Does your sales management job feel like an obstacle course? Do you have an endless list of tasks that demand your attention but do nothing to generate sales or to build customer relationships? Sales forecasting is one area where you may be able to save time, and actually deliver better forecasts.
Try following these three steps :
- Spend your time and energy on the forecasts where you have market intelligence that makes a real difference
- Put a system in place to handle the other forecasts.
- Measure the accuracy of your forecasts and the contribution that you and your team make to improve it.
How much value are you really adding when you spend time forecasting? We all like to think that we make a positive contribution when we work on an important task. Unfortunately, when it comes to sales forecasting, that often isn’t true.
Spending too much time might actually be making your forecasts worse.
One study looked at 60,000 sales forecasts in four large companies. It showed that forecasters made adjustments to 75% of the forecasts that they reviewed. Many of those were small changes, slightly increasing or decreasing the forecast. Those small adjustments took many hours but made no significant difference overall. They were just as likely to make the forecast less accurate as they were to improve it.
When forecasters made large adjustments they were much more likely to improve the accuracy. Even then, large decreases in the forecast were much more reliable than large increases. When they reduced their sales forecasts by a significant amount, forecasters improved forecast accuracy by 6-23%. Large increases improved the forecast accuracy by only 1-3% overall. Many of the large increases predicted sales that failed to materialize and the change actually made the forecast less accurate.
The reasons are easy to understand. We don’t tend to make large changes without good reason, a clear change in the market or an event that has come to our attention. In other words, market intelligence. We are particularly unlikely to make a large reduction without good reason. Being natural optimists, or coming under pressure to meet targets, we may see positive influences as being more significant than they really are. That leads us to increase the forecasts by more than we should.
What about the small changes? Many of them are driven by things that have happened in the last week or two that stick in our minds but actually aren’t significant. Most of us tend to over-emphasize recent events, and to pay less attention to longer term trends.
The only forecast adjustments that add value are those that are based on reliable market intelligence"
1. Spend your time and energy on sales forecasts where you have market intelligence that makes a real difference
How do you use that information to reduce the time that you spend forecasting and produce better forecasts? The first step is to focus your attention on those areas where you have real market intelligence. That could be:
- Customers that are making big changes
- Products that you are launching or phasing-out
- Changes in the market or
- Customers that are over-stocked or under-stocked and so might change their order pattern.
Set guidelines for yourself and your team. Just how significant does the change need to be before you make an adjustment? Have you seen similar changes in the past? What effect did they have on your sales? By focusing on market intelligence you will only make a small number of adjustments. That means that you can afford to spend more time on each one, collecting data and seeking input from everybody with useful information.
Make sure that you record the reasons for your adjustments and the logic that you used to decide on the size of your adjustment. That way you can learn from that experience in future.
2. Put a system in place to handle the other sales forecasts.
That takes care of the forecasts that that you prepare manually. Your planning and inventory systems need forecasts for all of your products though, not just the ones where you have market intelligence.
Computer generated statistical forecasts fill this gap. They use the history of your past sales and apply algorithms to detect patterns and trends to predict future sales.
Make sure that the system uses simple algorithms that are based on real-world factors such as:
- Seasonality - ice cream sells more in summer.
- Trends – products that are growing or declining.
- Intermittent sales – products that are purchased or used occasionally such as spares.
- No pattern at all. Many mature products really do show no pattern at all. Their sales vary randomly within a range and the best sales forecast is simply the average of past sales.
3. Measure the accuracy of your sales forecasts and the contribution that you and your team make to improve it.
When you have a process that takes advantage of real market intelligence and uses a computer generated forecast for the others, you are ready to measure the accuracy of your forecasts. That will help you to refine your process and to ensure that your sales forecasting efforts remain efficient and continue to deliver the best possible results.
There are many ways to calculate sales forecast accuracy. The one used most commonly in sales forecasting is MAPE – mean absolute percent error. MAPE treats all errors in the same way, it is just as bad to under-forecast as it is to over-forecast. The error is expressed as a percentage so that you can compare forecasts for different product groups and over different time periods.
Measuring MAPE is a great first-step. Comparing the accuracy of your adjusted forecasts with the accuracy of the unadjusted statistical forecast provides an even better insight. That measure, “Forecast Value Added” , shows you whether your forecasting efforts are delivering value or if you would do better to simply use the statistical forecast.
Sales forecasting is probably an unavoidable part of life for your sales team, but it doesn’t have to be a time-consuming distraction. Your contribution comes from your market intelligence, and that comes from the work that you do every day with customers. By developing a consistent way to capture your market intelligence, using a statistical forecasting system and measuring the accuracy of your forecasts you will save time and provide the business with the best possible forecasts.