Are you still forecasting sales the way you have been for years? Are your forecasts any more accurate than they ever have been? Does anyone take sales forecasts seriously? Perhaps it’s past time to consider a new method for forecasting sales … that actually works. Forecasting is done effectively in other parts of business, and that was not always the case. What can you learn about forecasting from other areas of business that will help you create better sales forecasts? Not surprisingly, a lot.
John was recently promoted to assistant plant manager. As part of his new job, he was recently provided with the production demand requirements for the next quarter. On first pass, the lead times appear reasonable, and John is confident of this factory team’s ability to produce quality products. Prior to his promotion, the factory floor had adopted a number of the key “best practice” manufacturing methods. The team understands lean thinking. They have flexible work cells and support a “one-piece flow” method of manufacturing. Average cycle time for work-in-process (WIP) for each work center is understood and tracked. Total lead-time for product production is also known. The team understands how to work through manufacturing bottlenecks and how to subordinate to the key bottleneck to increase throughput or balance the line. And, the company has the appropriate technology in place to meet the needs of the production process.
In addition to all of this, the Quality department has been actively supporting a “zero-defects” program for years. Continuous improvement is a way of life. In fact, John’s pretty sure that Dr. Deming himself trained the predecessor plant manager. Documented and audited production plans exist for all products in production. This, of course, includes the requisite inspection plans. All products released to production also include process capability studies including upper and lower control limits prior to their release to production.
So what is John’s problem given the near “heavenly” environment that he finds himself in? Well, John has never been held accountable for forecasting production output. While he is confident in the factory’s ability, this is the first time he has to present a shipment forecast to the senior management team, and he wants (in fact he needs it) to be right. In thinking about how to prepare his forecast, John considers what he has seen in the company before. He considers who else “forecasts” regularly to senior management. Well, actually all of the department managers forecast. Sales management forecasts sales, finance forecasts cash flow, engineering forecasts new product completion, human resources forecasts headcount. But everyone seems to use a different method. Upon consideration, John realizes that the sales manager forecasts most often and has been successful at staying with the company for many years, so John decides that he should consider emulating the sales manager’s forecasting method.
The next day John goes out onto the factory floor and asks each member of the production team what his or her production status is. How many units do they have in process at their work-center. He asks each production worker what they think the probability is (in percentage terms) of completing their units in the relevant timeframe. Based on the collection of that individual information, John can then build a factory output prediction based on the weighted-average of each of their probabilities.
This seems like a really good idea and Sales uses it all the time. But when John meets with his boss, the plant manager, Steve, and explains his ideas on how to forecast, Steve is in shock. Where on earth did John come up with such a forecasting method and why did he think it could possibly work. When John explains he copied it from the Sales Manager, Steve bursts out laughing. Steve explains that Sales can never accurately forecast anything, and their methods prevent them from ever being able to do so. But since no one expects Sales to get better at forecasting, they just adjust their forecast. But, John explains to Steve, one of the things he “learned” from the Sales Manager was that while he was interviewing his people to get their forecasts of production, he could also “motivate” them to try to do “better.” In fact, the Sales Manager claimed to have great success at getting his people to commit to increased “production” during those “motivation” sessions. John explained that maybe he could get the factory to do even better by emulating that management method. Steve just shook his head. Steve put his arm around John’s shoulder and began to help him understand what he was missing.
Steve explained that the culture in Sales Management has traditionally operated with a “strange” viewpoint on how to manage their process. But, since everyone seems to do it that way, nobody has ever questioned whether it makes any sense. Steve suggested to John that if the Sales Manager were to emulate the plant’s forecasting and management methods, they might have more accurate forecasts, and even more importantly, increased sales and operating efficiencies.
Let’s see how that might be true. First, we must be aware of a management truth:
You cannot accurately measure what you don’t understand. And you cannot improve what you cannot measure.
Your “friend” in sales management uses an “art” based process that has very low probability of predicting the correct outcome. Because sales management does not generally understand the process they cannot measure what their people are doing, and therefore they can’t improve their process. Several years ago the practice of sales management evolved a tool called pipeline management or funnel management as a way to try to improve forecast accuracy. By categorizing each “step” in the sales process into a stage, management felt they could better assign a probability to the prospect becoming a sale. These “improved” probabilities are “rolled-up” into a global forecast which sales management hopes will be reasonably accurate based on the law of large numbers. Even when it is somewhat accurate, there is no way to improve.
Shifting From Astrology to Astronomy
Sales management is always under pressure to produce better forecasts and more sales. Steve explained to John that once Sales hit upon the funnel system, he was always surprised that they did not see the analogy to plant floor work-cells. Instead of using “scientific” methods, sales management appears to continue to use the “astrology” theory of forecasting. As Steve explained, “Just as looking at the stars and attempting to predict a specific future doesn’t work reliably, neither do the sales forecasting methods now in general use.”
If instead, sales were viewed as a flow, then lean thinking methods could be employed, intrinsic patterns could be identified, and cause and effect relationships understood. Thus, sales could turn to “astronomy” (science based) rather than astrology to predict and improve sales. It is even likely that sales management could begin to use cause and effect management methods. What does Steve mean by “cause and effect” management methods?
Once you have a flow description of your sales process, you can experiment by changing an activity in your selling process and monitoring the effects through feedback to determine how the change effects the outcome (sales). Anticipating what the Sales Manager would say about “too many variables” to accurately identify cause and effect, Steve reminds John of how not that long ago, factory managers believed the same things about their complex processes. What production people learned is that the process works and even more importantly, when the output changes, the flow model allows them to identify what was the intrinsic cause. The same is true in a sales flow model.
If asked, Steve is convinced that he could demonstrate to an open-minded Sales Manager that customer buying process patterns exist and will result in natural sub-processes that can become useful feedback points in the flow model. Steve suggests to John that you could model the sales process in a similar way to the methods used to model the product floor with work-cells, work-in-process, latencies, yields, etc. Further, the same management concepts such as constraint theory, lean thinking, and continuous improvement methods that had been applied over the years to dramatically improve their product manufacturing could probably be applied to “manufacturing” customers . . . if someone just thought about it that way.
Steve points out to John that if the Sales Manager used a system similar to the one John had available, Sales could view their so-called sales reports in a whole new way. These new sales management reports could provide useful yield information from stage to stage as well as other in-process information such as throughput rate. Using these new process reports based on system yield and sales process rates, Sales could provide much more accurate forecasts than the ones they currently provide to management using their probability weighting system. Further, sales management would be able to identify where to focus to improve sales performance rather than running month-end and quarter-end sales incentive programs.
Which led Steve back around to asking John why on earth he thought he needed to go outside the data reporting system from the plant floor to get an accurate production forecast?