This is the fifth article of a six-part series about supply chain planning in online retail by Tommi Ylinen, our VP Product. The series will consider how to manage the unique challenges online retailers’ face, from assortment decisions to forecasting for different delivery methods.

<< Previous postNext post >>

It was only a decade ago that many senior retail executives were still looking at the world of e-commerce and saying, more or less; ‘it’s not for us’. How the world has changed since then! With the rise of e-commerce many traditional retailers have established an online channel to complement their core businesses. Since the whole e-commerce concept is still relatively new, and volumes are often quite small but can grow rapidly, the supply chain strategies chosen in response vary a lot. In this post I will examine how to optimise inventory levels while maximising sales across all channels.

Some companies have chosen to operate separate warehouses for online and for store channels. The reasons behind this decision are often a reflection of physical constraints: i.e. the warehouse currently responsible for delivering to stores can’t accommodate more stock, or the picking process (single items for online vs. cases for stores) differs too much between the two models. In such cases it is quite simple to just calculate forecasts separately for both warehouses and replenish them separately. However this is far from ideal from a stock-holding point of view. As you have to buffer against uncertainty in both warehouses you are very likely to end up having significantly more stock in the business than you would with a single-stock-pool-scenario. And still there is a risk – especially if your forecasting process isn’t very precise – that one channel sells better than expected and the other not so well, with the result that you end up with a stock imbalance: no stock for online and too much for stores (or vice versa). This can result in a combination of lost sales and mark-downs or in costly shipments between warehouses.

So, if possible, it generally works best to have only one pool of stock serving all channels. This way you can minimise your stock holding, ensure that stock is equally distributed to all channels, and that it is all sold at the best possible price. It still makes sense to calculate forecasts separately for online and for each store, as that produces the most accurate results. For instance you can then include events and promotions at individual store level and do the same for your online outlet as, typically, sales patterns vary across channels– for instance Christmas sales peak earlier online than in stores. The best warehouse replenishment forecast can then be built by adding the online forecast to the combined store-order forecasts – the latter can be calculated using store demand forecasts together with information on store stock levels, delivery cycles, lead times and so forth.

The question that remains is how do you ensure that all channels get their fair share of stock, especially in situations where you don’t have enough? In my experience the best solution here is to create a virtual ring-fence around part of the stock earmarked for online sales. Your stock can still be held in a single storage location in your warehouse, but part of it is allocated to online channels and the system treats it as reserved, so stores can’t draw upon it. It’s easy and efficient to send more stock out to stores as the season progresses, but recalling goods from the stores to the DC is always expensive and best avoided whenever possible.

This approach to allocation is best done if based on channel-level forecasts, which should be updated continuously. The starting point for online inventory before the season might even be quite crude – e.g. “forecast + x%” or a fixed allocation – as you don’t yet know how the item will sell in each channel – but as you start gathering actual sales data, the forecast for the season will be updated and so will the split between channels. As you then near the end of season you can slowly ramp down the ring-fenced amount in order to minimise end-of-season stock.

The ring-fencing approach also enables you to use different strategies for different situations – you can, for example, prioritise channels differently depending on the available supply; some companies might want online always to have 100% availability even if that would leave stores out of stock. To manage the process well throughout the season, a good exception management process is needed to ensure that all changes in demand lead to the actions necessary to meet the business’s goals.

In my next post I will discuss what implications and opportunities ‘click & collect’ has for supply chain planning.

Is your supply chain optimised for online?

Our ‘Supply Chain Planning in Online Retail’ ebook covers five steps to improve inventory management in your e-commerce operations. Read it to find out more.




Read more

Tommi Ylinen

Tommi Ylinen

VP, Product


Phone+358 50 433 5975

Tommi Ylinen, MSc. (Tech) launched and directed RELEX’s operations in the UK before joining the Helsinki office as VP, Product.

Subscribe to receive monthly blog updates

Stay on top of the latest trends in supply chain

Subscribe to receive monthly blog updates.