimize flow. Taking this approach, you can also create generic work cells through which products can flow to optimize the overall flow.
The main goal of value stream mapping is to create flow such that the percentage value-added time ratio (%VAT) is maximized. The %VAT is the total processing time required for a product family divided by the total amount of time it stays in manufacturing, stated as a percentage. This allows you to see for how much time the product was actually being worked on during the period that it was moving through the various operations in the shop.
Merwan Mehta, Ph.D.
A. While this seems like a straightforward question -- pallet loads vs. floor loads -- the answer is that it depends on your business. From a transportation expense perspective, floor loads will always result in a larger load.
Let’s look at your question in three areas: loading, transportation, and unloading. When loading the trailer at the distribution center, do you use fork trucks, pallet jacks, flexible conveyors, or extendable conveyors? Do you have the ability to sort the merchandise onto each pallet? The logistics department doesn’t want to ship by air, so they will prefer floor loads. How long do you wait for a trailer to fill up -- one day? two days? a week? What is the required frequency of deliveries from the DC to the store to maintain your level of service? When unloading at the store, is the trailer dropped or is it a live unload?
My experience is that stores are better able to handle smaller, more frequent deliveries. If product can be palletized by store zones, the pallet could be delivered from the trailer directly to the floor location in the store. I have seen various combinations of loading trailers. For example, one can co-load the merchandise for two stores onto one trailer to meet service levels. The front end is bulk stacked and the back end has pallets. The first store’s merchandise is on pallets that are unloaded live; the trailer is then dropped at the second store with a floor bulk load.
There is no one-size-fits-all solution for palletized loads versus floor loads. Determine your objective and constraints and cost out each method. In the end, it comes down to the number of footsteps and touches for your labor costs and transportation expenses.
Service takt time
A: As you know, takt time is the available time divided by customer demand. You never know the exact customer demand until the day is over; therefore, the takt time value is always an estimate that you would like to satisfy. History can help you get a fairly good handle on your customer demand estimate or you can use other sophisticated forecasting means to come up with a good estimate. For example, based on history you can estimate that on the Fourth of July you will need more attendants at fueling stations since more people would be on the road. Then based on the estimate of customer demand and your known lead-time to satisfy a customer, you can figure out how many attendants will be needed.
Talking about cycle times and lead-times, you are right in using the lead-times for satisfying a customer rather than the cycle time for applying the concept of takt time for services.
Sometimes, the way times are defined can cause confusion. The cycle time of a process can be the lead-time for the process, especially in services. This is because we define cycle time as the time needed for one iteration of the process to be completed and the lead-time as the time from the point of customer need generation to satisfying the need. One is based on the process, the other is based on the customer.
Unless these are the same or close to the same for services, you can have upset customers. This is because customers usually wait in front of the provider for their service, so if the cycle time for the performance of the process is not the same as the lead-time for the process, the customer will be upset viewing the wasted time. For example, if a plumber is fixing a broken pipe, the cycle time he or she needs to complete the process for the customer better be the lead-time that the customer is ready to wait for from the point the plumber arrives in the house to the moment the job is done.
Merwan Mehta, Ph.D.
A: The first step is to balance the work content. By adding a small conveyor, there will be WIP added back into the process. Standard work is a lean tool that will aid in balancing the work content. Standard work defines the interaction of people and their environment when processing a product or service. It details the motion of the operator and the sequence of action. Standard work also provides a routine for consistency of an operation and a basis for improvement. There are three central elements: takt time, standard work sequence and standard work-in-process.
Takt time is the frequency with which the customer wants a product or how frequently a sold unit must be produced. The number is derived by dividing the available production time in a shift by the customer demand for the shift. Work sequence is the specific order in which an operator performs the manual steps of the process. Standard WIP is the minimum amount of material or a given product that must be in process at any time to ensure proper flow of production.
Standard work provides a prescribed documented method or process that is sustainable, repeatable, and predictable.
Standard work is a tool to determine maximum performance with minimum waste through the best combination of operator and machine. Standard work helps eliminate variability from the process. It functions as a diagnostic device. Standard work also exposes problems and facilitates problem solving. It identifies waste and drives us to kaizen the process.
In a standardized work environment, production instruction is similar to continuous flow with verbal instructions, forecasting, and limited production locations. Product is manufactured to customer order with a defined WIP. The process is well defined to the work sequence for the operator. Machines are synchronized to approximately the same process speed. The lead-time for a product is predictable. Used as a tool, standard work establishes a routine for work and prevents backsliding. This makes managing schedules and resources easier. Relationships are also established between the operator, machine, and materials. Standard work provides a basis for making problems visual and obvious.
Elizabeth Cudney, Ph.D.