While reading a recent issue of Popular Science, I was intrigued by the article “The Santa Cruz Experiment: Can a City's Crime Be Predicted and Prevented?” (October, 2011). It’s fascinating and appears to hold great potential. However, what triggered a related thought to the challenge of accurate sales forecasting (i.e., your sales algorithm) was this: “One of the most common criticisms of predictive policing is that it will not tell police officers anything they don’t know already.”
When I talk to organizations about instituting a disciplined sales approach, and the need for good data gathering, resistance to software and information technology reflects a similar sentiment. Salespeople would rather operate on their own intuition than trust a computer model (or algorithm) to help them predict the close of a sale.
As the article points out, “An algorithm is a progressive series of calculations used to process and analyze large sets of data.” Said another way, it’s not foolproof in predicting an exact outcome; it merely points to the likelihood of something to occur based on past behavior (or available data). In my mind, that’s the Catch-22 in sales forecasts; if you can’t get people to enter the data, you won’t be able to get the algorithm to reflect a reasonable prediction.
Since I believe most customers have a defined purchase process, and most organizations can likely map a common sales cycle, I struggle to understand why it seems so difficult to build a more predictable forecast model. Rather than ‘gut probability’, often reflected as a percentage of likeliness to close (get the order), it seems to me that you should be able to weight factors against common data points (e.g., time lapse, stage of the sales cycle, past performance/seniority of the salesperson, etc.) and develop a more accurate sales forecast prediction.
My recent discussions with sales professionals, managers, and business owners and executives on this topic have given rise to the notion that the unpredictability that can’t be factored in is “people”. That applies to both sides of the sales equation; on the customer side and the vendor/supplier side. However, the Santa Cruz Experiment leads me to believe that even in high-touch sales situations, where long term relationships help drive business, some predictability could be factored in.
Read the article from Popular Science and let’s get some discussion going on this topic. Also, think in terms of those companies that don’t have the luxury of large software and information technology budgets. Can the average company develop and implement a predictive sales model?