I’ve always been a fan of Procter & Gamble CEO A.G. Lafley’s two moments of truth concept. The first moment of truth is what a consumer sees on the store shelf; the second is what the consumer experiences after they have bought the product. Today, I want to focus on the first moment of truth, on what consumers see on the store shelf, assuming there’s even a product there for them to see. It strikes me that we have gone through three phases in trying to solve the “out of stock” problem.
In the first phase, we mostly assumed that if a product was out of stock, it was the supplier’s fault-i.e., if only the manufacturer could forecast better, the problem would largely disappear. Or we thought, if only our supply chain was more nimble, if we did not have such long manufacturing runs and planning cycles, we could do a better job of replenishment.
In the second phase, many manufacturers came to the realization that they could not forecast accurately without more timely and granular data from retailers. Companies like WalMart and Target custom built what are now called Demand Signal Repositories (DSRs) to provide suppliers with near-real time granular data, such as POS sales by SKU by store, that made their historical forecast data less lumpy and less prone to the bullwhip effect.
This phase is still taking place. However, now instead of the retailer spending tens or even hundreds of millions of dollars to build these DSRs, companies like Retail Solutions (RSi) build the DSR for retailers like CVS and Walgreens for “free”. Suppliers who want access to the retailer’s sales and inventory data then pay RSI for it. There seems to be a tradeoff between the proprietary DSRs offered by WalMart and Target, and the off-the-shelf solutions offered by the RSI retailers. WalMart and Target make this data available to their suppliers for “free,” but they also have very high expectations regarding replenishment performance, and they are more likely to apply charge backs for “poor” performance. In contrast, Walgreens appears to have less demanding replenishment performance expectations, at least for now, but their suppliers have to buy the data.
In the third phase, we came to realize just how inaccurate in-store inventory systems are. At many retailers, in-store inventory accuracy is only about 65 percent. We also came to realize that poor in-store execution is a huge problem in achieving on-shelf availability, particularly for promotional items. Very often the inventory is present in the store, but not stocked on the shelves or end-of-aisle displays. RFID was supposed to solve this problem, but it hasn’t for a variety of reasons. It has, however, made this problem more visible.
The next phase, which is just beginning at some leading manufacturers, will include a better link between merchandising and supply chain operations. That phase will be the subject of tomorrow’s posting.