You may also be interested in:


Embracing AI-Enabled Analytics to Improve Forecasting

  • By Robert J. Zwerling, Lawrence S. Maisel, and Jesper H. Sorensen
  • Published: 1/28/2020


For FP&A to cross the chasm from the traditional world of a backward-looking spreadsheets to an unbiased predictive powerhouse with insights for data driven decisions, it will first need to embrace a new Mindset.

Mindset is that stair-step that moves FP&A from the old-fashioned spreadsheet/BI “reporter” to an unbiased predictive “strategist” that has a seat at the table. Mindset expresses the aspiration of finance and commits to the journey to becoming an analytics business partner—a journey that is not necessarily long, hard, or expensive, it’s simply disciplined.


Before exploring analytics, the backdrop of Mindset needs to be understood. Traditional finance transformation largely refers to building efficiencies around basic reporting and developing acumen around business operations so that FP&A can better collaborate and contribute to decisions. These capabilities are valuable, but only rudimentary. Finance is expected to deliver timely, accurate and accessible data that can support decisions. However, it doesn’t earn finance a seat at the table.

The analytics business partner goes beyond traditional finance transformation to provide insight and foresight to impact decisions. The former provides influence to tell the business something it doesn’t already know, and the latter impacts strategy of what might happen, as well as how to make the future happen. These high-value components break the bonds of finance as an expense-only trusted scorekeeper to a topline driven strategic partner.

With each generation progression, FP&A assumes a new persona in the Mindset Model, from the reporter to commentator to advisor to strategist. With each persona is a progression from data collection to data transformation to predictive intelligence, as well as the involvement with decisions. The reporter supports decisions with timely and accurate data. The commentator contributes to decisions by producing reports with efficiency and adding analysis toward a more informed decision.  The advisor influences decisions with insights that shed new light to begin a culture of data driven decisions. Finally, the strategist impacts decisions with unbiased forecasts and predictions.

With each persona, the technology also advances. A reporter uses Excel. A commentator expands the toolbox to include visualization tools. The advisor takes the next step to desktop statistical tools for use on limited data sets, and the strategist advances to analytics tools with AI capabilities to consume diverse sets of multidimensional data, such as product, geography, customer, etc.


With the Mindset Model in view, we can readily begin to discern why too many successful analytics pilots are not adopted. Causes include the concepts of misalignment of Mindset, resistance to investment, and risk of failure. Misalignment is the most prevalent cause, and it is when business management and finance have different Mindsets, which will be explored in the case below. 

Resistance to investment rejects the ROI from analytics and sees expenditures on tools and people skills for advancing to an Advisor or Strategist as a cost to be avoided vs. an investment. Risk of failure is when management is so worried about their personal path of advancement in the company that all things new, even when proven, are feared for risk of failure rather than an opportunity for success.    


A finance group for a Fortune 500 telecom company conducted a test of unbiased statistical forecasting of product demand over a one-year horizon. It was highly successful, with an ROI over 200X, but the solution was not pursued because the finance team would not relinquish the role of a reporter—we call this “Death by Excel.”

When times are good, demand will consume most forecasts, and excess inventory is something to be avoided. But when times turn down—well, you know what hits the fan! The problem is biased product demand forecasts done in spreadsheets—as are done in most companies, even those with demand planning tools.

Biased means forecasts generated by humans, who while having excellent qualitative intuition of the business, are under personal and political persuasion toward a desired outcome. Case in point, have you ever seen a spreadsheet tell you something you didn’t want to see?

Unbiased forecasts are those generated by mathematics, which have no concern about what a CEO may want and are calculated from the input data. 

Now, it’s certainly fair to argue that a human can determine what data is consumed by the mathematics and what mathematics are used, which can conspire to derive a desired outcome. This can happen when the tool is a spreadsheet. However, when implemented in modern AI-enabled analytics tools, these issues are mitigated and, remember, a primary purpose to use these tools is to avoid bias.

In the case at hand, a back test was done to learn if analytics would have predicted a downturn in demand that happened after the company had experienced 16 consecutive quarters of growth, and how far in advance the prediction would have been made compared to the internal forecast.

The results were dramatic. Analytics predicted the downturn nine months before the internal forecast and forecasted the demand 11 months in advance with some 98% accuracy vs. the internal forecast that was about 95% accurate. Had the company had the analytics back then and used it to make decisions, it would have been able to both adjust inventory and provide market guidance to mitigate impact on its profitability and resulting stock price.

The word “and” is emphasized because another deficiency in using analytics is, well, using analytics without human interpretation to make decisions. However, this is a story for another time.

If you think this predictive intelligence was warmly embraced, you would be wrong. Instead, staff circled the wagons to berate and bury the test. But why? For the job security one thinks one gets from Excel. The only one who knows the data when a spreadsheet is used is the spreadsheet owner. The only one who knows how to operate the spreadsheet is the spreadsheet owner. And when the desired forecast does not match the actual outcome, the spreadsheet owner blames the data. The lack of transparency afforded by Excel translates to job security. 

As the Mindset of the reporter is limited to Excel, this was the Mindset of the staff that sought to restrict the introduction of new technology. Management was out of touch with their data-driven imperatives. They paid lip service to analytics, but never pressed for it. Management’s Mindset of a strategist was confounded by their approach of NATO—No Action, Talk Only.


For FP&A to be relevant in 2020 and beyond, it will have to adopt advanced analytics and AI.  The Mindset of the reporter was fine 20 years ago, but not today. If finance does not aspire to be at least an Advisor, it will not have a seat at the table. If it does not dedicate budget to new analytics tools and people skillsets, then its value to the business will be that of controls, compliance and historical reporting. Call it whatever you like—the modern CFO, next generation finance, or the analytics business partner—these are all about the future of engaging a culture of data driven decisions with AI-enabled analytics. So, what’s holding you up?

For more insights, don’t miss the session, Using AI-Enabled Predictive Analytics for Unbiased Forecasting and Decision-Making at FinNext 2020, in New Orleans, March 15-17. Register for FinNext here.

Robert J. Zwerling is managing director of Aurora Predictions, providing intuitive analytics with AI software for finance and operations. Lawrence S. Maisel is president of DecisionVu, a specialized management consulting practice focused on delivering measurable performance improvement through innovative tools and capabilities development that supports effective decision-making. Zwerling and Jesper H. Sorensen are co-founders of the Finance Analytics Institute, an organization that teaches finance how to implement an analytics culture of data-driven decisions.

Copyright © 2020 Association for Financial Professionals, Inc.
All rights reserved.