Adapting to the times: Algorithmic forecasting

Within today’s global economy are links we didn’t know existed—in ways we didn’t expect. Impacts from events such as the trade wars, the Russia-Ukraine war, lockdowns in China, and the COVID-19 pandemic are challenging to predict and incorporate into business and financial forecasts. These unforeseen events have left CFOs wondering if algorithmic forecasting using predictive data models makes sense in our new, turbulent world:

• Is a predictive data model too abstract and unreliable without knowing the historical performance and influencing factors?
• Why do we need machines to forecast when people can do it just fine?
• Why does algorithmic forecasting matter when the pandemic has skewed data anyway?
  

So, has algorithmic forecasting run its course or retained its relevance today and into the future?

A hybrid approach: Increase predictive forecasting effectiveness

In the current economy, CFOs realize their organizations must carefully craft alternate, hybrid forecasting approaches to architect modern, best-in-class, end-to-end solutions. Hybrid approaches account for a mix of forecasting methods and help increase overall end-user and stakeholder model adoption. While sometimes a more simplified, traditional approach may be appropriate, the companies leveraging best and leading practices use algorithmic forecasting models selectively by considering market, geographic, volatility, and materiality to make decisions.

CFOs should react quickly and efficiently to provide data to make business decisions that align with enterprise strategic objectives. Defining a forecast’s level of detail and granularity is key to understanding what information is available for a specified decision-maker. When determining the proper forecast, leadership must consider three factors: