The Evolution of Omni-Channel Logistics

Traditional retailers were fairly slow to wake up to the huge threat posed by Amazon. But my recent visit to “Retail’s BIG Show,” put on by the National Retail Federation (NRF), convinces me that they are now wide awake.

One way bricks and mortar retailers have responded to Amazon and other e-retailers is by turning their stores into more of an asset with omni-channel commerce. Omni-channel is based on providing consumers with a more seamless shopping experience by better leveraging both stores and the e-commerce channels. The idea is that a customer should be able to go to a store, see something they like, and order it for home delivery. Or order the item online and pick it up at a store. The same sorts of permutations apply to returns.

Retailers can use new tools to decide whether it makes more sense to deliver an online order from one of their e-commerce warehouses or directly from a store. It will cost more in labor to deliver from the store, but perhaps less in transportation. Further, if it becomes clear that a store has too much of a particular type of inventory, it makes sense to deliver from the store, even if it costs more, rather than pay for transportation to move the excess inventory to another place in the supply chain where it is needed.

Manhattan Associates can help retailers make these fulfillment location decisions based upon a distributed order management (DOM) solution. Manhattan’s DOM solution provides visibility to all network inventory – e-commerce and store –and then provides rules-based decision making on where to fulfill from. JDA addresses the same problem from an optimization perspective. The company told me at NRF that its solution makes these channel/facility fulfillment decisions by instantaneously and optimally trading off the costs of labor, transportation, and inventory.

But a prerequisite of successfully deciding whether an order should be shipped from a store or a DC is inventory accuracy. In a warehouse using a warehouse management system (WMS), a cycle count should show at least 99.9 percent inventory accuracy. A cycle count not only proves that a company has the inventory it expects to have, it also shows that the inventory is in the correct slots, and that the correct number of items is located in each slot. In short, it shows that inventory has neither been stolen or misplaced.

I also saw folks from Checkpoint Systems at the NRF show. The company sells RFID and anti-theft tags (electronic article surveillance). They told me that the highest level of inventory accuracy the store floor can hope to achieve is about 96 percent, and then only through the use of item-level RFID tags. I believe this claim. It is far more difficult to prevent theft at a store than in a DC. Further, the store inventory system may be showing inventory that is really sitting in some shopper’s cart prior to check out.

If you were trying to do in-store fulfillment, a store associate would find at least four percent of the time that the inventory is not on the store floor. That is four unhappy customers out of every 100, which is too high, especially when you consider an unhappy customer may choose to take their business elsewhere in the future.

In the age of Internet retailing and fickle shoppers, I believe we may be greatly overestimating the lifetime value of a customer. But whatever the true lifetime value of a customer is, poor delivery capabilities will adversely affect it.

But inventory accuracy is not the whole problem for store floor fulfillment of online orders. Inventory accuracy does not mean location accuracy. Shoppers pull stuff off the shelf and put it back helter skelter. Therefore, you could achieve a high degree of inventory accuracy and still be unable to fulfill orders in a timely manner.

Item-level RFID will at least permit the items to be found. RFID scanners can be put in Geiger counter mode where they beep louder as you get close to the SKU whose number you have entered into the reader. But item-level RFID is expensive. Right now the only product categories it makes sense for is high-margin apparel and maybe electronics.

There are also customer experience and store management cultural issues that provide arguments against doing online fulfillment on the store floor.

There is likely to be a fair bit of experimentation around how to best practice in-store fulfillment. There will probably not be a one size fits all solution. But I believe most retailers will discover that if they want to use store associates to fulfill Internet orders, they will need to do it in specially designed backrooms, not on the store floor!

(Note: JDA and Manhattan Associates are Logistics Viewpoints sponsors)


  1. Your article is missing two important factors in this process.

    First, a store inventory accuracy of 96% would not result in 4 out of 100 customers being disappointed. You would simply set your parameters so that any search for items in a store would not register for item counts below preset threshholds. For example, in this case you night set a threshhold of 5 items so any search would not indicate available inventory in this store if 5 or fewer items were indicated as in the store.

    Second, a very important part of making pick up in store work is a task management system to assign the necessary tasks to the right people at the right time to pick and pack the article for pick up at the promised time. Store labor is too expensive and there are too many other priorities to let this process happen haphazardly.

Leave a Reply

Your email address will not be published. Required fields are marked *