Manual data gathering and harmonization burn up costly resources, yet the accuracy and reliability of forecasts remain inadequate. While the advantages of Predictive Forecasting solutions based on artificial intelligence are well known, firms struggle to implement them due to a lack of orientation and know-how. A carefully chosen, low-risk approach that enables firms to scale up their Predictive Forecasting functionalities quickly and sustainably restructure and consolidate their forecasting towards a more efficient and effective predictor of success.
CFOs and finance professionals currently face the Herculean task of predicting the impacts of a firm’s decisions in the current volatile environment. Without the ability to identify relevant drivers, calculate different scenarios and countermeasures, firms squander money on faulty forecasts that can endanger the successful realization of a firm’s goals and plans.
The unfortunate reality is that this pandemic will stay as an obstructive force in every sector, and that means forecasting tools must adapt as well. As the pandemic has shown, future crises can stem from a plethora of non-financial causes, so the integration of a wide range of external factors into forecasting is indispensable. Fortunately, recent developments have yielded a new generation of forecasting that uses the power of Artificial Intelligence (AI) to provide CFOs with reliable and frequent forecasts.
Technological and methodological developments have completely changed the process of forecasting. In areas where we now often see a high manual effort to collect and harmonize data, cost reductions and time savings are enabled by the automatic gathering of all relevant data and standardizing the forecast structure to allow fast and efficient validation.
New forecast methods lead to more reliability, unlike many forecasts that are often heavily influenced by political considerations and the internal bias of the forecasting party, factors that do not apply to algorithm-based forecasts. Lastly, the presentation of the forecasted values has become more user-centric and focused. Meaningful correlations of the investigated KPIs with internal and external variables are effectively communicated to the user, and the forecast values are integral parts of all management reports.
Whether in defensive phases or times of strategic expansion, the implementation of Predictive Forecasting improves the capabilities of a firm’s operational and strategic forecasts immensely. Both dimensions of planning benefit significantly from the expanded pool of internal and external variables that are included in the calculations. The intelligent algorithms solve strategic puzzles that have been challenging a firm’s decision-makers for an unnecessarily long time.
For example, a newspaper publishing company might struggle with a large amount of unsold inventory due to large fluctuations in regional sales. Traditional forecasts that cannot identify the relevant external variables might come to the erroneous conclusion that internal factors, such as the content of the newspaper or the advertisement budgets, are the reason behind this volatility. Smart algorithms, on the other hand, might detect that previously neglected factors, such as rainfall, local sports events, or vacation times, drive a customer’s decision to buy. Only with insights such as these can the problem can be tackled effectively.
Despite the strong benefits of Predictive Forecasting, its prevalence among firms is still relatively low. BearingPoint’s CFO 4.0 study shows that less than 40 percent of the sample firms have adopted or piloted advanced data analysis and prediction functionalities or are currently on the journey of implementing them. However, different studies are showing that over three-quarters of finance professionals expect Predictive Forecasting to improve their forecasting function and therefore estimate its relevance to increase even further.
The main factors that lie at the core of absent success in implementing Predictive Forecasting are:
The common theme in these factors is the lack of orientation and clear guidance. Fortunately, their soft nature allows them to be eliminated without much difficulty when a proven approach is chosen, and the required know-how is present.
The journey towards Predictive Forecasting can be overwhelming due to the variety of KPIs, BI tools, and methods. To alleviate this problem, we recommend splitting the process into more digestible steps. As a starting point, the implementation of Predictive Forecasting should be preceded by an assessment of the current forecasting landscape and business requirements. The opinion of all key stakeholders should ideally be merged with external experience to develop a tailored roadmap from a Predictive Forecasting Minimum Viable Product (MVP) up to the transformation of the entire Planning and Budgeting process.
By adhering to this method, businesses profit from a low initial hurdle, as a suitable proof of concept can be selected based on several criteria, and business expert-driven forecasts remain in place until the Predictive Forecasting reports reach a satisfactory level of maturity. After the successful implementation of a Predictive Forecasting MVP, the entire system can be scaled up fast with relatively low effort. Depending on the use case, scaling can be prioritized in one of two dimensions. Depth can be added by further training and refining the model towards better performance. Alternatively, the use case can be scaled up by broadening its scope (e.g., more KPIs, more variables).