Allocations are most effective when demand is the key driver of allocating inventory, this allows inventory to be optimized by location/SKU and for a projection of allocations into the future which can be used for workforce management and to identify exceptions before they occur. In some cases, demand is already forecast as part of the assortment planning or item planning process and in other cases new products may have no demand at all. Demand types allow for different approaches to creating demand and spreading it down to a Location/SKU level in combination with a sales curve, size curve and location curve.

Too busy to read the article, watch this video on how to set up and apply a location curve.

Demand types are predefined functions that create a forecast at a location/SKU level. There are currently three demand types to choose from. They can be selected from the drop down list on each allocation strategy.

Item plan demand type utilizes the forecast generated in item planning for a choice. Typically this is used for core items that have a lot of history and are closely planned and monitored over time for receipt planning purposes. The goal of utilizing the item plan demand type is to match the expected demand at a location/sku level with agreed forecasts from item planning.

Total sales demand type is typically used for a new item which may have been bought for a specific window of time with a specific quantity purchased. To easily generate a demand plan at the location/sku level, users can simply select total sales demand type and type either the total expected lifetime sales or just the purchased quantity into the demand column, this will then be spread to the location/sku level as the initial forecast of the choice.

ROS demand type stands for Rate of Sale and is defined as the average demand per location per week per choice (style / color), this can be entered in the demand column of the allocation strategy. This is another way of thinking about demand when a new product is introduced and may be selling for a defined period of time. Similar products can be used to derive an expected rate of sale, alternatively, the ROS defined in Assortment Planning can be used as an initial seed in the allocation strategy.

When defining an approach to allocating products, a mix of strategies could be used whether the choice in question has a lot of history or is a brand new product for which little is know. These strategies could also be defined by preference and familiarity with the existing products, for instance, many allocators are familiar with the performance of certain types of products and can intuitively set a ROS as a starting point.

Now that we have established the different demand types, the next step is to spread demand to the location/sku level. This is accomplished utilizing the size curve, sales curve and location curve. For more details on each of these models, see the following

Note that these models are easily defaulted and can be copied to multiple strategies.

The goal of setting demand type in the allocation strategy is to set the initial expectation of demand at a location/sku level. Once the choice in question begins selling in each location, it will automatically retrend the demand based on actuals to ensure the future demand aligns with actuals to optimize allocation of the remaining inventory units. This is important to understand as perfection in setting the initial forecast is not necessary.