Lean Planning for Medical Devices, Part I

Planning is a key part of any manufacturing business strategy, and is extremely important for an industry such as medical device manufacturing that supports life-and-death situations. Done right, Lean planning supports the Lean Flow business strategy by determining how, when and where a product is produced. Lean planning ensures customers get product on time, within stringent quality expectations and under cost targets, creating opportunities to sell the customer even more product. Planning done poorly will cost the company extra money, create chaos in the factory, and cause customers to find other places to buy product.

 

The “Lean” Journey

Most companies today, even some that believe that they are on the journey to Lean, use some very traditional methods to perform the planning process. The typical planning process includes the use of MRP/MRP II (manufacturing resource planning), a master production schedule (MPS), manufacturing orders, schedules, dispatches, hot lists, and parts kitting. In medical manufacturing, there are added regulatory burdens such as lot traceability and cumbersome documentation. For years, these techniques were all medium to large companies had available to them. Some did it very well and were successful; others did it poorly and had daily challenges to get product to market. In today’s fast-paced global marketplace, MRP and traditional scheduling techniques are not up to the challenges. That is why a vast array of advanced planning systems and sophisticated forecasting software tools have been created to “Band-Aid” fix the problems with this planning methodology.

Lean planning takes a different approach to the planning process. Is it necessary to scrap the entire MRP software tool and spend millions implementing new Lean software tools? No, because firms need MRPs item master, bills of material, costing, purchase orders and inventory data. What is needed is to take a different approach to the process of implementing some of the MRP tools, using the necessary fields/data, and looking at the entire factory differently.

Pull vs. Push

Lean planning uses the concept of “pull vs. push.” Push manufacturing tries to force processes and products to produce to a schedule, with multiple components trying to come to together at the same time based upon lead times, routings and bill-of-materials relationships. The traditional planning process tries to schedule each process with start and due dates, all with the assumption of infinite process capacity. When things don’t happen as expected, products build up, firms expedite, create hot lists, and jump through hoops to get things done. It takes a lot of work, people, paper, pain and money to get it done. All of it is non-value-added work. And MRP just keeps pushing work orders into the schedules of all the processes, regardless of what is happening downstream. In the end, many MRP-driven companies are only 60-80 percent on time to their customers based upon their own promise dates, and even worse if based on customer-request dates.

Lean planning still needs to collect demand and put it into a plan or sequence to produce, but orders are not pushed through the factory. Instead, they are pulled. Think of a Lean Flow factory as a large system of drain pipes, with each of the different flows of components/subassemblies and product being branches in the connecting pipes. Products are flowing off of the end of the last processes or assembly lines (through the largest pipe going to the customer) at the rate that needs to be produced to meet customer demand. If there is no product flowing through the end of the pipe as finished product, there is no pull of products through all of the rest of the system. No raw material can enter into the beginning of each of the connected pipes and be pulled, or flow, into the next pipe. If the rate of flow through the end pipe is increased then there will be more of a pull to flow more into all of the connected pipes. If a process could produce faster, but there is no place for it to flow because the downstream pipe is full, it can’t and won’t move downstream.

This flow, or pull, process has to manage demand, provide purchased materials to the processes and create a sequence of what will be produced. It must plan the rate of production to meet customer demand and manage the MRP data warehouse. Because there are no realistic alternatives at this time to MRP, most Lean Planning designs use the MRP software tool. Lean Planning just applies the tools differently to allow for a pull rather than a push system.

Actual customer orders are placed in order entry as before, but how they are presented to the factory changes. In a Lean Flow factory design, a dramatic change has been made; lead time has dropped significantly, as much as 30-99 percent, or an average of 75 percent. With dramatically shorter lead times, businesses can rethink some choices about whether to produce to stock or make to order (MTO). If the customer lead time is greater than the factory lead time, make to order becomes an option. If the customer lead time is still less than the factory lead time, then the business still will need to make to stock, but likely with lower finished-product inventory levels.

For MTO products, the planning is straightforward. Use the new lead times created by the Lean Flow factory redesign, plan to build the product within total lead time plus a little, provide the sales-order demand to the Lean Flow lines in the required sequence, and let them build it. There will be exceptions:

• Minimum run quantities that require aggregating demand from different sales orders or customers
• Sequencing rules that cause a short delay of hours or a few days, depending on the situation. Examples of a sequencing rule would be:

-Mixing the short-work-content-time products with long, to get a balance of time to the takt-time target;
-Sequencing the products requiring sterilization and a scheduled pickup time first, in order to be available for pickup on time;
-Sequencing the products together, if a group of products needs to be run in a batch, to create the batch as quickly and efficiently as possible; and
-Running a true mix of products all day long (ABABAB, etc.) may not be feasible due to regulatory lot-traceability and line-clearance issues. A better sequence might be to run all of the As, then all of the Bs, within the day. This should allow all products to be made every day.
 

• Material that is not kept in stock, even with a Lean materials strategy, must be procured. In this case, the product will not be started in the first process until the material arrives and is available. With a short, Lean Flow lead time, there’s no need to begin making the product until all material is available, and no place in a flow process for half-finished product to sit and wait for material.

The benefit of running all products, or as many as possible, as MTO is that there’s no investment in finished goods inventory (FGI). Some companies choose to make products to order, but put a small amount of consistent, high-volume products in FGI to meet the variability of customer demand. Medical devices can have a high rate of obsolescence or engineering changes, as well as a short shelf life. Making product to order reducing these risks.

To make products to stock, the FGI will be converted from a forecast-push-driven MRP model to a Lean Flow pull system. This requires creating an inventory of product that statistically has been sized to meet a company’s strategic target customer-service levels. Whether the service level is ensuring that 95 percent of the time a customer order can be filled from stock, or 97 percent of the time the product can be shipped within a predetermined lead time, the FGI is sized to meet the market expectations. An electronic kanban trigger mathematically is sized and put into the MRP system (using a reorder point, or ROP, field). MRP then signals a planner to make more product when the sum of on-order and on-hand drop below MRP’s ROP, or electronic kanban trigger quantity. So based on a pull—the consumption of FGI—more finished product is made in the factory, in a shortened Lean Flow lead time, to replenish stock. If no product is consumed by the customer, then none is replenished, or pulled, to stock, regardless of what the forecast predicts for sales. The opposite also is true: If product is consumed, a signal to produce is created and produced, regardless of the forecast.

A major component of the ROP level is the extra inventory that is statistically sized to deal with variation in demand. If the ROP was sized only on average demand, random variation in consumption would drive shortages, and the business would never meet its service levels and on-time-delivery targets.

The quantity made (minimum order quantity, or MOQ) when the ROP triggers is determined by how often the product can be made and its average daily usage. The replenishment of make-to-stock (MTS) product is driven by consumption, or a pull. Again, there always can be exceptions:

• A large export order that is six months of demand for limited number of part numbers, with plenty of lead time to produce, will be planned based upon availability of material and due dates. To wait until the FGI is pulled for shipment, have the electronic kanban (ROP) trigger and not fill the order complete, because of the unusual size of the order, makes no sense. Because of its large size and long demand-visibility horizon, this sales order should be made to order;
• Special sales events that will cause a huge spike in demand that the electronic kanban can not accommodate should have a “single-use kanban” created. Once the inventory has been created, it should be pulled out of available stock and applied to the sales-event order, so that the electronic kanban can support the normal daily demand for the product(s). If the sales-event inventory is not removed from the available status, it will be used to fulfill normal daily demand and the inventory will not be available for the sales event;
• For new product introductions, there will be no historical information to create an electronic kanban trigger or determine the correct minimum order quantity. The forecast will be the best available information until a couple of months of historical data can be accumulated. However, if the new product is a replacement for an existing product, the existing product’s historical data may represent the new product’s expected demand pattern; and
• There will be occasions when a new customer or market demand will increase the demand on a part number significantly. This sales insight must be communicated via sales and operations planning so that inventory planning may properly serve these new customers. Arbitrarily increasing the electronic kanban trigger (ROP) and MOQ will override the calculations, but this action might be required on an exception basis to better serve new customers and markets.

There will be other exceptions, but most products will be triggered using MRP’s ROP or electronic kanban to make sure product is available, strategic service levels are met and that the factory produces based upon consumption or a pull.

Planning rarely will interfere, expedite or create hot lists to move product through the factory. With the lead time reduced 75 percent by the Lean Flow factory redesign, there is no need to change priorities constantly. By the time you would try to expedite a product, it would be on to the next process or the customer. Without the need for a lot of planner resource time to expedite, the planner can devote more time to maintenance of the electronic kanban system, planning the short-lead-time, A-type items, managing outside processes (such as sterilization that can add a lot of time to the lead time), keeping a close eye on critical materials, ensuring proper regulatory documentation, and pursuing continuous improvement within the planning processes.

Preston “Jay” McCreary is a founding partner at FlowVision, LLC, a full-service provider of Lean and Flow manufacturing services. McCreary has been educating and implementing Flow Manufacturing concepts for more than 17 years, and he has more than 35 years of manufacturing management experience. His implementations have been used in a variety of industries, in more than 100 manufacturing companies.