Retail marketing through supermarkets is an intensely competitive business, based on large volumes and thin margins. If we take instant coffee as an example, it has attributes of
|Weight - 200g jar|
|Taste or Quality - powder, freeze dried, granulated|
|Brand - Maxwell House, Nestles|
A distributor anxiously watches the small blips in sales from week to week, eager to ascertain any change in consumer behaviour. The distributor wants to know how their brand is faring, and how each particular product is faring against its competition.
A productmap can be used to link a product and its attributes. Then specifying a brand will result in a set of products, while specifying a product will result in a set of attributes. Competing products can be found by specifying the product, remembering the attributes (using a REMEMBER function), then using those attributes to produce a set of products.
The retail marketing model will contain at least two productmaps - one to represent products, the other to represent stores and their attributes. Brands could also be treated as objects rather than just names - they have attributes of value, market dominance, mindshare.
The constrainable sets of products and stores coming from the productmaps are used to extract sales information held in HISTOGRAMs.