Note that Frequency divided by Tenure equates to Orders per Year. Monetary Value divided by Frequency equates to Average Order Value. And Latency minus Recency equates to Expectancy. Furthermore, ((Frequency / Tenure) x (Monetary Value / Frequency)) is mathematically the same as (Monetary Value / Tenure) which equates to Sales Revenue per Year. Therefore, the algorithm is simply calculating the customer's Sales Revenue per Year divided by the time until they can be expected to order again. The actual algorithm rounds both Tenure and Expectancy to the nearest six week period. This equalizes the time factor in them and provides a workable forecast of sales activity for the upcoming six week period.R·F·M = ((Frequency / Tenure) x (Monetary Value / Frequency)) / abs(Latency - Recency)