Tom Andel

SKU proliferation and other knotty problems
February 26, 2008

SKU proliferation, especially in retail, is every material handler’s nightmare. With square footage and labor at a premium, the more variations on a product theme, the harder it is to keep inventories balanced.

One reader facing that challenge posed this question to Modern’s editors:

“Is there research on the relationship between warehouse inventory levels and the number of brands for a product?”

I went to one of my best information sources for questions on research, Tom Speh, the James Evans Rees Distinguished Professor of Distribution at the Farmer School of Business at Miami University. He told me that although inventory rises with the number of SKUs, there’s no simple formula for calculation. Hardware stores have a great solution to this when it comes to paint. They mix it in the store to the customer’s desired shade to cut back on the number of SKUs they need to keep in stock. 

“The reason for the increase in inventory is the increase in total safety stock that is required as you add SKUs,” Speh told me. I’ll let him finish the technical explanation:

“Each SKU has its own safety stock (SS) to protect against extreme demands and delivery glitches.  So, if we had white cake mix, our SS might be 20 units.  If we now have white and yellow cake mix, and assume demand is split between the flavors, then the SS for each SKU might be 12 units.  Hence total SS rises from 20-24.  This is the old ‘square root law’ applied to inventory.  The SS increases with the square root of the number SKUs.  Thus, if 1 SKU has 30 units of SS, then if we add 3 more SKUs, for a total of four, the four SKUs in total would have 30 x the square root of 4 (2), or 60 units of SS.”

The math gets a lot more complicated, but that is the gist.  The underlying concept is that with one SKU the highest level of demand might be 40 units per day, but if we add another SKU in the same line, the max demand for each would not be 20, but something higher like 28 for example, because there is not that much reduction in demand variation (it is not proportional – it can all be proved statistically).

SKU proliferation is a phenomenon of both consumer demand and retailer promotions. One strategy logistics managers use to deal with it is packaging postponement. With postponement, they can decide to delay product differentiation to forecast total product demand rather than on an SKU-by-SKU basis. Delaying differentiation enables suppliers to better respond to retailer customers’ seasonal and promotional demands.

Some packaging service providers offer postponement as a service. Their customers will feed their production schedules to them. That way, these customers can focus on basic factory planning and manufacturing of their core product. The packaging providers will do packaging material requirement planning, specification management, order the material and package the product, and send the product directly to distribution. This can improve customer service levels. Plus, if production planning periods are six to eight weeks out, it’s hard to synchronize packaging production with that. You may make material that won’t sell.

So, I hope this offers that reader some food for thought. If you have another take on dealing SKU proliferation, I welcome you to add it to this forum.

In fact, if you’ve read or written a research paper on it, or some other knotty material handling problem, tell us that too. You may even want to nominate it for a research paper award.

Take the “2008 Outstanding Material Handling & Facility Logistics Research Paper Award,” for example. Through this award, The College Industry Council on Material Handling Education (CICMHE) and the Material Handling Industry of America (MHIA) are interested in recognizing excellent material handling and facility logistics related research papers published in the last two calendar years (January 1, 2006 - December 31, 2007). Papers will be judged on the basis of the following criteria:

--Impact on the state-of-the-art of material handling

--The degree to which the research advances the state-of-knowledge

--Innovativeness of research methodology

--Novel application exhibited

To nominate a paper for this award, provide an electronic copy (PDF strongly preferred) of the paper that was published in 2006 or 2007 along with a one-page (maximum) summary of why the paper should be considered by April 30, 2008 to Mike Ogle (mogle@mhia.org) at MHIA. The summary should include specific statements of how the nominated paper meets one or more of the criteria specified above. The entire submission must be in English.

Posted by Tom Andel on February 26, 2008 | Comments (1)