Nintendo's Wii video game was the "hot" item for the 2007 holiday season. People camped out for hours and trolled stores and internet sites in an often futile effort to find the game, which was in very short supply. In an article posted on the Harvard Review website, Yossi Sheffi, director of the MIT Center for Transportation & Logistics (CTL), argues that the shortfall could have been avoided with better use of demand management tools, particularly point-of-sale data. Sheffi's article was partially a response to a Dec. 14 New York Times article by George Harrison, senior vice president for marketing for Nintendo America. Harrison stated that the shortage occurred because Nintendo has to plan its production five months ahead and projecting that far in the future is difficult.
While acknowledging that accurate forecasting is a tough task that "has become trickier over recent years," Sheffi says that research conducted by CTL shows that companies can do a much better job at optimally aligning supply with demand by employing tools that are readily available. He specifically cites point-of-sale data. "At a recent CTL Demand Management Interest Group retreat for senior executives, attendees echoed general misgivings about POS data: that it is cumbersome and misleading," write Sheffi. "At the same time, there was wide agreement with new CTL research that such data, used judiciously, can greatly improve the management of demand."
Yet POS data remains an underutilized resource in demand planning, Sheffi says, noting that in a June 2007 survey of supply chain professionals, only about one-third of the 146 respondents said that they formally use POS data in planning and forecasting. More than one-third said they do not collect this data at all.
Companies need to spend the time and resources required to better understand and deploy this data, Sheffi says. He outlines several ways to use this information:
• Companies can analyze POS data for an initial period of a product introduction and get an early indication of subsequent demand. A CTL research team did this for shoes, and was able to improve the demand forecasts for the mix of colors and styles that the manufacturer was selling.
• Another option is to test the product in a small group of outlets to sample demand before triggering a full roll out. Companies can delay full distribution until they have these early indicators, or split production into smaller runs to give more flexibility.
• They can also use postponement, a strategy that postpones the final configuration of a product in the manufacturing process until better demand information is available.
• A further option is to improve forecasting by analyzing the POS demand pattern for the last selling season, paying special regard to the volume of product that was sold at marked-down prices.
"In each case POS data is crucially important because the true demand picture will not be revealed by shipment numbers. Indeed, basing supply decisions solely on shipment history can lead to significant errors. Even if surplus merchandise is being sold at knock-down prices, the manufacturer may be blissfully ignorant of the problem if it has no data on store-level sales. In such situations, surpluses sold at a discount are self-fulfilling because companies keep repeating the supply mistakes of the past.
Enjoy curated articles directly to your inbox.