I’ve been interviewing supply chain executives at consumer goods manufacturers to learn how they are using downstream data. A key focus area for these executives is using this data to improve the accuracy of their demand forecasts.
A couple of the executives I interviewed raised an interesting point about forecasting that I don’t remember coming across before: the importance of having an unbiased forecast. Forecasts, by their very nature, are rarely completely accurate. Historically, most companies focus on improving the accuracy of their forecasts. But making sure you have unbiased forecasts is also important.
An unbiased forecasting process is one in which the demand planning team errs on the high side just as often as they err on the low side.
There can be subtle, and sometimes not so subtle, pressures to err in one direction or the other. Consumer goods companies often have customer supply chain teams embedded with account teams at a retail customers’ headquarters. These teams can be pressured to produce forecasts that err on the high side, particularly for promotions. The idea is that you don’t want to sell out and have stock outs.
This practice can be counterproductive with more sophisticated retailers—i.e., those that measure both out-of-stocks and sell-through. Sell-through is a metric that helps a retailer prevent getting stuck with a lot of inventory after a promotion is over. Here is how sell-through works: if a store forecast says that 100 units will be sold and only 90 are sold, then there was 90 percent sell-through for that promotion. The retailer ended up with ten more units of inventory than it needed for its short term needs.
Retailers that measure sell-through often also use a metric called GMROI. GMROI is a calculation that assists retail buyers in evaluating whether a sufficient gross margin is being earned by the products purchased, compared to the investment in inventory required, to generate the gross margin dollars they are looking for. In short, for retailers that focus on GMROI, an overly optimistic forecast by the consumer goods company reduces the profitability associated with the product. This makes it less likely that the retailer will want to continue selling those products.
When forecasting is embedded in the Sales and Operations Planning (S&OP) process, other pressures can arise. During the budgeting process, many CEOs set “stretch” goals for their business units—i.e., goals that will be difficult to attain. This will cause the business units to have to “stretch” – work hard – to attain these goals. If these goals are communicated to Wall Street, and the company is gradually falling behind on its financial targets, forecasters might feel pressure from unit directors to produce forecasts that show these financial targets will be met.
Conversely, when these unit directors need to forecast revenues for the coming year, they want the forecast to be “realistic” (that means low) so that they can beat their numbers. Demand planners can get sucked into this game as well.
In conclusion, accurate forecasting is great, but having an unbiased forecast is also important.
