Accurate sales forecasts thanks to innovative AI solution
How will sales figures develop next year? This question is crucial for the strategic planning of companies in many industries. Up to now, forecasting methods were limited to heuristic estimates or simple statistic tools like moving averages or copying last year’s figures to the next. But now, AI is entering the stage – providing completely new analytical possibilities. In our Business Unit Business Intelligence, we have developed an AI-based model that delivers significantly more accurate forecasts than the standard functionalities provided by BPC in SAP BW/4HANA.
Many of our customers rely on accurate forecasts. These are the foundations for decisions in procurement, production, sales, and finance. The more precise the forecasts, the better a company can plan and optimize its supply chain. Forecasts help make processes more efficient, utilize resources optimally, and reduce costs – for example, by shorter storage times. Ultimately, business success largely depends on the quality of these forecasts.
In our Business Unit Business Intelligence, we have set the goal of developing a particularly powerful forecasting model using state-of-the-art artificial intelligence (AI). Thanks to AI, we can recognize much more complex patterns than with traditional methods. This allows us to comprehensively analyze historical data and link it with current market trends.
Time Series Forecasting with Neural Networks
For our model, we have implemented a neural network. Neural networks are very well suited to analysing multidimensional time-series data with complex (non-linear) dependencies. In doing so, both correlations across dimensions and recurring phenomena are recognized and continued as time-series forecasts.
Our neural network recognizes patterns in past sales figures and projects them into the future. The model not only extrapolates sales data for a single dimension, such as the development of sales in Germany next year based on the previous year’s data. It also identifies relationships between different dimensions. For example, it can predict how sales in Germany will develop next year if sales in France decline this year.
With our AI-based analysis tools, we can make sales forecasts for macroscopic dimensions such as sales organization, divisions, or product groups. Additionally, we can even make predictions for customer-item pairs, predicting how much a specific customer will buy of a particular product next year.
More Accurate Results than with SAP BPC in BW/4HANA
Our AI solution for time-series forecasting is currently in the proof-of-concept phase. We tested iton the sales data of one of our customers and compared the forecast results with those from the SAP Business Planning and Consolidation (BPC) software in SAP BW/4HANA. Specifically, we attempted to predict the year 2024 based on data up to 2023 and compared the results with the actual sales figures for 2024.
The result: Our model’s forecast shows up to 30 percent lower average deviation from the actual data compared to the BPC solution. Therefore, our model is significantly more accurate than the BW solution.
Author of the article
You need
advice or have
questions?
That’s our pleasure. We are happy to be of service to you. The best way is to write us how you would prefer to be contacted.
Kontakt Formular
