How does a network become a network?

A successful Manufacturer buys many plants and goes global. Many plants make many products; a global company distributes its products to many markets. The most tempting form of synergy is to bring the products of many plants to multiple markets; and to eliminate redundancies between plants. This thinking leads to the idea of centers of excellence or plant specialization. Great thinking. However, all these plants weren’t born together. Initially, their operating structure and processes were based on an isolated presence in the market. Hence, initially, post acquisition, plants continue to operate as autonomous operations; the way that plants distribute to many markets is by becoming clients and suppliers of each other. After all, the purchasing model is the oldest form of networking and everybody is ready for it. Hence, a network of sorts evolves where organizations in each market order products from plants everywhere, using a simple model of client and supplier.

This model, however, is not really a network. When nodes behave as clients and suppliers, that introduces unnecessary layers of intermediation that slow down the network. They leave in the network the same layers of customer service and purchasing that exist pre-acquisition, discouraging harmony between the nodes and thus remaining slow and unresponsive.

In a real network, entities with requirements are elegantly connected directly to the entities with supply and the latter respond spontaneously to the former. One simple model of a fast network is the messaging model of publishers and subscribers found in software object networks . In this case, the network is an entity unto itself. Nodes with requirements subscribe to what they need; nodes with supply, publish what they have; the network provides the service of funneling supply to where the need is based on a publisher-subscriber model. Hence, ‘client’ entities don’t order from ‘supplier’ entities. They (clients) simply declare their needs and, somehow, supply appears.

In a purchasing-based network model, the network is as powerful as the number of direct commercial connections established between nodes. In a real network, the power of the network is the number of possible combinations between nodes and is virtually unlimited. The only requirement for a real network is that a protocol of behavior be established to the messaging across nodes, a small price to pay for so much competitive power.

In the materials world where the goods circulating through the network are materials manufactured by plants that function as the ‘supply’ nodes and the ‘client’ nodes are sales and distribution offices at each specific market, the ‘agent’ that matches publishers and subscribers is the practice of supply chain management. In order to perform this role, however, it is extremely important that SCM teams not be ‘captive’ in any of the nodes of the network; SCM teams must be part of the network itself. If you think of a network as many sites – some being supply sites and others being client sites – SCM is in between the sites, i.e., amongst the sites, in the network.

The practice of SCM today has evolved from the simple MRP model of the 90’s and 2000’s and has separated from the practice of operational planning or production planning (which plans operations within plants). SCM now is really a form of supply network management and entails methods that are excellent at managing the ‘messaging’ of needs and supply through the network.

How does it work, then? It starts with an information system. This would be an Advanced Planning System or APS which wouldn’t ‘belong’ to any one ‘node’. It would be a federated system where a model of the network would be defined within its data. Client sites, supply sites, component materials and finished goods materials, routings for materials moving from supply site to supply site (raw materials going from some sites to other sites where they would be transformed into finished goods); or from supply sites to ‘client’ sites for distribution to the markets – they would be described in the data of this federated system.

Client sites would then place their needs in this system. Needs can be described as forecasts of demand; or as specific orders for finished goods. This system would then translate these requirements into supply requirements for the various plants; as well as the transportation orders to cause materials to move from site to site. Supply Chain Analysts – the professionals that practice SCM – would monitor the requirements for each plant; insure that requirements are in balance with available capacity; and would then monitor the flow of materials from site to site and insure no disconnects (e.g., a manufacturer of finished goods shouldn’t be preparing to make product before the necessary component materials are available and ready). SCA’s would also monitor demand fluctuations and either accelerate or slow down supply accordingly.

Information systems, however, are limited. They are great at manipulating and moving information and they are great at supporting business processes. However, by and large, systems are based on a model of the world as it should be. When the world does things that aren’t supposed to happen, information systems are not the right tools. The right tools are social networking or social enterprise tools which enable spontaneous connections amongst individuals across the network. These spontaneous connections are the necessary supplement to what systems can do. Individuals across the network come together quickly and spontaneously as emergencies and unexpected events occur and they figure out the best way to cope with those events.

It is this arrangement of SCA’s in an extra-enterprise setting, sitting amongst the various supply and client sites, equipped with advanced planning systems and social networking tools, that create the network, I.e., that make the many nodes acquired by a multi-national behave like an actual network of great flexibility and fast response.

A corollary of this networking model is that the nodes of the network don’t have to be part of the same enterprise or company. Theoretically, one can plug actual customers as ‘client’ nods; and external partners as suppliers.

Of course, the various supply sites have to perform. Publishers have to be able to satisfy the need of subscribers, they have to deliver on their promise. If they don’t, there is nothing that SCA’s will be able to do. If materials don’t exist, if they are not produced reliably, they can’t be distributed and demand can’t be met. Conversely, if client nodes don’t collaborate and provide good information about market demand, the supply network will aim at the wrong target and this disconnect will cause product shortages in the markets.