Skip to Content
Technical Articles
Author's profile photo Girish M P

Sales Forecasting & Planning using SAP Analytics Cloud


SAC provides a wide variety of Analytical & Planning Functionality that can be leveraged to help the supply chain management of a company to carefully plan their manufacturing and stocks of their products ensuring they can deliver and satisfy their customers. The aim is to bring about a balance between having sufficient inventory levels to meet customer demands without having a surplus.

Since SAC has both forecasting and planning capabilities in a single platform Demand Forecasting and planning can be handled with great ease.


Sales Forecasting & Planning

Time Series Forecasting :

For this blog post, we will take sales data for the past 4 years ( 2018 -2021 ). The below image shows the model structure used.


Model Structure

The Actuals version of the data is used to run a Time series Forecast Predictive Scenario. The below image shows the settings used for the Time Series Forecast. The Signal for forecasting is the Sales quantity and is Forecasted for each Product Sub Category and Region of Sales.


Time Series Forecast Settings


The Forecasting is done for 12 months in the future forcing only positive forecasts since a negative sales quantity doesn’t make sense. Adding influencers to the model will reduce MAPE which increases the accuracy of the predictions.

SAC provides recommendations and MAPE% for each of the entities such as increasing data source size to get better predictions, Follow the recommendations to improve the model

Once the Scenario is Trained and you are happy with the MAPE% The Forecasts can be saved to a private version of the same planning model which can be used as a base for the planning scenario.


Planning :

For this Blog Post let us consider the Sales quantity planning is done for the Product Sub-category and Region of Sales. The planning granularity is as per user requirements and can be done at any level. But there is a restriction with the number of entities for a forecast. The time series forecast only supports 1000 entities by default, Follow this blogpost to go beyond 1000 entities.


To help the Planners with a base set of values we can use the Sales quantity that has been forecasted. A simple data action with a copy step can copy the values from the forecast version to the plan version is used to act as a reference or guide for Planning future stock.


The below image shows the Demand Forecasting & Planning Dashboard developed with all the Predictive and Planning Features



Demand Forecasting and Planning Dashboard



Conclusion :

This Blogpost shows how to implement Demand Forecasting and Planning in SAP Analytics Cloud using the inbuilt Forecasting & Planning Capabilities.


Assigned Tags

      You must be Logged on to comment or reply to a post.
      Author's profile photo Antoine CHABERT
      Antoine CHABERT

      Hi, thanks for the blog! You might be interested by this one, that explains how to go beyond 1000 entities: Best, Antoine Chabert (SAC Predictive Planning product manager).

      Author's profile photo Girish M P
      Girish M P
      Blog Post Author

      Thank you Antoine for the informative suggestion ,I have incorporated the same in my blog for easier navigation

      Author's profile photo Antoine CHABERT
      Antoine CHABERT

      Thank you Girish.

      One more suggestion from my side: you can maybe "tag" your blog with the custom tags Predictive Planning and Smart Predict so that SAP community readers can retrieve it more easily. Just a thought.

      One click all Predictive Planning blogs:

      One click all Smart Predict blogs:

      Best regards


      Author's profile photo Martijn van Foeken
      Martijn van Foeken

      Hi Antoine,

      I checked the blog you refer to. Are there any other possibilities like simply extending the number of entities to create a time series forecast on the roadmap?

      Kind regards,

      Martijn van Foeken | Interdobs

      Author's profile photo Antoine CHABERT
      Antoine CHABERT

      Hi Martijn, thanks for asking!

      It's a valid ask and we are effectively thinking of ways this limit could be further pushed into the future. This is a bit early to say more but we have this as a 2022 investment theme.

      It could be also interesting to consider SAP's IBP for supply chain forecasting scenarios where huge scalability is required (typically NOT in the order of magnitude of thousands). SAP Analytics Cloud is traditionally serving more of financial planning & workforce planning scenarios compared to S&OP. This has to be balanced depending on the exact implementation requirements.

      Please allow two extra comments from my side:

      • With the release 2021.17 we are bringing the capability to filter out some dimension / hierarchy members. This is a great way to limit the scalability problems. Not every combination (entity) is de facto fit for predictive and this helps focus the predictive forecasting where it should be focused.
      • There are different possible approaches to predictive forecasting (which are greatly facilitated by the nice features of SAC/Planning) - top-down, "middle-out" and bottom up. Some of the scalability limits will be hit harder when we favor a bottom-up approach upfront (crossing many dimensions together). Sometimes though, it's not the most effective approach to achieve accuracy. It really depends at which level the predictions are needed and where their accuracy needs to be maximized. There is no definite approach to gauging the best aggregation level, experiments have to be done, and validated on the basis of known data.

      Finally, and as surprising at this might sound, this request not been raised so far in SAC's influence platform. SAC PM teams are monitoring this channel and having requests with many votes helps creating momentum.

      I think there's definitely an opportunity for you to raise this request if you feel like it 😉

      Best regards,