Engaging the sales force in forecasting sounds like a good idea, doesn't it?
Compared to everyone else in the organization, don't sales people have the closest contact with our customers? Therefore, shouldn't they know better than anyone else our customers' future behavior?
There are at least three problems with assuming a priori that engaging the sales force will improve forecasting:
- Do sales people really know their customers' future behavior?
- Do sales people have any motivation to give an honest forecast?
- Does improving customer level forecasts even matter?
For sake of argument, lets assume affirmative answers to 1 and 2 -- that the sales force has knowledge of their customer's future behavior, and provides an honest forecast of that behavior. Can better customer level forecasts help us?
For maintaining appropriate inventory and customer service (order fill) levels, we want a good forecast by Item / Location (where location is the point of distribution, e.g. a Distribution Center (DC)). As long as we have the right inventory by Item / DC, we don't have to care what individual customers are demanding.
If volume for an Item through the DC is dominated by one customer (or a small number of customers), then it could be helpful to have more accurate Item / Customer forecasts. Improving the forecasts for these dominant customers would likely improve the forecast that matters -- the Item / DC forecast.
On the other hand, suppose the DC fills orders for dozens or hundreds or thousands of customers, none of which is more than a small percentage of total demand at that DC. In this situation, positive and negative errors in the Item / Customer forecasts will tend to cancel each other out when aggregated to Item / DC level. So even if you can improve Item / Customer Forecasts, this is unlikely to make much improvement in the forecast that matters -- the Item / DC forecast.
[Note that some organizations utilize Customer level forecasts for account planning, setting quotas for sales people, etc. So there may be other reasons you want to do Customer level forecasting. Just realize that improving the Item / DC forecast may not be one of them.]
Efficiently Gathering Sales Force Input
If we are going to engage the sales force in forecasting, we ought to at least do this efficiently. Time spent forecasting is time taken away from building relations with customers -- and selling.
One way to gather input was suggested by Stefan de Kok of ToolsGroup:
...there is huge value in getting input from humans, sales reps included. That input however should be market intelligence, not adjustments to quantities. For example, let the sales rep input that their account is running a promotion and then let the system determine what the quantity impact is. Not only will the uplift become more accurate quickly, but also the baseline will improve. Ultimately it becomes a lower effort (but not zero) for the sales people and their forecasts become much more reliable. (source: LinkedIn discussion group.)
The idea is to minimize the time and effort from the sales force, requiring they provide information (promotional plans, new stores (or store closings), more (or less) shelf space, etc.) that can be put into the statistical forecasting models. But not requiring them to come up with specific numerical forecasts.
As always, the value added by these efforts needs to be measured -- are they making the forecast more accurate? If so, and your software can take advantage of the inputs, this is one approach.
Another approach is to provide the sales force with Item / Customer forecasts (generated by your forecasting software). Then have them make overrides when they feel they know something that wasn't already incorporated into the statistical forecasts.
This approach can be wildly ineffective and inefficient, when sales people are overriding every forecast, and not making them better. (Improvement (or not) is easily measured by FVA.)
The key is to train the sales people to only make changes when there is really good reason to (and otherwise, to just leave the statistical forecast intact). Eric Wilson of Tempur-Sealy achieved this by appealing to the competitive nature of sales people, urging them to "beat the nerd in the corner" and only make changes that they are certain will improve the nerd's statistical forecast.
Revenge of the Nerds?
There may be good reasons to engage the sales force in forecasting, we just can't assume this is always the case. When there are good reasons, focus on the efficiency and effectiveness of the inputs -- minimizing the amount of effort required to provide inputs, and measuring FVA of the results.
