This is an RFM eqaution I am having trouble with. Theres no way I can attach the spreadsheet to post so maybe the procedure help you understand my dilemma .

Recency, Frequency, and Monetary Value RFM is a relatively easy and inexpensive technique that facilitates a better understanding of customer behavior and creates segment strategies based on the customers likely future actions. RFM is not a true data mining approach. It uses historical information from three data categories, and the user makes an assumption that past behavior is a good predictor of future behavior. It only includes revenue and excludes costs to serve the customer, including cost of goods sold. It is a first step in placing customers into categories, and since it is inexpensive and easy to execute (an Excel spreadsheet can be used), it is a prudent first step to understanding customer data. It may be used as a first step to identify best customer segments for specific marketing strategies. The outputs of this process can then be input to more sophisticated techniques. As the name suggests, it relies on three variables: recency, frequency, and monetary value. Recency is the date of the most recent customer transaction. Frequency is the number of customer transactions with the organization within a specific period of time. Monetary value is the amount spent within the same specific time period. We will use the florist example at the beginning of this chapter to demonstrate the RFM technique. The florist has approximately 5,000 customers in a customer database (Figure 7.1). Each has a unique numeric customer number that links back to the customers identification and detailed product purchase information. The time period measured will be the past 12 months. RFM is basically a sorting procedure. The file will first be sorted on recency, and the results will place all of the 5,000 customers in an ascending order sequence, with the most recent date being the first record. That sorted output will be segmented into quintiles. Each quintile will be assigned a three-digit identifier. The top 20 percent of customers based on recent purchases will be assigned a 5 as the first three-digit identifier. The second quintile will be assigned a 4, and so on, with the quintile representing the oldest set of customers based upon recent purchase assigned a 1. Each quintile will contain 1,000 customers. The middle quintile from the recency sort will have a cell identification of 3_ _, with the last two digits not yet assigned. Sorting these quintile cells on frequency and dividing the result into quintiles again will yield five groups, or cells, of 200 customers each for each of the recency quintiles. In ascending order of frequency, these cells will have their second digit assigned. The same procedure applies to all the cells, but for illustration of this and the next sort (monetary), we will use only the middle cell. The middle quintile from the frequency sort will contain 200 customers and will have a cell identification of 33_, with the last digit not yet assigned. Sorting this cell on monetary and dividing the result into quintiles again will yield five groups, or cells, of 40 customers each. In ascending order of monetary, these cells will have their final digit assigned. Running this process for all cells will generate a total of 500 cells. The cell containing the 40 most recent purchasers who purchased the most frequently and spent the most money within the last 12 months will be identified in cell #555. These are the most loyal customers. The least loyal customers will appear in cell #111. They are the 40 customers whose purchase dates are the oldest, who have the most infrequent amount of purchases, and who have spent the least amount of money. Every other customer falls somewhere in between.