New product introductions with Oracle Demand Forecasting 1/3
This is my first blog post and I decided to start with one of the most asked features of Oracle Retail’s Demand Forecasting (RDF) package: Product introductions. It’s going to be covered in three posts and this is the first one.
Product introductions is a very common and frequent process for retailers. There are many variations of it, e.g. there are seasonal products which could be introduced every year (Santa Claus dolls for Christmas), variations of existing products (new taste coke) or completely new products (new mp3 player). How do big retailers manage the amount of work that needs to be done for new product introductions? How do they know how many new products to order in advance if they have no or limited historical data?
RDF is used to support this process. Before I start explaining how one can use RDF to automate some aspects of the process, let’s look closely at what parameters are usually considered by the retailer when introducing new products:-
- sales and gross margin on sales;
- store penetration, i.e. how many stores are going to be selling this new product;
- shopper penetration, that is market share in the target group;
- cannibalization and impact on the sales of the whole category, e.g. are the sales of classic Coke going to be affected by this ‘new taste’ Coke? If yes, how much?
- repeat purchases
RDF allows retailers to generate sales volume forecasts for new products by using a process called Cloning. Cloning allows users to generate forecasts for products by copying or cloning history from other products. Users can map items that have similar business cases, clone their historical data and begin generating forecasts. The products that have been cloned are called Parent Items. Usually, users can pick up to three Parent Items and select for each one of them a scaling factor. This factor will adjust the volume of the historical demand of the Parent Items according to the newly introduced item.
Cloning provides the ability to generate forecasts based on not only the historical data but also the promotional calendar, too. I.e. Sometimes the product that is introduced is set also on promotion and thus making the forecast process more difficult. In this case RDF is designed to look in the historical and promotional data that already exist and figure out the increase of the normal sales that the same promotion has on the Item. Then it will apply this increase proportionally to the upcoming promotion.
Vanilla cloning as it is described in this post applies to most product introductions. However, more complicated business requirements can be raised when discussing new product introductions. RDF can handle more complex functionality, too. Such topics will be discussed in the next blog post.