If you could reduce the amount of finished-goods inventory your company keeps in stock while simultaneously improving customer service levels, would you do it? Of course you would! But is that even possible?
Indeed it is, as some consumer packaged goods (CPG) companies have discovered. These manufacturers have found that they can achieve both of those objectives by shifting from a traditional make-to-stock operation to a produce-to-demand environment. It's not easy, since successful implementation of produce-to-demand requires a break with traditional CPG operations. Moreover, this approach is not appropriate for every company. But in the right circumstances, produce-to-demand can bring about remarkable improvements in finished-goods inventory, order fill rates, and supply chain costs.
This article will help you understand exactly what produce-to-demand is, why it is so powerful and attractive, where it fits and where it doesn't, and how to attain the savings it offers.
What is produce-to-demand?
Consumer packaged goods supply chains generally operate on a make-to-stock basis. As shown in Figure 1 (page 56), in a make-to-stock environment, demand plans are generated from the combination of sales history and planned future events, such as promotions and price changes. The demand plan then drives the creation of a master production schedule (MPS).
In this type of system, manufacturing plants focus on producing according to the MPS and on minimizing manufacturing costs, a practice that encourages them to make product in large batch quantities and to resist production line changeovers. In addition, in order to provide a 99-percent- plus case fill rate for customer orders, manufacturers hold finished-goods inventories that typically range from two to seven weeks of sales. This inventory is expensive to hold: working capital is tied up, there are storage and handling costs, and it's a challenge to manage stocks as items near their end-of-life stage. If this approach fails and a stock-out appears imminent, then plants acquiesce to short-term schedule changes that often are viewed as operational crises.
The classic make-to-order scenario, meanwhile, is not a useful concept for CPG products. These products generally are uniform regardless of which retailer-customer they are sold to, and manufacturers typically receive many orders per item per day. Nor does it make economic sense to assemble to order from pre-made components such CPG mainstays as food and cleaning products.
There is, however, a way for CPG companies to stay in the make-tostock world without incurring huge holding costs—operate with very small finished- goods inventories and the capacity to make product on a short cycle. This is the essence of produce-to-demand: reliably producing goods on short notice when needed, and having processes in place so that such an operation becomes routine rather than disruptive.
As Figure 2 shows, produce-to-demand involves some information flows that are fundamentally different from the classic make-to-stock model. The power to make production decisions shifts from the longer-term master production schedule to a daily schedule. Information about customer orders, local finished-goods inventory, and replenishment requirements at remote distribution centers flows directly to the production scheduling process.
Do you still need a master production schedule in a produce-to-demand environment? Yes, because you still have to plan the acquisition of materials for each plant. But the production objective is no longer attainment of the master production schedule, per se. Instead, the focus will be on making just the right amount of product, as needed, in order to pursue the ultimate objective of minimizing total supply chain costs while delivering the best possible customer service.
What about continuing to generate a forecast? That will still be necessary too. You must have something on which to base the master production schedule, and you will still have to plan for known future sales peaks. However, the demand plan will be a little less important to daily operations, hence forecast errors will have less costly consequences.
Is your business a candidate?
For produce-to-demand to be both feasible and economically attractive, a business must have particular kinds of production technology, product characteristics, supply chain organization, and information visibility.
Production technology. If you are going to substitute short-cycle production flexibility for finished-goods inventory, then production has to be reliable. Planners must be able to rely on the plant to make product on time, almost without fail. They also must be able to count on the plant to produce good-quality product at the expected rate, and to be able to change equipment over to make each product in a predictable and reasonably short amount of time.
While produce-to-demand does not require five-minute changeovers, it is not compatible with fourhour changeovers, either. Because it generally will lead to more frequent production runs than a traditional make-to-stock policy, the plant must be able to absorb the lost production time and make the additional changeovers without much impact on costs.
Which brings us to another potential challenge: having enough capacity to meet short-term spikes in demand with production rather than with inventory. In our experience, produce-to-demand can work with 20 percent additional capacity above the average utilization if six or seven days of finished-goods inventory is maintained. Customers can be effectively served from a plant warehouse with only two or three days of finishedgoods inventory when there's spare production capacity of 50 percent. That's a lot of flex capacity, but it is not out of the question, because the capital cost of most CPG processing equipment often is less than the value of the safety and cycle-stock inventory that can be eliminated in a produce-to-demand environment.
Of course, capital-intensive industries like semiconductors, paper, or steelmaking, which sometimes have billion-dollar plants, can't afford to have significant unused capacity. These industries can sometimes use a finish-to-demand approach. For example, paper companies will keep master rolls (also called parent rolls) of paper in inventory and then trim them to order. In addition to eliminating finished-goods inventory, this approach maintains the necessary high utilization level of the truly expensive assets.
Product characteristics. Products that are suitable for produce-to-demand manufacturing share a few characteristics. First of all, the elapsed time from the decision to make a product until that product is available to ship must be relatively short—generally less than the customer's expected lead time. Poor candidates for produce-to-demand are products with slow processes, such as those involving fermentation or long quarantine holds. Most consumer products have short production cycles and meet this criterion. Produce-todemand can work in some other industry segments, including high-volume packaging in pharmaceuticals and high-volume distribution of industrial supplies.
The economics of produce-to-demand are much better when there are few components that are unique to each finished product (see Figure 3). To ensure quick production turnaround, raw materials must be readily available. If relatively few raw materials are made into many finished goods (for example, the few natural cheeses from which all of the flavors of a cheese spread are made), then little or no incremental raw-material inventory is needed to support a variety of produce-todemand finished goods. In CPG companies, in fact, it's not unusual for just the packaging materials (for example, the containers or even the container labels) to essentially match one-for-one with the finished goods. All of the main ingredients are shared, and these collective inventories are thus quite cost effective.
Supply chain organization. In the most successful produce-to-demand environments, suppliers take on much of the responsibility for maintaining component inventories. Moreover, quick responses are critical whenever material is required. For those reasons, forming strategic, collaborative procurement relationships with suppliers is very important when produce-to-demand is involved. Ideally, the supplier operates as an extension of your manufacturing operation. Maintaining this type of relationship with suppliers can make a big difference in produce-to-demand's total cost. That's because the suppliers' willingness to be flexible in regard to production and delivery and/or to hold safety stocks will directly reduce the amount of inventory the finishedgoods manufacturer must maintain.
But there are circumstances where suppliers cannot—or will not—keep product in inventory for quick delivery on demand. In one actual case, the supplier was 1,000 miles away from the plant, and the material price was too low to justify an investment in just-in-time (JIT) warehousing near the plant. Even so, the economic case for produce-to-demand remained positive because the CPG manufacturer was able to keep critical materials in inventory on its own.
Here's another example of how produce-to-demand can work when suppliers can't keep component inventories on hand. Suppose it is difficult to get product cartons on time for a produce-to-demand operation. It may be possible to store safety stocks of the required cartons at the consuming plant in a way that preserves overall produce-to-demand economics. For instance, empty cartons can often be stored folded or nested, requiring little space compared with storing the equivalent finished goods. Moreover, they typically are quite inexpensive compared with the assets tied up in equivalent finished goods. (Better yet, the supplier may be willing to deliver on a consignment basis, so you don't pay for the cartons until you use them). Thus, quick-response material stock can be available at a relatively low cost to support almost-immediate production or a decision at the last minute to extend a production run.
Information visibility. As Figure 2 suggests, produce-to-demand requires new kinds of information visibility. It requires systems that make customer demand and finished-goods inventories visible to plant schedulers on an almost real-time basis. You might think that every company with an enterprise resource planning (ERP) system would have those capabilities, but they do not. Whether those capabilities exist or not depends on the individual implementation. In fact, there still are environments where these key data for produce-todemand are not immediately available.
This kind of data availability is more critical than is having an automated scheduling system. While an automated scheduling capability can be valuable, particularly if you want to reschedule finished-goods production every hour or two in order to use the very latest information, in most produce-to-demand environments scheduling is still mostly manual. Realistically, they are limited to rescheduling finished-goods production only once or twice daily.
What must change?
Not surprisingly, the list of changes required to convert to produce-to-demand begins with addressing any elements identified in the discussion above that are missing. Some may be fundamental to your business and unchangeable, but you may be able to pursue others. They include:
Following this prerequisite work, the specific changes necessary to implement a produce-to-demand approach then fall into the following categories:
Changing production planning and scheduling processes. Produce-to-demand typically requires making relatively modest changes in the production requirements/MPS planning process. This includes planning against lower finished-goods inventory targets, planning more frequent production based on smaller batch quantities (if economically feasible), and reducing the number of days in the planning horizon that are considered nominally fixed or frozen. Once all that has been addressed, the master production schedule will be viewed differently within the organization, as it is now only the starting point for the short-range schedule and a guide for acquiring materials (see Figure 4).
Plant-level planners will operate under a new paradigm in a produce-to-demand environment. They must now watch the daily customer order stream and redevelop the short-term schedule in a way that manages small finished-goods inventories and meets order requirements—all while still maintaining production efficiency. They are no longer just reacting to schedule- change requests that historically have come from central planners or management; they are now driving a very dynamic short-range schedule themselves.
You might wonder whether frequent changes in the production schedule create havoc on the production floor. The answer is no, because produce-to-demand plants are very flexible. In a well-managed produce-todemand environment, rescheduling affects production beyond an agreed time fence—anywhere from a few hours to a few days into the future, depending on how quickly changeovers are permitted and how quickly required materials can be brought to the production line. Thus, it reduces the number of daily "emergencies."
Rethinking inventory policies. Produce-to-demand is all about reduction of finished-goods inventory. Implementing it in a traditional make-to-stock environment requires recalculating safety- and cycle-stock levels to take into account the shorter lead times with which plants can now make product. It also requires that work-in-process (WIP) and materials inventory policies (both for ingredients and packaging) be reexamined to support production of finished goods under short lead times.
Improving supplier communications. Traditional production environments often rely on purchases that are driven by material requirements planning (MRP), with standard lead times that are based on the master schedule. While it's possible to retain that approach for less-critical materials in a produce-to-demand environment, it's necessary to intensively manage with suppliers those items that you use in very large quantities and those that are unique to each finished good. If you do not do that, then you will likely end up with choking masses of those items sitting at your plant. The flow of these kinds of components should be set so that there is very little lead time (ideally one day or less) from the time you finalize the need for the material to the time it's delivered to your plant.
Manufacturers that receive components on a JIT basis will already be familiar with the need for frequent communication with suppliers when short-cycle delivery is involved. For others that are used to taking an "order it and forget it" approach to material supply, the daily supplier communication required to manage the flow of the right materials to the plant for today's and tomorrow's production will be a new experience.
As is the case with scheduling, good software tools that support supplier communications can be helpful, but in many environments simpler options are sufficient. For example, a communications-oriented spreadsheet posted on a shared website, showing daily/hourly delivery requirements versus actual current material status, can work quite well.
Aligning the organization. Perhaps the most difficult change required when shifting to produce-to-demand is the conversion of operations management from a department-centered, cost-minimization mindset to a supply chain orientation. In the traditional environment, manufacturing management is principally evaluated on production cost per unit, and senior management pressures manufacturing to reduce conversion costs. In a produce-to-demand environment, by contrast, manufacturing management will still be evaluated on production cost, but now it will be in the context of total supply chain economics, including inventory, warehousing, and transportation costs as well as disposal costs for excess and obsolete inventory. Management must adopt the new perspective if the right implementation decisions for produce-to-demand are to be made.
Often, making that happen will require introducing new metrics. For example, we have seen companies that have adopted produce-to-demand begin to track supply chain inventories in new ways (materials at suppliers, materials at plants, work in process, finished goods at plant, finished goods in distribution centers). Many of them also begin to compare the actual number of changeovers each week compared to the number planned in the original weekly schedule. These kinds of metrics assist in tracking the impact of produce-to-demand on plant changeovers and efficiency, on total pipeline inventory, and on the return on total assets.
It is also necessary to evaluate the quality of planning staff at headquarters and, most critically, at manufacturing plants. Because the complexity of production scheduling increases in the produce-to-demand environment—both plant efficiency and inventory must be managed, and rescheduling is more frequent—capable planners are an absolute requirement.
Test for success
The best way to convince skeptical management to adopt a produce-to-demand style of manufacturing is to conduct a pilot implementation that shows what this approach can accomplish. Such a test will not be costly; as noted earlier, the new processes can be executed without specialized software, and produce-to-demand can be implemented for an appropriate single product category at a single plant to demonstrate its economics.
But pilot operations are valuable for more than just changing attitudes. They are important laboratories in which to test the design of produce-to-demand processes for a business, a plant, or a product category. This will allow you to see just how well the combination of new processes and policies works, and to make adjustments to those processes and policies before rolling them out more widely. A successful pilot also provides insights into special challenges (for example, promotional demand or international products), and it often identifies tools and conditions that will be needed for a wider implementation and generation of sustained savings.
Produce-to-demand is a results-oriented approach to operations, a way to provide high levels of customer service at a lower total supply chain cost than traditional make-to-stock. It has been very successfully implemented in parts of the CPG industry and makes sense in some other industries, but it clearly is not for everyone. Implementing it successfully requires meeting several prerequisites as well as some careful rethinking of planning and operations processes and of inventory policies. Successful execution, with smaller inventory buffers, requires well-managed operations led by competent staff members. We believe that in the consumer packaged goods industry, the better operators will over time move a significant fraction of their operations to produce-to-demand, and in so doing will gain a distinct business advantage over less-sophisticated competitors.