How often have you heard a sales executive commit to a top-line revenue number by handicapping the forecast? It happens in executive meetings around the world. As a quick refresher, handicapping the forecast is usually done by weighting the pipeline revenue number. Weighting the forecast is the simple process of multiplying the opportunity “probability to close” by the opportunity revenue and aggregating that “weighted revenue” across all opportunities—and “Abracadabra!” you have your final revenue number.
In a large company, it is quite impressive when you can rely on such a simple process to forecast a revenue number 3 months in advance. In my software career, from Siebel to Right90, this simple method has been incredibly accurate at calling the top line revenue number. This isn’t anything new, and many folks have espoused this method.
So, sales forecasting is easy. Call it done, right?
Should I build 55% of the forecasted units?
WRONG! Is the sales forecast used only for calling the top line revenue number? No. It is used across the organization, including by operations teams to determine how many units of which product they need to build.
Example: If Customer A is forecasted to buy 1000 widgets and the probability of that opportunity is 55%, does that mean the factory should build 550 units?
The customer will either buy 1,000 units or 0 units. Building 550 is the worst of both worlds: if you win the business, you don’t have enough to fulfill the order (and therefore lower customer satisfaction or lose the business all together). However, if you don’t get the order you get stuck with 550 units of left over inventory for that product. In this case, a weighted forecast doesn’t work. Instead, a raw, bottoms-up forecast is a better method.
Denis Pombriant, of Beagle Research, explains this issue in great detail in a recent white paper. Let’s take another example of where weighted forecasting is not enough.
Existing business vs new business
For many companies, 80% of revenue comes from ‘run-rate’ or ‘existing’ account business. It goes to reason for these companies, forecasting existing business is more important than forecasting ‘new business.’ In the case of existing business, there is no probability.
Example: We all know HP buys chips from Intel. Do you think there is an “opportunity” for HP computers in Intel’s CRM system with a probability associated with it? NO – Intel won that business a long time ago, but Intel still has to forecast the units of each specific chip that HP demands each month. In this case, Intel needs a raw forecast of exactly how many units of which specific chip that HP demands in which month (or week).
This requires a bottoms-up forecasting approach by units, by account and product managers familiar with the HP and Intel relationship. After capturing the raw forecast, Intel likely leverages a statistical forecast and management judgment in arriving at the final unit judged forecast. This is another example of where the weighted forecast method is not sufficient.
There are many other cases where a weighted forecast does not provide all that you need. However, as I stated at the beginning, weighted forecasting is a valuable PART of a sales forecast. This is why we recently added weighted forecasting functionality to the Right90 sales forecasting application, enabling companies to create a complete sales forecast. A complete sales forecast in best-in-class companies involves some combination of
- a bottoms-up raw forecast from direct sales folks
- a bottoms-up raw forecast from distributors or partners (the channel)
- A product manager’s forecast
- A statistical forecast
- A calculated forecast (based on probability weighting or some sort of historical measure)
- Management vetting (scrub or review) of the raw forecast
- Executive judgment
Sales forecasting is part art and part science; best-in-class companies use all information available to them to arrive at a trusted and actionable sales forecast.
