A company could achieve high forecast accuracy numbers just a decade ago, leveraging only internal data. However, at the speed at which the market and economy are changing, not including market-and economic data would render your tactical and strategic uncompetitive.

What does a demand planning tool traditionally do?

The most commonly-used forecasting model employed in a planning tool is exponential smoothing, a model developed in the 1940s. It uses historical demand data to generate forecasts. The main disadvantage of this - especially in market forecasting - is its inability to correctly factor in and handle market shifts and trends. This method is best suited for passive demand forecasting. As it frequently predicts future patterns that resemble current ones, it's ineffective in long-term forecasting necessary for building informed strategies to drive business growth.

In recent years, newer planning tools have been utilizing AI-based models, such as the Prophet model developed by the data science team behind Facebook. 

But is this sufficient?

All, if not most, forecasting models built into planning tools (even the newest AI-based planning solutions) are solely based on OR restricted to using straightforward univariate models. The algorithm will only base the forecast on historical data of the forecasted variable. They cannot calculate the impact that market or economic variables have on the internal sales data, resulting in not considering the potential market trends and regional business cycles. 

With univariate forecasting models incapable of doing that, demand planners risk missing the opportunity to discover and incorporate leading indicators into their aggregated forecast. This factor is crucial to getting clarity over precisely what impacts your market, even more so in turbulent times.

‍

It isn't a case of discounting the importance of your organization's historical sales, as it undeniably plays an integral role in further optimizing your demand forecasting.
Instead, the emphasis is on stressing the importance of how weaving in external macroeconomic factors can help you get a personalized forecast that evolves in real time with your sales data.

‍

How do planning tools hold up in times of volatility?

We live in times where volatility is a given, impacting many industries agnostically. One of the pitfalls of solely using univariate models in one's forecasting process shows up here — simply depending on historical data as a critical input would limit your forecasts. An added advantage of using multivariate models is meeting volatility challenges by conducting scenario analysis. If the goal is to determine or plan for demand fluctuations, first identify your weakest link and its probable impact.

To compound this, many questions tend to emerge in turbulent times. Instead of scrambling to find answers, running a few potential scenarios allows you to experiment with different adjustments without interrupting the present one. Answering what-if questions based on simulation rather than gut feeling would be a wise approach to generating accurate forecasts. 

How to improve your demand planning and forecasting?

It isn't a case of discounting the importance of your organization's historical sales, as it undeniably plays an integral role in further optimizing your demand forecasting. Instead, the emphasis is on stressing the importance of how weaving in external macroeconomic factors can help you get a personalized forecast that evolves in real time with your sales data.

Here's an example. Vehicle sales have evolved throughout the years, affected by countless factors such as regulations, prices, trends, new products, seasonality, competition, growth, recessions, promotions, and interest rates. These factors did not affect manufacturers equally, which supports the need for a customized, personalized forecast.

To estimate anticipated demand, we recommend active demand forecasting. This approach employs internal and external data - the latter being quality data sets that capture the monthly or quarterly macroeconomic changes. (Such an approach can serve newer businesses well, especially those needing more sales data.)

By automating the detection of leading indicators from these data sources and executing top-shelf model aggregation, Indicio measures the impact of internal and external factors to predict sales accurately at any horizon in any given situation. Running these multivariate models and analysis in parallel, your team can combine sales forecasts and market information and let the two validate each other. 
‍