How CFOs can challenge their teams to deliver better forecasts

CFOs need to deliver faster, more accurate forecasts. It’s up to them to decide what insights they want and then challenge their teams to choose the data and methodologies to make better decisions.

Today’s forecasting, planning and analysis teams have access to significantly more data than was conceivable just a few years ago. While your company may not be enhancing the use of this data, you can be confident that competitors and market disruptors are trying to get an edge by doing so.

It’s up to CFOs to challenge their teams to identify new data and methodologies that can deliver timely forecasts that reach deeper for the insight decision-makers need. While change isn’t easy, the tools and capabilities are there, and the time is now.

Forward-looking companies are building strategic forecasting capabilities that use artificial intelligence (AI) and machine learning to improve accuracy and enable leaders to make better-informed choices more quickly.

Based on our experience helping companies plan and implement better forecasting and planning systems, the following provides some field-tested guidance to help your teams get it done.

Think creatively about your data sources

To bring more meaningful, quality data into the forecasting process, finance teams need to think differently about where they look, both internally and externally. Data used in the past may need to be challenged, as traditional variables, such as fuel prices or interest rates, that used to be sufficiently reliable may no longer be good enough.

When brainstorming new data sources, finance teams may need to reach out across business functions. This may seem obvious, but many companies don’t do this well — often, we find ourselves helping companies connect the dots across siloed businesses and functions. Ask questions to reveal the blind spots: What external sources related to business drivers is the company tracking that finance could benefit from? Where do teams need to dive deeper into internal data to get a more future-minded view?

These are a few examples of how nontraditional sources can be used:

• Marketing routinely tracks buyer sentiment through social media and consumer search engine trends. In forecasting hotel occupancy, the number of consumers searching vacation destinations online could be a more forward-looking predictor than current airport passenger data.
• Supply chain teams track weather events to enhance logistics. This data can also be meaningful for forecasting financials and operations in the short and long term, as analysis may show a pattern of increasing frequency and severity of bad weather, as well as changes in buyer behavior in response.
• While pandemic-related uncertainty remains, more companies are using smartphone mobility data to track and predict shifts in buyer behavior. With its geographic specificity, mobility data will continue to be valuable, especially for managing through supply chain disruptions and demand swings.

Once you identify different data sources, your forecasting team can perform a variety of statistical analyses to see which metrics have more explanatory power. But as you include more data and variables, managing it gets highly complicated with traditional tools, such as Excel. That’s where AI and machine learning come in.

Deploy machine learning to adapt

Even after you have better data flowing in automatically, setting up forecasting models isn’t a one-and-done process. When the impact of variables is constantly shifting, as they do in the real world, your forecast accuracy may veer off course unless you keep rebalancing variable weights, introducing new variables or removing those that are no longer statistically significant.

But when you set up your data, models and time horizons in an AI-assisted system, this all happens automatically, whether you have 10 or 100 variables. With machine learning, it’s collecting data, comparing results between actual vs. plan and then continuously adjusting the variables to improve the accuracy of your latest forecast.

Teams that worked with EY on transitioning to new data and AI technology can update their forecasts in hours when it used to take weeks. The time they once spent wrangling data can now be spent conducting scenario planning, assumption analysis and simulations to understand the impact of potential actions and drivers on the business. This enables the company to evaluate dashboard-driven variance, conduct root cause analysis, identify emerging risks and recognize the difference between one-time events and sustained changes to the business.

Realize the value of strategic forecasting capabilities

Beyond helping companies manage through crises and uncertainty, CFOs with access to AI-assisted forecasting and scenario planning capabilities can realize significant value for their people and businesses. These capabilities help leaders make better-informed strategic and tactical decisions about managing capital so they’re not just placing bets without robust analytical support. Finance teams are freed from tedious spreadsheets to think more strategically (which doesn’t hurt with retention). And, of course, investors tend to reward companies that reliably meet or beat earnings estimates.

Click here to learn how the EY Forecasting and Scenario Planning Team can help CFOs tap into all their data to better analyze different potential futures.

This article originally appeared on CFO Dive.