New Order Flows Make Omni-channel Initiatives Tricky

To support omni-channel initiatives, retailers have had to implement a technology most commonly known as distributed order management (DOM). These omni-channel programs often support new order flows, for example order online and have the order shipped to the customer’s home from either an e-fulfillment warehouse or a store. But as I thought about some of these new flow paths, I began to have doubts that DOM, as I understood this technology, could really fully support all these order flows.

Traditionally DOM solutions are built on top of a common, real time view of inventory across all company stocking locations. Because of this common view, a DOM can say, in effect, “this store is closer to the customer than that store, and if the warehouse does not have the inventory, we will save money if we ship from the closest store to the customer.”

The problem is that real-time inventory data is not all that is necessary to fully operationalize the order online/ship from store flow path. What if a customer is ordering in Boston and the DOM shows inventory in stores in New York City and Charlotte? NYC is closest to Boston, so a traditional DOM would have the order dropped to the NYC store for fulfillment. But what if the Charlotte store is about to put the product on a steep markdown? In this case, the most profitable fulfillment source would be Charlotte. But as I understood DOM, DOMs do not contain merchandizing data. If that was true, how could retailers fully support these new order flow paths?

In talking to Brian Kinsella, Vice President of Order Management at Manhattan Associates, I learned that my understanding of DOM was outdated (Manhattan Associates is an ARC Advisory Group client). Their solution includes merchandizing and other forms of data that do help to operationalize newer omni-channel flow paths.

Brian walked me through their DOM’s data model. First of all, their solution has replenishment layers. So the e-fulfillment warehouse could be one type of replenishment inventory, front of the store stock (the presentation layer) could be another type, and back of store stock could be yet another layer. By segregating the inventory into layers the order flow paths can be operationalized in different ways. Perhaps, for example, you never want to sell the last two of a particular SKU in the front of the store – presentation layer stock – because it will adversely affect store sales of that item.

On the merchandising side, Brian tells me that many retailers have a 0-6 schema that reflects where a SKU is in it markdown cycle with one side of the scale representing full pricing and one side reflecting deep liquidation mark downs. By importing this data into the DOM, and understanding when a product will move from one score to a lower score, the DOM can make the decision to ship from a more distantly located store that is further along in the mark down cycle.

These merchandising flags can also be used to operationalize returns flow paths in new ways. “DOM does not direct a customer to return an item to a particular location (we assume they’ll just do whatever’s easier for them). However, if an item ends up where it does not belong, e.g.  a drop ship item ending up in a small footprint store, DOM will source the  next outbound customer order for that item from that ‘wrong’ location, thus continuing to clean up these undesirable inventory positions.”

How profitable a customer is should also affect flow paths. Brian told me that their DOM has come to encompass the full transaction ordering history that a customer has with a retailer. Thus, a customer service representative (CSR) can look at a customer’s profitability and ordering history and be empowered to expedite certain shipments for profitable customers.

Finally, Manhattan Order Management also includes labor availability – by day of week by store location. So if a retailer with 500 stores has 14 stores that could ship distressed inventory to the customer at the same price (they are in the same parcel zone so the cost and delivery times are the same), they can assign that order’s fulfillment to the store with the most unused labor capacity. Brian told me that initially some of their customer’s did not turn this functionality on, but learned the hard way it was important, and then did end up turning this capability on.

Last year ARC did a survey of retailers on their omni-channel initiatives (available to ARC clients only). There were 177 respondents. We asked “which of the following technologies do you believe you need, that you don’t have, to drive a more successful omni-channel initiative?” Of 15 technologies, DOM was the solution most often mentioned the most important technology they did not have. When you start to think about operationalizing new flow paths, the reason for DOM’s primacy is clear.