The Challenge: Defining the Data Problem
When you’re responsible for an entire organization’s financial forecasting, you’re collecting and confirming a lot of mission critical data – everything needs to be accurate, timely, and up to date.
Before using SAP Analytics Cloud, the team calculated projections by using multiple on-premise tools to plan, visualize, and predict in order to generate a forecast.
One of the line items in the Profit and Loss statement was Travel & Entertainment (T&E) expenses.
While travel has been significantly reduced globally, Corporate Controlling still needed to account for this change and plan on company’s performance. The overall goal was to set up a stable, predictive model to deliver a Central Forecast for T&E expenses that would help plan into the future.
But there was a problem: Since SAP’s organizational structure is continuously evolving, Corporate Controlling needed to adjust data structures frequently. This created multiple inefficiencies and frustrations for our analysts, who were spending an excessive amount of time developing and updating forecasts, rather than focusing on delivering high value insights.
The amount of time and effort also affected decision makers who relied on timely information to make significant business decisions.
In order to gain a more stable forecast, the team needed to improve their predictions and reduce inefficiencies by:
- Flexibly selecting the optimal dimensions
- Automating labor-intensive components
This is where Predictive Planning’s flexibility and automated machine learning came in.
The Solution: Predictive Planning
Predictive Planning, an integration between Smart Predict and SAP Analytics Cloud for planning, enables users to forecast revenue or costs at scale. By looking at historical data in a planning model and then leveraging the machine learning capabilities of Predictive Planning, teams can efficiently extract real-time data, uncover patterns, and provide a smart, trusted baseline for planning activities.
Here’s how it works: Predictive forecasts are created from dimensions (i.e. product, country). This enables flexible forecasting at different levels of granularity and provides their team with greater control and accuracy of predictions.
Predictive Planning is:
- Smart. It automates data-driven enterprise planning
- Self-service. It empowers business users to self-service predictions
- Trusted. It provides full transparency on the outcomes
In SAP Corporate Controlling’s case, they were able to use the governed connectivity of SAP Analytics Cloud to gain access to their data in real time, analyze their data, and collect predictions using minimal human intervention.
Corporate Controlling team already had a predictive model in use to forecast T&E expenses in a central forecast. However, that predictive model was segmented per profit center, which was a constantly changing structure, bringing time-inefficiencies into the predictive model setup process.
By defining measures (the actuals of SAP’s travel expenses) and toggling dimensions (accounts and cost of sales, instead of profit center), the team was not only able to get more exact and reliable insights into the company, they also automated multiple tedious processes, creating efficient workflows.
Corporate Controlling also found that they could quickly compare all the possibilities of predicting the travel costs based on various measurements in their hierarchy, such as account dimension, cost of sales dimension, and various forecast horizons, and more.