Supply Chain Planning in Online Retail

Although the basic rules of demand forecasting and inventory optimisation work independently of the particular channel in question, there are indeed a number of differences with traditional retail.

Let’s first consider a purely online retailer, delivering orders from its warehouse against online sales. From a forecasting and ordering point of view this is straightforward. There is only one channel to forecast, and all inventory is reserved for that channel. However, the cool thing about online is that you tend to register more data as a matter of course, as all shopping happens digitally through computers or mobile devices. For instance, you could support forecasting or assortment management by introducing page-view information from the website into the process. Pricing is another area that is even more critical to online than to bricks-and-mortar retail: if your competitor sells the same product at a lower price, your customer will probably buy from them. Having this data in your forecasting system could prove useful.

Retailers with both an online and a high street presence face a slightly more complex situation. For them to manage their supply chains accurately channel-level forecasting is key – after all, different channels have different sales patterns. For instance the Christmas peak typically arrives a bit earlier for online sales than for store sales. Some retailers have separate warehouses for digital and physical stores, which makes stock management simpler, but inevitably increases the total stock holding in the business as items are double-stocked. It could also be the case that a product runs out of stock in one warehouse but is still piled high and overstocked in another. From a supply chain planning point of view the better option is to have a single pool of stock. That then raises the question of how to manage availability between channels. In this case the answer is virtual stock ring-fencing.

Many retailers offer both home delivery and click & collect. Often it makes sense to forecast these delivery methods separately as different channels – for instance home delivery orders can be fulfilled from a separate ecommerce DC whereas C&C orders could be delivered from the store stock. In this case good analytics is vital so that the business gets a proper overview of the split between the delivery models, both at item, and at store level. Sometimes this information can be even acted upon and combined with assortment management; if some items are routinely ordered online and then collected from a particular store, it might make sense to add those items to the store’s range – preferably automatically.

Read more about solutions for e-commerce and omni-channel retailing.