With S/4HANA, the traditional SAP enterprise system has evolved from a transactional system into an intelligent digital core. This new generation application empowers the enterprises with extensive information, live data and analytical capabilities that are more beneficial for seamless day to day operations. Furthermore, most of these applications can be accessed from various devices such as mobile, tablet or a personal computer.
In addition to its regular competence to handle complex business operations, the new suite of applications is considered as “Intelligent Enterprise” as it is equipped with several other capabilities such as handling simple and repetitive tasks in an automated fashion, making recommendations on complex business issues and uncover patters and insights based on historical data. These features allow the enterprises to take the right decision at the right time.
The three key components of Intelligent enterprise are Intelligent suite (S/4HANA application suite), Digital Platform (HANA Database) and Intelligent technologies (Leonardo). Embedding the enormous capabilities of intelligent technologies such as machine learning with S/4HANA suite of applications, has transformed the whole suite into a self-driving and a self-learning autonomous system.
SAP S/4HANA Suite comes with built-in analytical capabilities to enable enterprises to monitor and analyze their day to day business operations and issues. The embedded real-time analytics offer a full coverage of monitoring and analytical features across several core areas and business processes.
In the interest of this blog, I would like to focus on the embedded analytical capabilities in Cloud Sales combined with machine learning capabilities. This is termed as Embedded Predictive Analytics. Predictive analytics as per the definition, extracts information from historical data and uses it to predict trends and behavior patterns.
In context with a sales department of an enterprise; this functionality within the digital core allows the sales staff to make necessary predictions on sales performance easier than before and can save immense time and effort spent on gathering the historical data and consolidating it into a meaningful report.
Currently, the embedded predictive analytics functionality in cloud sales supports the following business scenarios –
- Sales Volume Prediction: This scenario can help the sales managers to forecast the achievable sales volumes and develop the sales plans periodically. With the “Sales Performance – Predictions” app (business catalog “Sales Planning” and business role “Sales Manager”), it is possible to compare the achieved sales volumes with the predicted values (based on predictive modeling). The associated sales plan acts as a benchmark to analyze the extent to which the sales targets are being achieved. The analysis can give predictive insights to measure the current sales performance.
- Quotation Conversion Rates: This scenario provides reliable predictions for the sales manager or sales representative to monitor the probability of a sales quotation being converted into a sales order. By leveraging machine learning capabilities, enterprises can gain predictive insights into quotation conversion rate by comparing actual, planned and predicted results. With precise predictions, enterprises can focus more on value-generating aspects of their business.
- Predicted Delivery Delay: With this scenario a sales representative can monitor the current delivery performance situation, and instantly recognize the effect of the delivered-as-requested ratio of sales orders, to prevent a critical delay of delivered goods, and thus increase customer satisfaction.
The comparison of actual, planned and predicted sales values is possible only when the following items are in place:
- A “sales plan” is setup and released.
- A “predictive model” is successfully trained and set to active status.
Sales Plan: The “sales plan app” can be used to create, change, release, and display sales plans of an organization. In a sales plan, you can set sales targets on various dimensions such as sales org, distribution channel, division, customer, material etc. for a specific planned period. These planned values are later consumed to compare the actual sales values to the sales targets and predict the future performance by using the “sales performance – predictions” app.
The planned sales targets must be maintained in an excel format and uploaded into the sales plan app. The excel template to maintain the data is available within the same app to download (sample excel template shown below)
Predictive Model Training
To understand how the training of a machine learning model works, it is necessary to know the basic concepts of predictive scenario and predictive model.
Predictive scenarios are standard business scenarios that are available as an out-of-box solution within S/4HANA Cloud. These scenarios depict the predictive use cases of an enterprise with specific business goals, type of prediction required for business operations and the historical data required for the prediction.
Predictive Model on the other hand is a model used for predicting outcomes based on a historical data set and are linked to specific predictive scenarios. For a model to predict an accurate outcome, it must be trained at regular intervals with most recent data. Re-training an active model multiple times with most recent data will ensure the predictions are precise and meaningful.
Creation of multiple versions to a model is possible with a different status. Further, the system allows to review the quality and the training status of a model and delete any inactive model version(s) if necessary that have never been used or set to active.
An active model version can be put to a quality check and decide if it denotes the business data by analyzing the provided responses in relation to a business use case. The predictive model performance can be improved by continuously exposing it to new data sets as they are available.
There are two indicators available in the predictive model app to evaluate the predictive power and prediction confidence. These indicators can be used to evaluate the performance of a model version.
- The quality indicator “KI” evaluates the explanatory power of the training model when exposed to a data set. A perfect model would possess a KI equal to 1 and a completely random model would possess a KI equal to 0.
- Whereas the robustness indicator “KR” defines the prediction confidence and strength of the model. The degree of robustness corresponds to the predictive power of the model applied to an application dataset.
Analytical list page to demonstrate “Sales Performance – Predictions”
The sales performance app currently offers only ‘sales volume’ as the KPI measure. it refers to the total invoiced sales achieved during a specified period which is based on invoice items or debit memo items (debits) that have not been canceled.
The app is a good example of analytical list page. It can help to examine the business data from various perspectives, investigate a root cause for a specific issue using the drill downs and provide a detailed analysis of planned, actual and predictive figures. You can access this app on your Fiori launch pad under the business catalog ‘Sales Analytics’.
Below is an overview page with header (red) and content area (green) displayed for the key performance indicator (KPI) “Sales Volume”.
- Header area provides a visual representation of the data (chart form).
- Content area provides the detailed values of the key parameter selected in the header area.
The header area can be displayed in two types of filters i.e.
- Compact filter – Traditional display of filter values to get the desired output
- Visual filter – graphical representation of filter values (details covered below).
The visual filters have 3 options –
- Line Chart – used to analyze the time related data such as year/month (sales volume)
- Donut Chart – non time related parameters such as sales org, distribution channel, division etc.
- Bar chart – non time related such as customer or customer group, material or material group etc.
The adapt filter option at the top right-hand corner allows to customize which parameters can be applied to the KPI (sales volume) in the header area. Users can navigate to choose the chart type (line / donut / bar) or the value measure (actual / planned / predicted) as required for each parameter separately such as year/month, sales organisation, material, customer etc.
For non-time related series, it is also possible to sort the values in an ascending or descending order.
Additionally, selecting a parameter such as a specific sales org, customer or material from the header chart display area, adds up as a filter value for the final output. The data is refreshed automatically, and the information in the content area is updated based on the selected filters.
The chosen filter values can be saved as a variant so that the selected layout can be easily accessed next time.
The content area reflects the data based on the selected parameter from the header area. It can be displayed in 3 different views
1. Chart and Table view (default)
2. Chart view
3. Table view
Further, within the content screen the users can choose to display the data in different types of chart options available such as vertical or horizontal bar chart, pie chart, line chart, heat map etc.
Also, the table view format allows to export the content into an excel format.
The “view by” option allows to choose additional dimensions to the content area. For example, the data displayed for year/month is added with an additional parameter “material” as shown below.
Hovering over the chart area and selecting a specific item can display a detailed summary as shown below.
The content area can be maximized if required with a flexibility to move the columns and sort the values with a different parameter.
Further, you can also access other related apps by clicking on ‘actions’.
With changing market conditions and customer behavior, enterprises can no longer focus solely on delivering the best product or service and stay ahead of their competition. To succeed in a competitive environment, organizations must also pay attention to several aspects of their stakeholders (customers, employee, suppliers etc.) and understand the market trends and insights to make right decisions at right time.
In order to anticipate the business risks/issues, exceed the customer expectations and empower their own business teams, enterprises must adapt new technologies that are scalable, practical and which can support to achieve the business objectives. Embedded predictive analytics with S/4HANA Cloud offers predictive insight by offering some of the most relevant predictive scenarios and models to deliver insight into real time business issues and make better and timely decisions across the enterprise. These analytical apps are easy to navigate and offer various functions as explained above to achieve the desired result