In my last post on the topic of demand planning, I mentioned that sales or demand forecasting is one of the most important elements of the demand planning process. After all, without some projection of how much inventory your customers will need, planning for demand becomes a pretty futile effort.

Certainly then, if you want your company to enjoy success with demand planning, you must strive for demand forecasting accuracy–which as you will know is more easily said than done. In this post, I want to talk about some common demand forecasting techniques and their limitations. At the same time, I want to suggest some benefits of using a system which blends two primary demand forecasting techniques; stochastic and causal forecasting.


Some Common Demand Forecasting Techniques

The three fundamental techniques commonly used for demand forecasting are stochastic, causal and empirical. Let’s take a brief look at each of these:

Stochastic Demand Forecasting: The stochastic approach makes an assumption that demand patterns are random and therefore draws on historical data. Transaction data from the past is analysed and used to predict future demand patterns. The demand planning accuracy of a stochastic approach is impacted by changes in the market, which can’t be detected through analysis of historic information.

Causal Demand Forecasting: As its name suggests, causal forecasting bases analysis on factors which might directly impact customers’ ordering behaviour. Many factors can be taken into account, for example, the time of year, weather conditions, sales promotions, and even the popularity and availability of product SKUs.

However, demand planning accuracy is limited when using causal analysis alone, since it’s extremely difficult to identify all causal factors and weight them in terms of impact on future demand.

Empirical Demand Forecasting: This is basically the use of market knowledge to predict demand. It relies on people, rather than software to perform a demand forecast. In reality, this method of forecasting is typically used to augment automated forecasting, with the intention of using human intuition and experience to hedge against statistical limitations.

In practice, this method often decreases, rather than improves demand planning accuracy, due to the influences of human emotion, which make it difficult to be truly objective in evaluating the strength of an automated forecast. This is why many supply chain organisations struggle with collaborative demand forecasting initiatives such as consensus planning.


A Better Recipe for Demand Forecasting Accuracy

Removing empirical forecasting from the equation, automated demand forecasting, even when left untouched by human hands, varies greatly in terms of accuracy. I’ve already mentioned the limitations of stochastic and causal forecasting. However, when a solution is used that has the sophistication to utilise both techniques, we start to get somewhere.

Advanced demand planning platforms, analysing data at the level of customer transactions can perform primarily causal analysis, while also employing stochastic calculations to fill in the gaps, if you like. In effect, combining both techniques helps to cancel out the limitations of each.

The result, while far from perfect, can be a higher degree of demand forecasting accuracy than is achievable with more rudimentary applications, such as those packaged within ERP systems.


Find Out What the Vendors Think

There is no shortage of demand planning software on the market and each vendor has different ideas about how to maximise demand forecasting accuracy. While some may choose to combine stochastic and causal analysis at customer transaction level, others may take a different approach.

If you get a chance to chat with some vendors in an environment where they are prepared simply to offer opinions, it’s well worth tapping into their expertise as part of your software evaluations. One way you can do this is to reserve your place at the Supply Chain Leaders Insights event scheduled for October 26th, 2016.

This unique conference is a very real opportunity to benefit from vendor expertise and knowledge on a purely advisory basis. You’ll also get to network with other supply chain practitioners who may have overcome similar challenges to those in your organisation.

More details of the event will be coming soon, so remember to check here regularly for updates.