Having an accurate forecast is critical to your company’s bottom line. While we caution against evaluating a demand planner’s performance solely on MAPE, a widely used metric to measure forecast accuracy, we believe it provides a good starting point for a closer examination of your S&OP processes. If, for example, your forecast is consistently inaccurate, your culprit may just turn out to be bias; that is, something particular to your workflow could be causing your demand planners to regularly over- or under-estimate demand levels. But thankfully, there are a few simple steps you can take to reduce the havoc bias may be wreaking.

 1. Invest in a tool that will detect patterns in forecast error

Having a high degree of error doesn’t necessarily mean your demand planners are in the wrong, as there are many reasons this may occur. High market volatility may simply be inherent to your industry, for example, or your industry may operate in an extremely competitive landscape. But you should never see a pattern in forecast error. If you do identify a pattern in the errors made, either you’re using the wrong forecasting method or you might be dealing with bias. That’s why we strongly recommend using a tool  that monitors the forecasts you’re producing — and not just the final consensus forecast, but each one of your forecasts — and can detect patterns in the conclusions drawn.

 2. Don’t be afraid to make adjustments to your forecasts

Identifying patterns on each and every forecast made across departments — not just the final one — enables you to make appropriate tweaks to your forecast as you go. If you as the demand planner notice the marketing team is consistently optimistic about how much demand a new product will generate, work with them to fix this bias over time, so that eventually their forecast is adjusted down and, therefore, is more accurate. If your sales team, on the other hand, always provides a conservative estimate in order to ensure they meet their sales goals, make sure they adjust their methods so that their forecasts are more accurate. It’s true that forecasting is a numbers-driven endeavor, but it’s also one that requires an expert’s judgment and guiding hand to ensure all departments are constantly evaluating their contributions and then making the appropriate changes to their forecasting methodology.

3. Measure Forecast Value Add (FVA) regularly 

It’s imperative that you start your forecasting process with a strong statistical baseline forecast — and for that, having the right tool is critical. Yet since the next step, arriving at an accurate consensus forecast, requires collaboration and the input of various stakeholders across departments, it’s worth looking at how your error rate changes throughout the forecasting process by measuring the FVA of each step.  This will ultimately help you understand how effective each step was in strengthening the accuracy of the final forecast. Perhaps you’ll discover that significant value was added, but perhaps you’ll find the opposite. In that case, you would be better off trusting the statistical forecast and focusing your team’s energy on implementing strategies that improve demand planning for products with a lower forecastability.

4. Periodically run a Forecastability Analysis 

Running a Forecastability Analysis helps you determine the following: if you’re using the right forecasting methodology, if you’re forecasting at the right level of your product hierarchy, and if you’re focusing on the right products. In addition, it will provide you with insights into demand drivers, such as seasonal factors or other trends, and will offer strategies for dealing with harder-to-forecast products. While there’s no doubt that an initial Forecastability Analysis will go a long way to improving the accuracy and efficiency of your demand planning, you should conduct these periodically for the best results. You should, for instance, re-evaluate your processes with the emergence of a new competitor or any major shift in the market.

The common theme among all of these steps is that you must focus on the process of forecasting, not just the end result. Bias is a result of a flaw within the workflows used to arrive at your forecast, and by employing the unbeatable combination of advanced software and expert analysis, these steps help expose and either correct that flaw or reduce its impact.

Should you want additional insights and guidance, schedule a demo with one of our Forecast Xperts to see how John Galt’s solutions can maximize efficiency across your demand planning process.