digital blue chart analytics forecast


Algorithmic forecasting in a digital world

Improving the forecasting process with predictive analytics

CFOs are in a prime position to challenge the way the enterprise looks at and consumes data. In the area of forecasting, they can champion an innovative, data-driven approach that will help people predict the financial future of their business. By modeling the potential impact of important decisions, they can help generate smarter insights and stronger business outcomes.

Forecasting processes that involve people working symbiotically with data-fueled, predictive algorithms

Human fascination with the future is part of our evolutionary heritage. Those who can foresee and navigate risks have always been more likely to survive than those who can’t. That’s just as true in business, where the ability to predict what’s ahead continues to separate performance leaders from everyone else. That’s what forecasting is all about, yet it’s surprisingly difficult (and often expensive) to do.

Traditionally, forecasting has been a mostly manual process with people gathering, compiling, and manipulating data, often within spreadsheets. There’s another way. Organizations are shifting to forecasting processes that involve people working symbiotically with data-fueled, predictive algorithms. It’s all made possible by new technologies—advanced analytics platforms, in-memory computing, and AI tools, including machine learning.

Keep reading below to learn more and download the sixth report in our Crunch time series, Forecasting in a digital world, for the full story.

So what exactly is algorithmic forecasting

The basics

Algorithmic forecasting uses statistical models to describe what’s likely to happen in the future. It’s a process that relies on warehouses of historical company and market data, statistical algorithms chosen by experienced data scientists, and modern computing capabilities that make collecting, storing, and analyzing data fast and affordable.

Beyond the basics

Forecasting models offer more value when they can account for biases, handle events and anomalies in the data, and course-correct on their own. That’s where machine learning comes into play. Over time, forecasting accuracy improves as algorithms “learn” from previous cycles.

Models are also more valuable when they’re grounded in richer, more granular data. In some cases, that might involve using natural language processing, which can read millions of documents—including articles, social posts, correspondence, and other text—and feed them directly into the algorithms. People can’t do that, not at the speed and scale required.

The magic

The real lift from algorithmic forecasting comes when it’s combined with human intelligence. Machines help keep humans honest, and humans evaluate and translate the machine’s conclusions into decisions and actions. It’s this symbiotic relationship that makes algorithmic forecasting effective—especially when humans are organized to support and share their findings across the enterprise.

Bottom line

Algorithmic forecasting doesn’t create anything out of thin air, and it doesn’t deliver 100% precision. But it is an effective way for getting more value from your planning, budgeting, and forecasting efforts. We’ve seen companies substantially improve annual and quarterly forecast accuracy—with less variance and in a fraction of the time traditional methods require—while building their predictive capabilities in the process.

Getting from manual to algorithmic forecasting

The ripple effect

How work changes

With algorithmic forecasting, Finance does more insight-driven work and less manual drudgery. Instead of spending their time grinding through spreadsheets, humans get to bring their expert judgement to the process. Leading finance organizations are already using automation tools to help with manually intensive work like transaction processing. Automating routine forecasting tasks is another area ripe for improvement.

How the workforce changes

Your finance talent model should evolve to keep up with changes in how work gets done—and that will likely require a different mix of people than you have in place today. Algorithmic forecasting depends on collaboration among Finance, data analytics, and business teams. Once they hit stride, these teams can move across the range of forecasting needs, embedding capabilities in the business and driving integration. These teams are integral to establishing an algorithmic solution that can work for the business, bring insights to life within the organization, and support continued business ownership of the outcomes.

Our experience has shown us that some finance professionals are simply better at forecasting than others. They’ve learned to set aside their biases and look at the bigger picture objectively—and they have a knack for understanding algorithmic models and uncovering flaws that others may miss.

You’ll need storytellers, too—folks who really understand the business and can translate analytical insights into compelling narratives that trigger appropriate actions.

How decision-making changes

Making choices becomes a more interactive process with advanced forecasting techniques, resulting in smarter, more informed decisions, even on the fly. In-memory computing, predictive analytic software, and visualization tools enable management to easily and quickly ask “what if” questions and produce a range of scenarios to help them understand potential impacts on the business.

How the workplace changes

Forecasting isn’t limited to Finance. Functions from marketing to supply chain to human resources all have needs for predicting the future to drive important decisions. While CFOs may not lead function-specific forecasting, they should help shape these forecasting initiatives since Finance will inevitably use the outputs they generate.

A shared forecasting infrastructure—even a physical Center of Excellence—can help improve collaboration and coordination while providing efficiencies in data storage, tool configuration, and knowledge sharing. And once the organization develops the forecasting muscle to solve one problem, the capability can quickly be extended and applied in other areas.

“Foresight isn’t a mysterious gift bestowed at birth. It is the product of particular ways of thinking, of gathering information, of updating beliefs. These habits of thought can be learned and cultivated by any intelligent, thoughtful, determined person.”

-Philip E. Tetlock in Superforecasting:
The Art and Science of Prediction

Common applications for advanced predictive models

Many of the companies we work with have begun their digital finance journeys by investing in cloud, in-memory computing, and robotic process automation. Others have broadened their ambitions to include advanced analytics, with an emphasis on forecasting. They want to create predictions that enable faster and more confident decision-making. This is where those digital investments can begin to pay off. Traditional approaches to forecasting can take far too long, cost far too much, and generate too little insight about potential future outcomes.

The most common applications for algorithmic forecasting we see today are:

Top-down planning

  • Target setting 
  • Integrated financial statement forecasting
  • Working capital forecasting
  • Indirect cash flow forecasting
  • Demand forecasting
  • Competitive actions and implications
  • Tax tradeoffs and revenue/profit implications

External reporting

  • Market guidance
  • Earnings estimates

Bottom-up forecasting

  • Product-level forecasting
  • Market or country-level forecasting 
  • Direct cash flow forecasting

Function-specific forecasting

  • Customer retention
  • Inventory optimization
  • Employee retention and attrition modeling



Case study

Is fast growth a financial problem?
Yes, if you can’t explain it.

The FP&A team for a global consumer product manufacturer frequently outperformed their guidance to market analysts. Problem was, they couldn’t explain the unanticipated growth, and holding credibility with the executive team, board, and industry analysts was a key priority.

The team suspected sandbagging was the source of their headache. Individual business unit (BU) leaders made their own bottom-up forecasts as part of the target-setting planning process, which was used for performance incentives. Finance had no objective way to verify or push back on the BU leaders’ numbers.

What happened next The toolkit
Finance leadership asked Deloitte to help them develop an objective, data-driven forecasting approach. Within 12 weeks, Deloitte’s data scientists designed a top-down predictive model that incorporated the company’s internal historical actuals and external drivers for each global market, including housing starts, local GDP metrics, commodity prices, and many additional variables.

The model enabled the FP&A team to deliver a second-source forecast based on external macro drivers that aligned to market expectations and provided insightful, accurate projections across the P&L, balance sheet, and cash flow statements. Planners also gained the ability to quickly create growth, recession, and other scenarios using desktop visualization software.
The Finance team received the fully functional predictive model built on an open source platform, which enables their CoE to manage and model forecasts on an ongoing basis. Leadership gained an objective, transparent, and visual conversation starter for discussions with business units about new opportunities and upcoming challenges for their markets.
Looking ahead
Leadership sees this as a game changer, and not just for their top-down financial forecasts. The business segments do, too. Following the successful delivery of a prototype focused on corporate FP&A, Finance has added data scientist capabilities to its talent model. Socializing the results with the business has created significant demand to dive deeper and extend the solution to the business segments and regions. Additionally, the client has taken steps to industrialize the model and to provide business users greater visibility into the driver assumptions and relationships to the financials.

The company continues the journey to scale the solution within the organization on their own. The client’s FP&A leader reports, “This is a good news story. We made an investment [to prototype algorithmic forecasting], proved the concept, and got the business interest. We created significant demand for a new Finance-developed forecast capability that will help the organization improve forecast accuracy and efficiency."

Making predictive forecasting a reality at your company with PrecisionView™

Traditional forecasting methods can be excessively manual and prone to human bias. PrecisionView—Deloitte Consulting LLP’s financial forecasting solution with integrated process and modeling capabilities—can help change that. With PrecisionView, finance leaders can leverage data science and pre-built algorithms to create deeper insights and better predictive models into their total enterprise, business units, geographies, and products.

What can you gain? More accurate, customized modeling and improved performance. Learn more by downloading the PDF.

It's crunch time for finance

Want to learn more about topics related to finance in a digital world? Explore other reports and guides in our Crunch time series and read case studies about digital transformation in the finance function. Whatever your interest, one thing is clear: From cloud computing and robotics to analytics, cognitive technologies, and blockchain, a new class of digital disruptors is transforming how the work of finance gets done.

stopwatch technology
Did you find this useful?