Your focus will vary depending on how your business views inventory placement.

While these terms are often used interchangeably, Logic’s perspective is they are two tightly linked but separate processes which work together to solve the same complex puzzle: How do I meet my customer’s expectations?

Forecasting is the use of historic data to predict future outcomes. Demand Forecasting is a theoretical, statistical and almost academic activity where you have infinite supply and product travels across time and space unrestricted. A forecast attempts to quantify what is possible if everything went perfectly. From there, the Supply Chain solutions work to optimally achieve that ideal picture by determining the timing and quantity to supply. Our viewpoint is your focus will vary depending on how your business views inventory placement. If your customers expect a fully stocked shelf at the location in which they shop, and they’re not willing to wait for it or make a trade-off, an accurate demand plan is more critical than an accurate sales forecast.

To determine if a sales forecasting or demand planning viewpoint is better for your business, consider the following:

How do your customers shop?

Do your customers come into the store, knowing exactly what they need or want, and expect it to be on the shelf week after week? Or do your customers browse, looking for something new, but are willing to make a trade-off on colors or sizes? Do your customers use social platforms to discover new products, and place mobile orders from a variety of retailers? Or do they trust your assortment is the right mix for them, and want you to pre-emptively fulfill their shopping list?

If your customers expect inventory availability for an exact product, a demand planning approach will help you continuously purchase the appropriate amount of goods to support fulfillment. In this case, inventory accuracy is of paramount importance, and the source from which you forecast should change accordingly. With a demand planning viewpoint, we recommend the use of warehouse transfers as a forecasting input to bulk purchases, as what you are really servicing is supply, not sales. At the supply destination, i.e. the store, the sales forecast is used to predict sell-through rates and thereby reduce overstock potential.

 Is it expensive to be wrong?

As we move to a single inventory pool for ecommerce and brick & mortar stores, forecasting where is gaining traction over forecasting what. The cost of reverse logistics is going up, and we’re all working to predict how a customer wants to buy, not just what she wants to buy. When it’s expensive to be wrong, item / location-level sales forecasting is warranted. This enables the prediction of where a product is likely to sell, so that inventory can be placed accordingly. But when inventory is expensive and bulky; such as treadmills, appliances or boats; demand for the item is likely lumpy, and the forecasts are essentially useless. Rather than working to meet sales, you are working to control inventory. If you must own the inventory (rather than order a shipment from the vendor) consider holding the goods in the warehouse, and triggering the system to ship one only when stock hits zero. For such items, a customer is likely to prefer home-delivery, so why have more than display stock on hand? This also helps with an extended aisle strategy, where the sales floor can carry a few models, but associates can place orders online for customers looking for a model with different features, or different colors.

Do you have statisticians, or people who can learn statistics, on staff?

Most of us are too biased to make the right decisions about how our business is actually going to behave, especially in verticals where products are intended to provide an emotional response. If your team lacks statistical experience, rather than focus on forecasting, our experience has shown pairing a statistical forecast with automated replenishment reduces on the incidence of stock outs by focusing on supply. Let your in-house resources apply the market intelligence to improve the forecast, but in most cases, the focus should be inventory planning, not sales forecasting. Often-times the forecast itself can be owned by an external partner, freeing up your team to solve business problems instead of system problems.

What is your sales velocity?

Forecasting engines simply work better when they have access to more data. Low velocity items are notoriously difficult to forecast, and are often better managed through stock settings than forecasts themselves. To meet your customers’ expectations for low velocity items, you need to watch service levels, not forecast accuracy. When it comes to building financial plans, however, Logic advocates for the use of forecasting engines to provide a realistic trend line from which to baseline open-to-buy forecasts. Forecasting at merchandising level – such as department or division – provides opportunity to apply lifts at an aggregate that can trickle through the category. This is especially useful when building promotional forecasts.

How good is your data?

It’s cliché, but we’re going to say it: Garbage in = Garbage out. It’s critical to leverage analytics and data-driven processes to make disciplined, process-driven merchandising decisions, and to do that, you need clean and consistent data. We are pleased to say, gone are the days where data cleansing is a highly manual, labor intensive task. Today’s machine learning algorithms can quickly scan massive data sets to determine, and automatically remove, anomalies. As the algorithm gets to know your business better, it can even add in trends and factors to provide an even stronger starting point for your forecasting engine.

In summary, remember it all starts with knowing your customer, and who knows them better than you? First define a reliable picture of what is possible leveraging as much information as possible, and next determine how you will use that information to better service your customer. Whether you call it demand forecasting or demand planning doesn’t really matter if your customers get the experience they are looking for.

Tagged with: , ,