Utilizing inventory flow models with suppliers

Farris, M Theodore II

A common challenge facing logistics managers is the problem of reducing inventory without adversely affecting such other components of the logistics strategy as cost effectiveness, product availability, process and materials flows, and timely delivery. Two common approaches for managing inventory levels are schedule optimization and cost changes. During the past decade the management of cycle time has increased as a means of inventory management. A useful technique for managing inventory levels through cycle time analysis is "flow modeling." The purposes of this article are to introduce flow modeling, describe the steps in the flow modeling process, and explain how flow modeling can be useful to inventory reduction programs.

The article is divided into five sections. The first section discusses the historical development of flow modeling from the early 1960s to the present. The second section describes the five steps of flow modeling. The third section focuses on key components of flow modeling. Next, specific applications of flow modeling are identified and discussed. Finally, implications of flow modeling for practitioners, teachers, and researchers are examined.

WHAT IS A FLOW MODEL?

Flow modeling is a tool to assist the management of cycle time at all levels of the organization and across all members in the supply chain. Flow modeling identifies both the time and cost associated with a process. Multiplied by a scheduling vector, the flow model will emulate the entire process. The flow modeling technique offers valuable insights to many of the strategic and operational decisions critical to successful operation in a competitive environment. The flow model utilizes components from common business decisions and may be used in both service industries and goods manufacturing.

Historical Development

Early conceptualization of flow modeling has been traced back to Hughes Aircraft's "inventory flow models" in the 1960s. Raytheon's refined use was detailed in a Harvard University "Leisure Products, Inc." case study in 1972. In 1977, Production Magazine cited Sirianni's modeling work at Westinghouse.1 Sirianni suggested using flow models as a method to calculate inventory requirements. In 1978, Gregory2 cited the lack of methods to calculate work-in-process inventory levels, particularly in long cycle environments. He also proposed the use of flow to balance inventory investment against other demand. Developmental papers by Colley3 and Davis4 furthered the concept by increasing the level of detail within the flow model by breaking each activity into geometric segments. Colley and Davis suggested using the flow models as a means of goal-setting and performance measures. Flow modeling by companies such as Bendix and IBM is cited in the 1975, 1978, and 1983 American Production and Inventory Control Society (APICS) Conference Proceedings.5 Wantuck6 points out the flow model shape is germane to a given business configuration and concludes that no one flow model fits all situations. In addition, Wantuck suggests the flow model can be used to do more than calculate optimal inventory requirements, it can be used to identify overage inventory. Clark7 proposes a macro perspective by linking flow models to corporate balance sheets and income statements. Hughes8 further adds to flow model capabilities by tracing multiple flow models on a single chart representing different types of inventory in the product cycle. He recognizes the difference between "operational" inventory and "optimum" inventory and suggests extending flow models beyond the immediate plant. The use of flow modeling continues to be an effective technique to understand and analyze any process. The contribution of this paper is to identify how inventory flow models may be used and extended throughout the supply chain.

Step I: Gathering the Right Information

Primary components of a flow model include:

Cumulative lead time analysis of the time segments that make up a process.

. Cumulative value analysis of the build-up of value over time.

Schedule stated in terms of the daily going rate (DGR).

. Build structure reflecting relationships of components.

Collecting accurate and valid information requires interaction with manufacturing, production, design engineering, scheduling and planning, inventory and production control, purchasing, finance, and the accounting organizations in the company. The initial collection of information is the most critical aspect of the flow modeling technique. It requires a repeated effort to define and fine-tune the information. After the correct sources have been determined, the gathering of information becomes routine.

Step II: Begin with Cumulative Lead time Analysis

Analysis of cycle time often utilizes a sketch of the cumulative lead time. Cumulative lead time is the identification of all the elements that make up the total lead time of a part or assembly. It may take a variety of forms originating with a simplified diagram representing quoted lead times. Mather likens a cumulative lead time analysis to a pipeline flowing materials and products from vendors through the plant and off to customers.9 Further maturation of cumulative lead time diagrams will offer increasingly more detail for each major component or subassembly.

Lead time analysis diagrams can be quite extensive and extend across the entire supply chain from the purchase of raw materials all the way to the final end consumer. A typical manufacturing firm holds 30 percent of its inventory investment in procurement inventory, 30 percent in operational inventory, and 40 percent in distribution inventory.10It is important to do a complete analysis of the entire supply chain to capture all inventory reduction opportunities. Lambert suggests "while interdepartmental integration is important and beneficial, it's only a first step. Clearly the biggest payoffs from all this will come through integrating the entire supply chain."l Development of a cumulative lead time diagram is a critical first step in cycle time analysis. It allows the user to understand the time frames of each activity within a process.

Step III: Reflect Dollar Build-Up

Mature cumulative lead time diagrams aid the understanding of the production process, but they are limited in quantitative value as they only reflect cycle days and not the value of the inventory. Measuring only changes in time may imprecisely report cycle improvement.

Consider a scenario in Figure 3 where two workers reduce cycle times in their respective processes. Assume each process originally takes 50 days. Worker A removes 20 cycle days from the raw materials end of the pipeline. Worker B removes 10 cycle days from the finished goods end of the pipeline.

It would be inappropriate to assume Worker A has achieved twice the benefit of Worker B just based on cycle days. It is intuitive that finished goods will maintain higher value than raw materials. Assume for this example that finished goods are valued three times higher than raw materials. Ten days removed from the finished goods portion of the process removes as much inventory from the process as 30 days removed from the raw materials end. Many cycle time reduction efforts focus solely upon the number of days removed from the cycle without regard for the value added during the cycle period. This assumption limits cycle reduction gains by misdirecting reduction efforts and causing suboptimal results. It becomes imperative to consider the multi-dimensional attributes of each segment within the process.

A second dimension provides more precise quantification of cycle day value by representing the cumulative investment of value for each stage in the cycle. A multi-dimensional representation may be differentiated from the traditional pipeline chart by designating it as a "flow model." In short, a flow model = cumulative lead time + cumulative investment valuation. At this point in the design, the model graphically creates a two dimensional pipeline where the shape represents inventory investment in a process. This relatively simple technique offers many opportunities to manage the overall business perspective.

Adjusting for Daily Going Rate (DGR)

Initial calculations for the flow model design are based on a daily going rate of one unit per day. To allow for easy computation for fluctuations in the rate of production, the flow model is comprised of a series of daily snapshots throughout the production cycle for manufacturing a single unit. Inventory levels reflect quantities held to support the manufacture of the single unit. A third dimension can be added to the model to adjust for a daily going rate larger than one by multiplying the variables along the Y-axis by the daily going rate. After adjusting for the daily going rate, the flow model takes a three dimensional shape (Figure 6). Most flow models are viewed on a two dimensional basis with supporting numbers reflecting the daily going rate.

Step IV: The Review Process

Peters states that "all of us must start `thinking in wholes' and across boundaries, not about 'my' job, 'my' function, or even 'my' company. We must, in fact, reorient all of our structures and procedures, and our attitudes from parts to wholes; vertical to horizontal; exclusion to inclusion; transactions to relationships; functional task expertise to multi-function project execution; borders to `spider webs' (networks)."12 The flow modeling technique supports this philosophy. In order to achieve a reduction in total cycle time, a company must shift its operational focus from individual functional silos to how operations interact and flow together. A manager must identify the real business process and ask fundamental questions about the way every piece of information and material flows.13

A review should be conducted with all organizational players to ensure correctness and encourage communication. When initially viewed by all participants, there are often conflicting opinions on what the true process should look like. This is not unusual and may be the result of functional silos. It signals suboptimization of the process and opportunity for improvement. Over time, definition of the process will evolve. Joint review provides valuable interaction between functional areas and helps define a commonly agreed process for all organizations.

Check Proportions

Pipeline diagrams often are not drawn to scale. Proper proportionality in the flow model diagram will further emphasize the relative importance of each segment in the process. Notice the difference in the twenty days required to accumulate parts at the vendor in Figure 7 when the graph is drawn to scale compared to the unproportional pipeline diagram from Figure 5. The use of proper proportion helps to graphically identify the relative significance of each segment.

Step V: Add More Detail

Flow modeling is a dynamic evolutionary process that continuously seeks to fine-tune more details in targeted areas.

KEY COMPONENTS OF FLOW MODELING

Management by Eye Technique

Readability of the flow model may be enhanced through the use of shading common key activities. One successful practice defines each process using the following classifications:

1. Value Add (adding value to product in form of labor or the addition of parts).

2. Materials Movement (adding time/place utility to the product or components).

3. Strategic Inventory (holding product at strategic locations in the process to accommodate the manufacturing process).

These classifications can quickly identify the responsibilities of functional areas. The example in Figure 8 classified activities by the functional organizations of manufacturing, distribution and transportation, and production and inventory control.

Flow modeling supports the "management by eye" technique by creating a single summary of the process flow in terms of time and financial measurement as well as identifying which functions manage each segment. The management by eye capability of flow modeling is helpful to identify processes with high variability. Van Amstel14 bases his pipeline analysis on the fact that "the longer and more complex the pipeline, the greater are the variations within lead times." Flow modeling reflects accumulations of strategic inventory to protect the process from the high variability. Reduction in variability will play an important role in cycle reduction. Flow modeling may be utilized to seek out these pockets of high variability.

Both Detail and Overview Capability

Flow modeling may help to identify the process for a single part in manufacturing and logistics operations. Users of single part flow models include department managers, manufacturing engineers, or purchasing buyers.

The flow model may be extended by combining many parts into a higher level assembly or product. The complexity of the flow model will increase if all the information from each part level is utilized. It may be advisable to utilize averages in appropriate areas to allow for a simpler perspective of the higher level flow model. Users include product managers, purchasing buyers, and higher level management.

The flow model may be extended further to show the entire supply chain and identify logistical relationships between customers and suppliers. It may be utilized to identify strategic placement of inventory in the channel to maintain serviceability or reduce critical lead times. The seams (points where two processes come together) generally will offer the greatest opportunity for cycle improvements because each process will tend to increase cycle time to protect the uncertainty associated with the other process.

Multiple product flow models could conceivably be combined to reflect a business segment as long as the product schedules were dependent on one another. The flow modeling technique may be utilized to understand a process from many different levels, ranging from detailed through overview.

Determining Optimum Inventory

The flow model can quantify the optimum level of inventory required under the current cycle process. "Optimum inventory" is defined as the total amount of inventory that must be loaded into a process to support the manufacture of one unit per day. Total optimum inventory may be calculated by determining the area inside the flow model. Each unit area will measure one value unit times one time unit and will represent one value unit of optimum inventory required to support the process. The magnitude of financing required to support production or manufacture of a new product or business is often underestimated. The flow modeling technique provides a realistic estimate of inventory investment requirements.

The changes in the optimum inventory level required for a different production rate may be calculated by comparing values from each flow model after multiplying by the daily going rate.

Utilize as a Tracking Tool

Changes in inventory levels may be explained by investigating three basic variables: cost, schedule, and cycle times. Inventories will increase if there has been an increase in scheduled production, cost, or cycle times. Conversely, they will be reduced if the schedule is cut back, costs drop, or cycle times are shortened. The flow modeling technique may be utilized to summarize the impact of multiple changes to schedule, cost, or cycle times. It may be adjusted to accommodate varying production rates. In short, it may be utilized as an effective tracking tool reflecting changes to the process or business.

Developing a Compression Curve

The shape of a simple flow model reflects the build-up of product value over time. As flow models become more complex, all components must be combined, or compressed, together into a single line to summarize the build-up of value through the entire cycle. An example is shown in Figure 9 where the two operations from Figure 7 have been combined to create a single line.

Compression curves differ in curvature depending on the nature of the operation. Operations with high materials content relative to the amount of labor added will have a convex shape. Those with high labor added content will have a concave shape. Compression curves provide visibility to the inventory and cash flow needs of the overall process.

Measure Inventory at Risk

When a customer places an order commitment on a supplier, the customer has an obligation to receive and pay for the completed order, or in the event of a cancellation, compensate the supplier. This obligation represents risk by the customer. In the event of a cancellation, the customer should expect to compensate the supplier only for the amount of value added. Extreme situations of 0 percent cancellation charge or a 100 percent cancellation charge are financially unfair to one of the parties. Disputes arise when attempting to determine equitable financial responsibilities.

The flow modeling technique and compression curve shown in Figure 10 provides the visibility of the process to derive an unbiased assessment of the amount of value added at any point in the process. This measurement has been called "inventory at risk," which is the maximum risk a customer will incur in terms of inventory or cancellation charges in the event an order is canceled. For ongoing business relationships, use of the flow modeling technique and compression curves may be useful to help resolve disputes and ensure fair treatment of both parties. In the example, $1,800 would serve as a starting point for discussion of cancellation charges. Utilizing a flow model, both parties would have access to common quantitative information prior to discussion of cancellation charges. Other factors, such as scrap value or alternative use for the canceled order, would modify the cancellation charge from the initial starting point.

Engineering changes may cause cancellation (or postponement) of supplier orders and may result in significant addition to overhead costs. In 1985, prior to cycle reduction efforts, each Northern Telecom location (on average) made an engineering change to an existing product once every two hours. The cost of these changes accounted for more than 20 percent of manufacturing overhead.15 In a situation such as this, the ability to measure "Inventory At Risk" would be applicable to highlight the cost to cancel an order.

Postponement Analysis

The flow model may be utilized for postponement analysis when rescheduling delivery dates. Upon notification of postponement, it is assumed the supplier will remove the product from his process and hold it until it is time to reinsert it into the process to coincide with the new delivery date. (The alternative is for the supplier to complete manufacturing the product and hold the higher valued finished product for later delivery. Unless there are technical limitations preventing the interruption in the process, it is not advisable to continue to add value to the product as this will result in higher inventory carrying charges.) The compression curve is useful in determining the value of inventory and holding charges to be incurred. Holding charges should be calculated to cover the period of time the unfinished product will likely be held.

Quantifying Cycle Improvements

Flow models may be used to quantify the following benefits from cycle improvements:

One time invenlory reduction.

Annual savings in carrying costs.

Unit price reduction.

Reduction in "Inventory at Risk.

Calculating One-Time Inventory Reduction

A change in cycle time will result in a one-time reduction in the level of inventory carrier to support the process. The savings will be in the form of excess inventory assets that may be reduced over time. Graphically, a reduction in cycle time shifts the compression curve toward the origin of the graph as shown in Figure II.

The one-time reduction in inventory can be quantitatively determined by subtracting the area under the revised compression curve in Figure 11 from the area under the original compression curve. The amount of the reduction represents the amount of capital that will become available for an alternative use.

Calculating Annual Savings in Carrying Costs

In addition to the one time reduction in inventory, there will be a reduction in the annual costs to carry the inventory that was previously tied up in the process. After excessive inventories are eliminated, the annual savings may be calculated by multiplying charges associated with carrying this inventory by the one time savings in inventory calculated above.

Calculating Unit Price Reduction

The costs of carrying inventory are included in the final unit price of the product. Reduction in inventory carrying charges may be reflected in a lower unit price. The annual inventory cost savings should be spread over all of the units produced annually. Unit price reductions may be shared between customer and supplier. The customer benefits from a lower unit price. The supplier benefits from increased profitability and a more competitive price.

Calculating Reduction in Inventory At Risk

When the compression curve moves toward the origin, inventory at risk is reduced. The change in the average inventory at risk can be a general benchmark to measure improvement. Average inventory at risk with the original curve is 50 percent of the unit price.

Summary

The benefits quantified from the example are shown in Table 1.

INDUSTRY UTILIZATION OF FLOW MODELING

Many companies have utilized the flow modeling technique for a variety of applications from:

Reduction in inventory risk.

Improving supplier/customer relations.

Designing a new process.

Re-designing a current process.

Competitive analysis.

Supplier sourcing decisions.

Reduction in Inventory at Risk

A world-class manufacturing electronics company has been using flow modeling to analyze the degree of risk associated with channel relationships. The company has modified supply contracts to utilize flow models as a means of settling disputes concerning cancellation or scheduling changes. Contract terms require updating the flow models twice a year. All improvements are shared equally between the parties.

Improving the Supplier/Customer Partnership

A domestic manufacturer of printers has utilized flow modeling to improve the process flow with suppliers. The company is encouraging suppliers to implement flow models with their sources of supply.

After implementing a flow model to understand the process, bottlenecks, and associated risks, the printer manufacturer contractually agreed to accept full financial responsibility in the event of complete order cancellation (to a maximum of $40,000) for strategic raw material stores held by a key supplier. In return, the supplier agreed to hold additional strategic stock at his expense to accommodate unexpected schedule increases by the customer without extending lead times. Sharing strategic risks reduced cycle lead time by over 17 percent.

Designing a New Process

Concurrent engineering has had a dramatic effect increasing productivity and reducing costs for many companies. Flow modeling was used when designing assembly and test operations to house the IBM 3990 Storage Control Unit. In-house cycle manufacturing time dropped from 45 days for its predecessor 3880 to two days. Space to house the operation required one quarter of that for the 3880.16 IBM initially constructed a flow model for the 3880 process build, then drew boundaries on the flow model to represent the cycle reduction goals. Purchasing lead times extending beyond twenty days were subject to challenge and potential re-sourcing. Any in-house manufacturing process extending beyond five days was re-engineered. The end result was an order-to-finish process contained within a twenty-day time span that provides quick response to the market.17

Re-Designing A Current Process

A major airbrake manufacturer used flow modeling to map the manufacturing process. The flow model identified queues in front of 27 test bottlenecks. Information from the flow model was utilized to justify the expense to rework the process and reduce it from 14 days down to four days.

Competitive Analysis

A producer of cathode ray terminals (CRTs) utilizes flow modeling to reverse engineer competitor products. The competitor's product is disassembled and estimates for manufacturing cycle time and cost are combined into the flow model. The competitor's flow model is printed on a plastic transparency and overlaid on top of the producer's own CRT manufacturing process for comparison.

Supplier Sourcing Decisions

An airframe manufacturer requires suppliers to include a flow model when they respond to a RFQ. Flow models incorporate many of the logistics activities into the total all-in cost analysis. IBM-Rochester, winner of the 1990 Plant Malcolm Baldrige National Quality Award, extensively utilized flow modeling for sourcing decisions for the AS/400.

SUMMARY

The flow modeling technique is a tool to help focus management direction. Used correctly, the technique can be beneficial to all members in the supply chain as summarized in Table 2.

Characteristics of a flow model include:

Can range from simple to complex details.

The result of multiple function inputs.

The result of many decisions.

Segments are typically managed well.

Opportunities are usually found in seams between processes.

Principle benefits of flow modeling include: . Ease of understanding lead time.

Identifies critical processes.

Usable as a negotiation tool to reduce "inventory at risk" and to ensure: fair treament of both parties.

Offers detailed perspective of a product.

Offers overall perspective of business including multiple members in the supply chain.

The flow modeling technique simplifies analysis of the relationship between inventory and changes in cycle. It incorporates cost, schedule, and cycle time; and may be an effective tool to understand and help strategically manage the supply chain.

NOTES

lRobert F. Huber, "How Much Inventory Should You Have?" Production Magazine (March 1977): 62-64; and N. C. Sirianni, "Inventory: How Much Do You Really Need?" Production Magazine (Nov. 1979): 70-74.

2Donald J. Gregory, "Zero Base Inventory Profile System," in 21st Annual APICS Conference Proceedings (Falls Church, Va: American Production Inventory Control Society, 1978).

3John L. Colley, "Inventory Flow Model," Colgate Darden Graduate School of Business Administration, University of Virginia, 1977. 4Edward W. Davis, "Inventory Flow Models," Colgate Darden Graduate School of Business Administration, University of Virginia, 1980.

SJ. Nicholas Edward, "Target Inventory Levels," in 18th Annual APICS Conference Proceedings, 1975; Kenneth A. Wantuck, "Calculating Optimum Inventories," in 21st Annual APICS Conference Proceedings, 1978; and James T. Clark, "Inventory Flow Models. . . Preparation for Zero Inventory," in 26th Annual APICS Conference Proceedings, 1983, pp. 85-103.

6Kenneth A. Wantuck, "Calculating Optimum Inventories," in 21 st Annual APICS Conference Proceedings (Falls Church, Va: American Production Inventory Control Society, 1978).

7James T. Clark, "Inventory Flow Models. . . Preparation for Zero Inventory," in 26th Annual APICS Conference Proceedings, 1983, pp. 85-103. 8R. Hughes, "Computer Automatic Profiling System," IBM Manufacturing Technology Digest 6, no. 1 (1988).

9Hal Mather, "Want To Be More Competitive? Concentrate On Attacking Non-Value-Added-Wastes," IE Magazine (May 1991): 22-26, 73. 10Bernard J. LaLonde and James M. Masters, "Logistics: Perspective for the 1990s," typescript.

ll 1 Mitchell MacDonald, "Integrate or Perish," Traffic Management (Oct. 1991): 31-36.

12Tom Peters, "Time-Obsessed Competition," Management Review 79 (Sept. 1990): 16-20.

13E. J. Muller, "Turbo Logistics," Distribution 89 (March 1990): 28-36.

14 M. J. Ploos van Amstel, "Managing the Pipeline Effectively," Journal of Business Logistics 11, no. 2 (1990): 1-25.

15Roy Merrills, "How Northern Telecom Competes on Time," Harvard Business Review 67 (July-August 1989): 108-114.

16Roy Merrills, "From Frame to Finished Product: The 3990 Manufacturing Team Fine-Tunes Its Process," IBM Panorama (May 1988): 22-23. 17Robert C. Creese and Ted L. Moore, "Cost Modeling for Concurrent Engineering," Cost Engineering 32 (June 1990): 23-27.

M. Theodore Farris II University of South Alabama-Mobile

ABOUT THE AUTHOR

M. Theodore Farris II is an assistant professor of logistics and transportation at the University of South Alabama in Mobile. He was awarded his doctorate in business logistics and management information systems at Ohio State University in 1994. Prior to pursuing an academic career, he was employed with International Business Machines Corporation and the INTEL Corporation.

Copyright Council of Logistics Management 1996
Provided by ProQuest Information and Learning Company. All rights Reserved