Case Study - Ben and Jerry's

1. If I were to design Ben & Jerry's data warehouse I would use several dimensions of information. The first dimension would consist of the company's products; ice cream, frozen yogurt or merchandise. The marketing department has to know which products are selling, if Ben & Jerry's didn't know that their T-shirts are selling out as soon as they hit the stores, then they wouldn't be able to take advantage of the opportunity to sell the shirts. The second dimension would consist of the different areas of sales; US, Canada, Mexico, or Europe. I am not sure if they sell their ice cream in Mexico, but with data collection they can find out if their ice cream would be a better seller in the hot climate, rather than pushing for greater distribution in Canada. The third dimension would consist of the "specifics"; where the sale was made, when the sale was made, and who purchased the product. This information can help in the design of the product to focus on the buyer; it can tailor flavors to seasons, and packaging to buyer who looks for the better-looking product. If Ben & Jerry's could know when a season was coming to an end in a specific area, then they could forecast the need or the decline in need and speed up, or slow down distribution to those areas. The focus of the information is that it needs to be useful, and almost any information is useful.

2. Ben & Jerry's would need a huge number of files. The files I would create would be Store, Product, and Employee files. The Store file would contain information about each store that sells Ben & Jerry's, it would tell me everything about every transaction with a Ben & Jerry's product. The primary key for the Store file would be the store number, or code. Some of the foreign keys would be; delivery dates, order dates, sales, and customer complaints. The Employee files primary key would be an employee number. The foreign keys for the Employee files would be; start date, wage, shift, employee address, paid time off, and absences. The Product file would have a primary key of product name, for example, Cherry Garcia. The foreign keys would be product type (ice cream/frozen yogurt), amount in stock, amount sold, ingredients, and comments. Comments sounds like a useless bit of information, but I think it could be one of the most useful. When it comes to a decline in ice cream sales a comment query could tell managers exactly what flavor