Having designed a data warehouse the next step is to understand and make business decisions from your data warehouse. Business intelligence is nothing but analyzing your data. An example of business analytics is shown through the analysis of results from a product placed on sale at a discounted price, as commonly seen in any retail store. If a product is put on sale for a special discounted price, there is an expected outcome: increased sales volume. This is often the case, but whether or not it worked in the company's favor isn't obvious. That is where business analytics come into play. We can use Analysis Services 2005 to find out if the net effect of the special sale was to sell more product units. Suppose you are selling organic honey from genetically unaltered bees; you put the 8-ounce jars on special — two for one — and leave the 10- and 12-ounce jars at regular price. At the end of the special you can calculate the lift provided by the special sale — the difference in total sales between a week of sales with no special versus a week of sales with the special. How is it you could sell more 8-ounce jars on special that week, yet realize no lift? It's simple — the customers stopped buying your 10- and 12-ounce jars in favor of the two-for-one deal; and you didn't attract enough new business to cover the difference for a net increase in sales.
You can surface that information using Analysis Services 2005 by creating a Sales cube that has three dimensions: Product, Promotion, and Time. For the sake of simplicity, assume you have only three product sizes for the organic honey (8-ounce, 10-ounce, and 12-ounce) and two promotion states ("no promotion" and a "two-for-one promotion for the 8-ounce jars"). Further, assume the Time dimension contains different levels for Year, Month, Week, and Day. The cube itself contains two measures, "count of products sold" and the "sales amount." By analyzing the sales results each week across the three product sizes you could easily find out that there was an increase in the count of 8-ounce jars of honey sold, but perhaps the total sales across all sizes did not increase due to the promotion. By slicing on the Promotion dimension you would be able to confirm that there was a promotion during the week that caused an increase in number of 8-ounce jars sold. When looking at the comparison of total sales for that week (promotion week) to the earlier (non-promotion) weeks, lift or lack of lift is seen quite clearly. Business analytics are often easier described than implemented, however.