AutoID in Supply Chain
Computers are used to process, analyze and display information about many processes involving moving “things.” Entering the information about the status of moving things requires repeated data entry which is cumbersome, expensive and error-prone.

Consequently, many automated systems have been developed to perform this data entry task. Together these systems are referred to as automatic identification (“Auto ID”) systems.

Example application areas include product identification in consumer goods industries, swipe card access systems and GPS/on board computer systems used in trucking. The availability of such data entry systems has provided the impetus for development of more sophisticated decision support and control systems.

Supply Chain Trends and Challenges

The term “supply chain” refers to the series of players and activities that take part in the movement and transformation of raw material “in the earth” into finished goods at the consumers’ hands. Many analysts also include reverse flows (returns, disposals, recycling) in the definition of the supply chain. Clearly, the term “chain” is a simplification of the complex web of suppliers, sub-assemblers, manufacturers, distributors and logistics providers who are the primary actors in managing the physical flows “from womb to tomb.”

The management of supply chains can be summed up in two challenges: (i) being able to optimize entire systems, rather than subsystems and (ii) managing the variability inherent in supply chain operations. The first challenge stems from the restricted view of managers who are constrained by corporate boundaries, limited responsibilities and lack of supply chain-wide visibility. Thus, the best they can do in most cases is to suboptimize.

Variability management has always been important but is increasingly critical due to several current trends:

  • Globalization – Requiring longer and more complicated supply lines, inventory systems and distribution networks.
  • Outsourcing – Involving more entities in the supply chain.
  • SKU Proliferation – Resulting in demand desegregation with the resulting increased coefficient of variation.
  • Shorter product lifecycles. Resulting in lack of historical data, reducing organizations’ forecasting ability.

In addition, the whole external environment has become less predictable. The rate of technological change is high, competition is intense, and new risks, such as terrorism and the government actions designed to fight, it introduce new costs and uncertainty. In trying to understand the ways in which variability is dealt with we distinguish between two types of variability that might influence a supply chain operation:

  • External variability brought about by unexpected demand or supply fluctuations,
  • Internal variability brought about by imperfect internal process control. Traditionally, demand uncertainty is managed by using a forecast, coupled with fixed level-of-service guidelines, to set operational parameters (such as inventory reorder levels and shipping schedules). Supply uncertainties are handled by using a yield forecast to inflate manufacturing requirements or set inbound buffer inventory.

Naturally, both demand and supply are random variables whose various possible realizations are best characterized by a probability distribution, while the control systems which guide supply chain operations require specific numbers in order to act.

Consequently, all forecasts tend to be “wrong” and the resulting mismatch can be handles either by using buffer inventory or by building into the system a degree of ad hoc responsiveness, or agility.

Sources of internal variability typically include errors in determining the amount and location of inventory, errors in process yield or errors in predicting effort and time required for different operations. In addition to affecting stock holding and operational costs, these errors also hinder the ability to accurately determine whether demand can be genuinely satisfied, limiting the proactive capabilities of the organization. The impact of internal variability on the ability to manage demand fluctuations is significant, and it is the combination of these two factors that is addressed by Auto ID systems.

Many companies have realized that the key to success in environments with significant variability is to complement predictive capabilities (that is, forecasting and estimates of consumption and operational times) with a highly reactive capability as part of a single strategy for variability management. In order to be most effective, however, such a response capability must be reflected throughout the supply chain

One of the most important tenants of responsiveness is the ability to detect a variation, recognize its cause and act accordingly [12]. Clearly, the earlier a problem is detected the more time there is to recognize its nature and formulate an action plan. In fact, this is the promise of a relatively new type of supply chain management software tools dealing with event management.

Supply chain event management tools evolved from process control and are an outgrowth of supply chain visibility software tools. Visibility software flags deviations from plan (late shipment, missing components, lower-than-expected inventory, etc) so that logistics operations managers can act on it. Supply chain event management tools have added some capability for intelligent and (sometimes) automated response.

The Achilles’ heel of all such systems is the data acquisition – event management processes are completely dependent on the availability of accurate and timely data from suppliers and service providers as to where shipments are, what the current inventory level is, and where is it located. This is where AutoID can add crucial value to supply chain operations. AutoID can improve the data acquisition process, make sure that deviations are captured earlier and that the data are more complete and accurate – thereby giving supply chain managers more time to recognize a problem, assess its potential impact and take corrective action. Note that this is not only an issue of human reaction time. Early detection means that more variables can be manipulated and more options are open for a systemic response. For example, early notification of a delay of a shipment on the railroad may mean that it can still make its deadline by trucking it, while a later detection of the problem means that a critical part may have to be air-lifted.