Real-Time -- One Domino at a Time

Real-time data can prevent distortions in product demand forecasts. But as Network Services' CIO learned, it isn't something that can happen all at once.

My company, Network Services, is getting ready to start working on its next four-year strategic plan. The "real-time" or "agile" enterprise is sure to come up as a topic of conversation. This has been an ongoing discussion in the company since our last strategic planning exercise, when we were urged by consultants to implement a real-time supply chain and order management system. It was a tempting scenario, and let me explain why.

Network Services is a distributor. It provides products and supply chain services to companies that need food service disposables, janitorial and sanitary supplies, and printing paper. We serve customers such as Baskin-Robbins, Premier Health Care and Starbucks. We sell products from manufacturers such as Georgia-Pacific, Rubbermaid, SC Johnson and Solo Cup. As distributors, we are the humble middlemen who serve both ends of the supply chain.

Twenty years ago, we called what we did "sellin', warehousin' and truck drivin'." Now we get to call what we do "supply chain management". I am ever so grateful to have such a fine phrase to use when people at a party ask me what my company does. But even though the name has changed, the realities of what we do have not changed that much. Change takes time. We learned this lesson very clearly in our last strategic planning exercise. It was the northern autumn of 1999, and we had Arthur Andersen working with us to facilitate and guide our strategy sessions.

Under Arthur Andersen's tutelage, we discussed what it would be like to detect changes in customer demand more quickly and react faster than our competition. We enthused over the possibilities inherent in tracking business processes as they happen. We saw how we could fine-tune our operations, maximise our efficiencies, lower our costs and increase inventory turns.

In the supply chain world that we distributors live in, the key to success is to find ways to counteract the distortion in the supply chain known as the "bullwhip effect". It occurs when small fluctuations in product demand by the customers at the front end of the supply chain get distorted as the information is transmitted back up the supply chain. The result is larger and larger swings in product demand.

All members of the supply chain feel the costs of the bullwhip effect. Manufacturers add extra production capacity to satisfy an order stream that is much more volatile than actual demand. Distributors carry extra inventory to cover the variability in order levels. Transportation and labour resources increase because extra capacity is needed to handle the periods of high demand and then sit idle during periods of low demand.

A synchronised supply chain that could dampen down or eliminate the bullwhip effect is based on a constant flow of accurate sales information from the companies at the front of the chain to all other companies in the supply chain. This information sets the rhythm that the other companies should move to.

However, it is one thing to provide customers and manufacturers with summary reports three months after the end of the last quarter. It is another thing entirely to provide people with daily data that is both highly specific and very accurate. This is sensitive stuff. As our CEO says: "It's only a matter of time before some company turns it against you."

The second thing we discovered is that using real-time data and being truly agile requires a high degree of internal coordination between departments. Different areas of the company often have conflicting goals. Inventory managers are motivated to reduce inventory. Salespeople are motivated to sell everything they can to the customers. Credit people are motivated to prevent sales that could result in hard-to-collect or impossible-to-collect customer invoices. Senior management has to create complementary incentive plans for all these groups, and then each group needs to understand the perspectives of the others in order to coordinate effectively. That's a tall order.

Page Break

We realised that taking our vision of the real-time, agile company from dream to reality was not something that could be done all at once. It was not just a matter of installing some software and teaching people how to start up the system. That would actually be the easiest part. The hard part would be learning new ways to do our jobs based on what real-time systems and data can do for us. A sudden flood of data would not make us any more efficient until we learned what to do with it.

As it turned out, our budget constraints at the time actually worked in our favour because they prevented us from embarking on a heroic but ill-conceived and hugely expensive ERP or e-commerce project. They helped us see the wisdom of using the "crawl, walk, run" approach as the best way to move toward the real-time enterprise.

In 2000, we built and rolled out the first version of a suite of e-commerce and supply chain management systems, and we have been enhancing and extending them each year since. These systems collect and move data such as purchase orders and invoices between our customers, our suppliers and us. They move data using batch cycles that happen every 15 minutes to two hours. Using these batch cycles, it is relatively cheap and easy to electronically exchange data between our systems and the systems of the companies we work with.

Using these batch cycles during the past three years, we have deployed supply chain systems that provide daily, updated sales history information and make it available over the Web. We and our customers can now monitor sales activity by customer location, product and manufacturer during any time period from one day to three years. This data has become critical to us and to our customers for planning, negotiating and managing procurement activities. We are starting to share this daily sales data with some manufacturers too. There are issues of trust and confidentiality (as in noncompete and nondisclosure agreements) to be worked out with each manufacturer when we share this data. It takes time, but it is happening.

Our customers are learning to use the data to plan their purchases and negotiate better deals with us. We are learning to do better sales forecasting and, in turn, negotiate better deals with the manufacturers we buy from. Operations people are learning to use the data to do better logistics management. Credit people are learning how to monitor sales and accounts receivable in a more timely manner. We are making progress one step at a time.

So where are we now? I'd say we are at about a fast crawl. For the past year or so, we have been rolling out an efficient transaction processing system that allows us to share five key documents (purchase orders, invoices, advanced shipping notices, product masters and price books) with more and more of the companies in our supply chains. As this project progresses, it will bring us into the walk phase: the creation of truly synchronised supply chains. Hello real-time, sense-and-respond organisation - here we come.

Mike Hugos is CIO of Network Services Company in the Chicago area and the author of Essentials of Supply Chain Management