Do you work with a financial planning & analysis (FP&A) team? If so, then you’ll know there’s excitement in the air. Why? First, because there’s no other group within an organization (other than the CFO and CEO) with a view into the operations and the financials like FP&A. Next, as CFOs continue to unleash the true value of finance, FP&A groups are destined to expand their roles as trusted advisors to business partners. And don’t forget, FP&A teams play a critical role in driving innovation for the Office of Finance.
In fact, we recently sponsored a webinar with FP&A Trends Group titled “The Power of Driver-Based and Predictive FP&A.” The webinar featured Alex Beired, Global Director of FP&A for Guardian Industries, and yours truly. In the webinar, we covered a range of topics, including the use of driver-based forecasts and the potential for predictive analytics and machine learning (ML) for FP&A. Here’s a quick recap of the discussion.
Driver-Based Forecasting Aligns Operations & Finance
After a brief introduction from FP&A Trends CEO & Founder Larysa Melnychuk, Mr. Beired shared Guardian’s multi-year journey towards a driver-based forecasting process. For manufacturing organizations like Guardian, driver-based forecasting emphasizes the factors that physically drive the business, such as specific products and/or plant-level volumes, price changes, and mix. Why’s this important? Because it enables finance leaders to trace the movement of inventory, material costs and product volumes across the internal supply chain – which can be worth millions of dollars in shareholder value.
Mr. Beired then reviewed Guardian’s key transformation factors. As Mr. Beired noted, developing and communicating key transformation factors is a critical step not only for finance teams but also to enable clear communication, collaboration and alignment with business partners. Guardian’s key transformation factors are noted below:
Given Guardian’s vertically integrated global business, Mr. Beired and the team concluded that investing in a comprehensive corporate performance management (CPM) solution was a critical step to modernize and simplify key processes like the financial close, reporting and budgeting, planning & forecasting. And with the OneStream platform in place, Guardian embarked on a multi-year finance transformation journey, which is shown in figure 1 below.
Figure 1: Guardian Industries Journey to Driver-Based Planning
The Game Changer: Going Driver-Based
As figure 1 illustrates, 2017 was a “game changer” for Guardian. Why? That’s when Guardian implemented advanced, driver-based planning processes. What does this include? According to Mr. Beired, the driver-based forecast process “fully aligns commercial planning and operational planning.” Here are some additional important factors Guardian considered for its driver-based unified business planning process:
Mr. Beired concluded with a brief overview of Guardian’s performance dashboards, built in OneStream, which help analyze actuals vs. plans down to the underlying drivers.
Telling the Story with Visualizations
Ad depicted in figure 2 below, Guardian’s Performance Dashboards provide visibility into how critical business drivers like price, mix and volume are trending vs. the plan. And with drill-through capability, Guardian’s FP&A and business teams can quickly understand underlying operational details that comprise each variance and then take corrective action. As Mr. Beired noted, “Visualizations help us quickly understand the business story to take action. And we expect advanced analytics to add further depth to our dialogue with business partners going moving forward.”
Predictive Analytics and Machine Learning (ML)
The webinar then shifted gears to predictive analytics and machine learning for FP&A. And with so much market buzz on these topics, I sought to help the audience dispel fact vs. fiction.
I kicked off the presentation with what finance people love the most – data. And what does the data tell us? Well, despite FP&A’s excitement about the promise of predictive analytics or machine learning, finance lags behind other functions in adopting these new technologies, as figure 3 illustrates.
Figure 3: Market Survey Results for Data Science and Machine Learning
What’s Holding Finance Back?
So what’s holding us back? Based on speaking with finance teams around the globe, I think it comes down to something simple – finance people do not yet know what predictive analytics and machine learning means for them.
Yup. It’s just not yet clear how these advanced models will fit into the day-to-day forecasting process. Here are a few examples of questions I’ve heard to further illustrate the point:
In addition to these questions, here are a few other misconceptions that cause confusion for finance teams.
I then shared some key differences between predictive analytics and machine learning.
Comparative Breakdown: Predictive Analytics vs. Machine Learning
Predictive analytics looks at historical data to produce a forecast. Such models leverage well-known statistical approaches and are therefore inexpensive to deploy. And while predictive analytics models may not yield the same accuracy as machine learning models, they are powerful – especially when forecasting general trends for scenario analysis or strategic planning.
Machine learning models take predictive analytics further. How? Machine learning models leverage human intuition.
For example, consider a retailer’s demand forecast for product A in region B within store C. Whereas predictive models rely exclusively on historical volumes to generate the forecast, there’s so many other factors to consider. Like what? How about the weather in region B? And what if store C built a new parking garage? Or what if a major competitor opens a location across the street? Don’t factors like these also impact product demand?
Of course they do. And machine learning models are designed to capture scenarios like the examples above.
But these machine-based forecast insights also come with a cost. What cost? Machine learning models are hard to scale across a sophisticated organization. In other words, there’s not one-button to click. Not to mention, machine learning also requires data scientists to work side by side with finance teams.
Bringing It All Together
Ultimately, it’s not a question of whether predictive analytics or machine learning is better or worse than the other. As I shared with the audience, the bigger and more important point for finance teams is to know what you’re trying to achieve with advanced analytics and then select the right technique.
Both predictive analytics and machine learning offer FP&A teams a new way to ask “why”. And there’s nothing bad about having an unbiased forecast scenario to help drive dialogue with business partners.
In fact, that’s part of unleashing finance!
For more on Guardian’s journey and to learn about the future of FP&A, click here to watch and listen to the replay of the webinar.