Measuring Performance in Chaotic Networks

mThink Knowledge - Posted on 25 July 2003

Authored by:

William Walker, CFPIM, CIRM;

PDF File:

ASC5_wp_walker.pdf

Agilent Technologies

Business process orientation as a discipline, and supply chain event management as a technology, build a risk management framework for chaotic networks.

In a chaotic network, a different set of network nodes are interconnected when fulfilling each order-to-delivery-to-cash cycle. The central idea is that a particular node can "plug" into the supply chain network and "play" for a single order. Technology in the form of reverse auctions, supply chain collaboration applications, and supply chain event management along with the increasing ease of open, standardized interconnectivity has become both the driver and the enabler of chaotic networks. Technology allows a supply chain network to be defined and operated within a broader context of organizational relationships.

The dilemma is that, on the one hand, an organization trying to make inventory and capacity investment decisions in a data-starved environment is doomed to fail. On the other hand, once an organization connects to the network, it is expected to deliver in a competitive timeframe. There are investment risks prior to the network connection, delivery and quality risks during the connection, and performance-ranking risks after the connection. But without network throughput, there is no return for the organization.

This white paper briefly discusses one solution to this dilemma being the combination of business process orientation (BPO) discipline and supply chain event management (SCEM) technology. BPO discipline is used to define a standardized process interface for the single order scenario. Once an organization gets connected to the network, it will know what to expect. This is important because the process must work perfectly for each single order-to-delivery-to-cash cycle. SCEM technology is used to continuously sense the network while the organization remains disconnected. This is important because it helps to eliminate discontinuity at the connection, and it helps to minimize the investment risk. A good analogy is the way the revolutions per minute of a car's engine and the rpm of a car's drive shaft must be closely matched in order to shift gears smoothly. When there is a big mismatch, the car jerks, the clutch overheats, and the car's forward motion may be stopped. When a mismatched organization connects with the network, the supply chain jerks, incremental costs overheat and throughput may grind to a halt.

The Nature of Networks

Consider a middle node in a supply chain network. This node buys from upstream nodes and sells to downstream nodes. This node may simultaneously participate in several different supply chain networks. And this node plays a different role by different rules in each of these networks. In some networks the node will be a trading partner, meaning that it adds a core value and its financial success is largely dependent upon the end-to-end throughput of that specific network. In other networks the node will be a nominal trading partner,¹ meaning that while it provides an essential material or completes an essential connection, it is easily substitutable and its financial success is largely independent of the end-to-end throughput of that specific network.

A supply chain network has a finite life-cycle. It may exist only for a single order- to-delivery-to-cash cycle fulfilling a single customer order, or it may exist over a continuum of order-to-delivery-to-cash cycles fulfilling many customer orders. Network relationships can be seen to fall into one of three classes from this perspective:²

In a static network, the same set of trading partners are used to complete every order/replenishment-to-delivery-to-cash cycle throughout the network.
In a switched network, from time to time different trading partners are switched-in to complete a particular order/replenishment-to-delivery-to-cash cycle throughout the network, while other trading partners are switched-out.
In a chaotic network, a different set of trading partners is used to complete each order/replenishment-to-delivery-to-cash cycle throughout the network. At any time, a middle node may be simultaneously participating as a trading partner or as a nominal trading partner in a static, switched, or chaotic supply chain network configuration. For example, a component supplier might simultaneously sell directly to its largest customer, sell significantly through a distributor network, and sell opportunistically through a reverse auction. In this example, the component supplier is a trading partner in a static network, a trading partner (or a nominal trading partner) in a switched network, and a nominal trading partner in a chaotic network.

At an elemental level the middle node needs to be able to "plug" into the supply chain network and "play" for a single order. If the network is switched or static, the node stays and plays for a number of customer orders. All four subcycles must be completed before the node "unplugs" from the network. The subcycles are:

The upstream seller to middle node information-to-physical distribution subcycle
The middle node to upstream seller information-to-cash payment subcycle
The middle node to downstream buyer information-to-physical distribution subcycle
The downstream buyer to middle node information-to-cash payment subcycle

The successful middle node must learn to manage the relationship, the investment, and the risk as it plugs and plays to different customer orders in different supply chain networks. And if there is need for a return, then a reverse supply chain network must be in place with each of its own subcycles.

A Framework for Managing Risk

But, what management team can remain effective while disconnected from the network? The team's ability to forecast, deliver, and manage risk when only occasionally connected to a supply chain network distributed over geography and time is difficult at best. The answer to this question is the combination of BPO discipline and SCEM technology. BPO discipline is used to drive a mutual agreement around the definition of the end-to-end processes that every potential trading partner agrees to use. SCEM technology is used to sense the actual state of the network using real-time data. Capacity and inventory investment decisions and risk management strategies can then be grounded in actual boundary conditions at the time of switching into or out of the network.

Potential nominal trading partners get included within a chaotic supply chain network community by adherence to connectivity standards, standard process definitions, and common performance measures. There is usually some degree of prequalification. For example, raw-material suppliers invited to participate in reverse auctions are first asked to respond to a request for proposal (RFP). The RFP is used to screen out any potential raw-material supplier who is unable to source the required technology within a competitive lead time. Only those raw-material suppliers who meet certain minimum criteria are given password access to the reverse auction. When one of these suppliers wins the bid, that supplier is expected to immediately connect with the network for an information-to-physical distribution sub-cycle and an information-to-cash payment sub-cycle using the standardized process. The raw-material supplier's ability to respond with a competitive RFP in the first place is dependent upon that supplier's risk management in investing in capacity and inventory prior to the reverse auction.

BPO, as developed by Kevin McCormack and William Johnson, extends earlier work that recognized that competitive results are accomplished through horizontal processes across the end-to-end supply chain network (SCN) rather than through vertical organizational structures. McCormack and Johnson use the constructs of SCN values and beliefs, SCN process view, SCN process structure, SCN process job, and SCN process measures to define standard processes and common performance measures across legal company boundaries.³ McCormack and Johnson's research shows that as the trading partners are able to improve their level of BPO maturity, then network connectedness improves, network conflict diminishes, network esprit de corps improves, and network business performance improves. McCormack and Johnson define the process constructs as:

SCN values and beliefs: In a chaotic network, each potentially connected organization shares in a common view of the end customer and believes that all the trading partners involved in fulfilling that order are customer-focused.
SCN process view: In a chaotic network, each interface between potentially connected organizations is an extension of the standard process. This is achieved by sharing a common vocabulary and by using a mostly graphical representation to define process flows.
SCN process structure: In a chaotic network, each potentially connected organization understands the framework that defines the horizontal team and the shared responsibility that will be used to implement the customer order-to-delivery-to-cash cycle and its sub-cycles.
SCN process job: In a chaotic network, each potentially connected organization recognizes and respects the person responsible for owning the whole process. This person generally is an employee of one of the core trading partners, and remains responsible while other parts of the network are fluid.
SCN process measure: In a chaotic network, each interface between potentially connected organizations aligns with a common set of performance measures that reflect end-to-end performance for that particular customer order-to-delivery-to-cash cycle and its sub-cycles. BPO makes the transition of potentially connected organizations transparent to the end customer as they connect and disconnect with the chaotic supply chain network. Transitions are transparent in the sense that standard end-to-end processes with common performance measures and the same person responsible for the whole process are used to fulfill each order.

Sensing the Network

In many supply chain networks, the midstream trading partner relationships are fairly static and integrated by a single ERP application. A few dominant customers and suppliers may also be integrated into the ERP. In such a supply chain, the network is driven chaotically along the network fringes. Smaller downstream customers connect to the network, buying at random times and from random geographical locations. Smaller upstream suppliers sell opportunistically to the network at random times and from random geographical locations. Midstream and reverse-stream nominal trading partners, including logistics service providers, information service providers, and financial service providers, plug and play and unplug with the network under a variety of scenarios.

In such a supply chain network, the ERP system is incapable of matching supply with demand at the network extremes. The nominal trading partners are incapable of managing their risk because of being starved for supply data even while connected to the network. And the deselected nominal trading partners are incapable of making future investment decisions because of being starved for demand data while being disconnected from the network.

The solution is a second information network that all current and qualified potentially connected organizations monitor continuously. This second information network becomes the context for the SCEM application. SCEM is an economical, Web-based, wireless technology that can interface with disparate, cross-enterprise systems. While the SCEM information network is grounded in the ERP database, it defines a much broader connectivity context for those who need to shadow the core relationships (see Figure 1). The SCEM information network provides the primary means of managing risk and measuring performance in a chaotic supply chain network environment. The connectivity investment per node for SCEM is less than the connectivity investment per node for ERP.

Figure 1: Overlay of the SCEM and ERP Information Networks

SCEM software uses Web and wireless connectivity to inexpensively sample and normalize forecast, demand, inventory, capacity, and supply data from the ERP database and from all other data warehouses in real time. Normally, SCEM is configured to trigger alerts to the trading partners based on exception-reporting. These alerts signal that some action must be taken in real time to assure the promised service level to the end customer. But in this application the SCEM technology is used to "sense" the state of the chaotic portion of the network for those qualified nodes just connecting to the network and/or just disconnecting from the network. The technology allows demand-tracking on the relevant SKU groupings to make timely investment decisions for inventory and capacity. While a particular bid may still be lost and access to the network denied, at least each qualified nominal trading partner has access to the planning information necessary to manage its own level of risk.

Defining Performance Measures

Clearly the data definitions for sensing supply chain networks get a little tricky. First, chaotic network participants are not interested in becoming overwhelmed by all the data available from the whole universe. The relationship of the bill of materials (BOM) to the supply chain network and the definition of "equivalent throughput" establish a practical means for focusing demand and supply data into an appropriate information window. Equivalent throughput uses the product BOM to establish equivalency between fully manufactured downstream flows and partially manufactured upstream flows.⁴ For example, a distributor of semi-finished product only needs to sense the network over a specific range of SKUs and only within the network echelon(s) corresponding to the proper BOM level(s). The SCEM software has to be able to filter and normalize the feeds from different data warehouses. But it will be unable to detect problems caused by data corruption or data inaccuracies.

Second, while potential chaotic network participants have been qualified for this supply chain network, they are also sensing and participating in competing supply chain networks. This raises the question: Which information is proprietary? The answer is to differentiate among information that is public, information that is semiprivate for nominal trading partners, information that is private for trading partners, and trade secrets that must remain closely guarded within a single organization. Forecasts, demand, supply, capacity, and inventory are forms of semiprivate information when they are expressed in units over time. Forecasts, demand, supply, capacity, inventory, discounts, costs, and profitability are forms of private information when they are expressed in dollars over time.

Network-sensing works best when a few global performance measures become standardized in their definition. Reaching a consensus on data definitions is hard work and requires a commitment from all parties. Figure 2 shows a dashboard consisting of four standardized performance measures — demand, inventory, capacity, and throughput — for three scenarios. In the left scenario, the organization is disconnected from the network and information-starved. In the center scenario, the organization is using SCEM technology to sense a potential connection with the network. This might be the case when a contract manufacturer gets ready to bid in a reverse auction. In the right scenario, the organization plugs smoothly into the network and uses BPO standard process constructs to play the order-to-delivery-to-cash cycle. Once the organization unplugs from the network, it returns to the middle scenario of sensing until the next connect opportunity. An organization that simultaneously participates in multiple supply chain networks would maintain separate dashboards for each network.

Figure 2: A Dashboard to Facilitate Plug and Play

Summary

While a core set of stable trading partner relationships integrated through ERP technology usually dominates a supply chain network, the same supply chain may operate more efficiently when organized as a chaotic network of nominal trading partners along its fringes. The information technology investment to "sense" the network is less than the information technology investment to be fully integrated with ERP. In addition, doing business simultaneously within multiple supply chain networks becomes prohibitively expensive when each network utilizes a different ERP solution. The risk-return equation for such nominal trading partners is a big dilemma as they attempt to plug and play and unplug from the network. Do they have enough of the right inventory? Should they risk their capacity for this one order? If they invest ahead of demand, how long will they have to finance their investment? And, when they win a single order connection to the network, will the return be worth it?

This white paper has briefly explored the use of BPO as a discipline and supply chain event management as a technology to build a risk management framework for chaotic networks. BPO is used to set process standards and operational expectations at the interface. This allows each order-to-delivery-to-cash cycle to be executed perfectly when a different nominal trading partner is plugged into the network. SCEM is used to continuously sense the network regarding relevant demand and supply information. This prevents major discontinuities when an organization plugs into or unplugs from the network. In this way the level of risk can be managed to a level commensurate with the expected return.

Endnotes

1 Kevin McCormack and William Johnson with William T. Walker, Supply Chain Networks and Business Process Orientation: Advanced Strategies and Best Practices, Boca Raton, FL: St. Lucie Press, ©2003, page 107.

2 Ibid, pages 112-113.

About the Author

Title:

Supply Chain Architect