Supply Chain Frontiers issue #39
Almost two decades after integrated demand signals (IDS) technology was introduced to improve supply chain performance, is the process finally on the brink of realizing its full potential? The MIT Center for Transportation & Logistics (MIT CTL), and the Pennsylvania State University Center for Supply Chain Research in collaboration with MIT CTL industry partner General Mills, organized a conference to find out.
The Capturing Strategic Advantage from Integrated Demand Signals conference took place on the MIT campus on November 16, 2010. About 30 enterprises were represented at the event, including major consumer goods companies such as Coca-Cola, The Hershey Company, Johnson & Johnson, Kraft Foods, PepsiCo, Procter & Gamble, and Wal-Mart Stores, as well as carriers, technology vendors, and research firms.
Such a gathering is long overdue. As a speaker from a well-known consumer goods manufacturer pointed out, an industry study published in 1993 set the goal of reducing the amount of inventory in the supply chain from 100 days to 50 days by the year 2000. Last year a study from the same source estimated that the inventory level was 98 days — a reduction of just two days in 17 years.
MIT CTL Research Affiliate Dr. Larry Lapide voiced similar concerns at the event. Wal-Mart introduced its Retail Link® sales data retrieval system back in 1992, yet upstream applications of such shared point-of-sale (POS) sources remain limited, he said. Lapide framed the central research questions for the conference by asking, “How do we use downstream data to improve upstream operations?”
Upstream IDS applications have moved at a snail’s pace for a number of reasons. Some retailers have embraced the technology by making POS data readily available to suppliers at no charge. But these enterprises tend to be the exception rather than the rule; some offer access to selected partners for a price, while others refuse to allow general release of their data to third-party syndicated firms. Sometimes restrictions are imposed by merchandisers and marketers who control data distribution but fail to appreciate the dramatic supply chain gains possible when data is shared with trading partners.
Operations folks are also culpable, particularly when they neglect to make a strong case for IDS projects even when there is hard evidence of the benefits. Internal departmental silos can prevent the data from being used effectively. Indeed, one of the most striking observations during the event was the importance of involving manufacturing, marketing and sales, and supply chain in IDS applications.
Still, the stars seem to be in alignment for fulfilling the technology’s promise. The Great Recession has put companies under intense pressure to review and streamline their cost structures. The collapse in consumer confidence is prompting many enterprises to question the value proposition of their offerings. At the same time, margins are being squeezed by reduced disposable incomes and commodity price inflation. On the positive side, recent developments in data storage and analysis mean that companies are able to process more demand data than ever before.
A number of leading organizations are taking the initiative. For example, a $40 billion-plus food company is shifting from a “plant-centered” to a “shelf-centered” supply chain mindset. In the past, POS-related projects were not actionable and the required business processes did not exist. The demand signals originating from purchase decisions at multiple stores were aggregated at various distribution centers. As a result, the replenishment data received by manufacturing plants was often three to 10 days old. Compounding these inaccuracies was the inability of retailers to measure on-shelf availability.
The company launched an IDS project to look at what infrastructure was required to address these issues. It needed to build a demand signal repository (DSR) to cover three primary operational areas: in-store forecasting, replenishment, and logistics efficiency.
As the project team delved deeper into the company’s supply chains, it came up with some “staggering” findings. For instance, it analyzed the inventory generated by one product line across 1500 stores for the second half of 2009. The stores and associated DCs stocked about 30 million pieces of product, but there were only 21 days in each year when more than one million pieces were sold. The result underlined how the loss of reliable and timely demand signals had created layers of excess safety stock.
These excesses were addressed by improving the flow of demand data. The manufacturer now receives POS data for the previous day’s sales at 2 a.m., and about four hours later has a forecast that is integrated into a vendor management inventory system. The improved information supports a number of important process changes. For example, the company worked with customers to identify high- and low-velocity items to make sure that fast-moving “A” category units receive the highest inventory coverage. The strategy represents a paradigm shift because it involves service rates for some items that were intolerable under the old inventory regime. The company achieved store-level inventory reductions of 45% to 70%, and cut working capital requirements throughout the system.
The latter result is particularly important. Addressing cash flow and working capital issues attracts interest from management outside of operations. Moreover, wins in these areas help to build momentum for IDS initiatives.
This was one of the many valuable insights that came out of the case studies presented at the conference. Another lesson learned was that demand data can include many different types of streams, such as DC withdrawal numbers and even data derived from retailer coupon programs. Making sure that the right people use the data is crucial, as is policing its quality. A major manufacturer found that the consistency of demand data is as important as its accuracy.
Looking ahead, a panel of industry analysts commended the companies for their IDS work, but suggested that no organization has yet to achieve truly revolutionary change. What one analyst called “perpetual pilot syndrome” still limits the degree to which meaningful change takes place. The analysts suggested that more radical applications of IDS technology are needed, such as harnessing pattern recognition technology to generate a more accurate picture of demand.
For more information on the Capturing Strategic Advantage from Integrated Demand Signals conference, contact Jim Rice, jrice@mit.edu .