A Supply Chain Problem I’d Like to Have

I recently wrote a piece called “The Reset Economy” where I argued that for various macroeconomic reasons demand may not bounce all the way back after the recession ends.  Consequently, there will be too much capacity in many industries.  I believe this is the biggest supply chain problem many companies will confront over the next few years. 

However, in certain industries, demand will always be hard to meet for certain “hot” products, a problem many companies would like to have in this economic climate.  Nevertheless, for the industries where this happens, significant money is left on the table. 

Oracle briefed us recently on the new and enhanced capabilities of their Value Chain Planning suite, and they introduced me to a type of constrained supply scenario I had not come across before.  Here are some of the industries where supply constraints occur, and how these industries typically respond:

• Semiconductor and high tech components: Suppliers typically respond to cyclical surges in demand by putting their customers on allocation. Every customer gets some portion of the products they want, but nobody gets everything. More powerful or more profitable customers get more of what they want than other customers.
• Fashion retail: They also allocate how much of a hot product goes to the various stores in their chain.
• Seasonal demand industries, such as fertilizer in the spring, back-to-school supplies in the fall, and turkeys for Thanksgiving.  In many of these industries, suppliers build ahead to meet the seasonal demand. Factories may run for months in the offseason, the goods are warehoused and then sold in a very short time period. If you work in these industries, “sense and respond” and “demand-driven” supply chain management has always sounded ridiculous. In many cases, the goods have already flowed to the stores by the time the selling season hits. If the forecast was wrong, and the store sells out quickly, there is only a limited amount of time available to send more goods to the store, or to shift product from stores with excess inventory to those that have sold out.

And finally, here is the new variant Oracle introduced me to. 

• Movie and video game retail chains: When a blockbuster movie comes out, the selling season is short. At many retail outlets the main constraint is their facings (floor-level shelving devoted to a particular product). Retailers are unwilling to have too high a proportion of their facings devoted to a single movie. Consequently, they sell fewer videos than they could have. Their historical demand then shows what was sold, not what could have been sold. This leads to future forecasts for blockbusters that are too conservative. One method for dealing with this is to conduct a certain number of demand tests. Certain stores are told to use many more facings than their policy would normally allow. These demand tests can help determine what the true demand could have been and how much money the chain is leaving on the table. This gets the retail chains thinking about new policies and methods to respond to blockbuster demand.

For retail chains that have these types of constraints, or consumer goods manufacturers selling to those chains, the ability to forecast at different levels, including very granular zip code forecasts, is very important.  At a higher level of aggregation, a national or DC level forecast, you will have better forecast accuracy.  However, if you can get down to zip code forecasts, or other more granular groupings of stores, better replenishment and allocation decisions can be made.  For example, below is a screen shot Oracle provided me from a recent briefing.

What this slide shows is how well certain genres of movies sell in certain zip codes.  Horror sells particularly well in zip code 60018, which is Rosemont, Illinois.  If you are visiting Rosemont, think twice about going outside on nights when there is a full moon!