SAP Analytics Cloud Planning : performances issues what can you do before contacting SAP?

Dear customers and partners,

When encountering performances issues,

what can you do before contacting the support of SAP or your Customer Success Partner?

Start by checking your local PC resources, your network speed, number of Chrome extensions and set power mode on battery on “best performance” for more detailed information see System Requirements and Technical Prerequisites – SAP Help Portal

During the User Acceptance Tests process, use the role configured with most accesses and not the Administrator role that isn’t relevant to your planning data collection process

Once you’re in the optimal conditions for your tests, assess what is slow, describe items perceived as slow, current time and expected time

This comprehensive KBA will help you 2511489 – Troubleshooting performance issues in SAP Analytics Cloud (Collective KBA)

For more “developer” profiles, open Google Chrome Developer Tools and analyze what is slow:

✓ Query data at backed
✓ Calculation on the fly
✓ Render and display data

Once you have localized the areas where you need performances improvements,

please find below some examples of features of SAP Analytics Cloud Planning that you could leverage in a short term period.

Implement security in the model and its dimensions (thus reducing the Private version “playground” of the end user)

Readapt “performances costly” data actions

  • Check security in user’s role (note that the data action triggers only the entities the user is entitled to)
  • Check the scope of the calculations and reduced it if needed
  • Use member set selection
  • Try to avoid too many If statements in the Advanced formulas (please find illustrated examples there SAP Analytics Cloud Advanced formulas best practices)
  • Use parameters of the data action to narrow the scope of the execution especially during a cross model copy process
  • Use function “Aggregate_write to” when cross copying from one model to another
  • Check the amount of data to be published in the details of the version

Review your data collection stories 


  • Define a Summary page and Use hyperlinks to go to detailed information (more levels if needed)
  • Use “mass or fluid data entry” modes to ease data collection for end users


  • Filter as much as you can, implement filters in documents rather than creating generic documents without selection possibilities in dimensions as drill-down to all details can be expensive
  • Take care at drill state – filter as many dimensions as possible
  • Limit the number of charts per page – max 6-8 parallel queries can be run by browser (depending by version)
  • Split big tables in multiple tables with fewer accounts
  • Propose to the end users to try “Add member” instead of configuring tables with “Unbooked data”
  • Less is more – request only what you need
  • Limit the numbers of descriptive columns into the table – easy to be read
  • Avoid formatting rules for tables with large datasets
  • Reduce as much as possible the cell references


  • Use Restricted measures instead of Conditional Aggregations in the model

Review your reporting stories 

  • Filter first (avoid loading all and filter after)
  • Split the information in more level of details:
    • Define a Summary page with global indicators – use conditional formatting to highlight a wrong result
    • Use hyperlinks to go to detailed information (more levels if needed)
    • Limit the number of charts per page – max 6-8 parallel queries can be run by browser (depending by version)
  • Limit as much as possible the linked data sources
  • Use page filters instead of individual filters per chart
  • Use cascading effect in case of multiple filters or input controls
  • Enable “unbooked members” only if needed
  • Rebuild the story in using “lazy loading” approach (tables one after another)
  • If your performances issues are related to the front end of SAP Analytics Cloud, try out the Optimized view mode more details in this blog and please be aware of the current limitations Optimized View Mode (Beta) Limitations (

On a mid-term basis, you may also reconsider the configuration of the Planning model with the following recommendations


Size of the model and its dimensions combinations

Often customers and partners mix up the concept of dimensions and properties

Check if

✓ all dimensions are required to define the collected information, I’ll give an illustrated exemple with the gender in a coming blog about HR Planning

✓ some dimensions can be transformed into properties of other dimensions

✓ the theoretical Cartesian product of your model with this document Calculation of the theoretical Cartesian product of the Planning model (isn’t it too large ?)
✓ the occupancy rate of the data in the model is relevant and split the model in several if the combination of all dimensions leads to large cube with low sparsity

Model calculations

✓ Try to find the right balance between on the fly calculated measures in the model and calculation done by data actions
✓ Use hierarchies instead of calculations
✓ In the classic model avoid accounts formulas, lookup formulas, exception on aggregation, try to initialize data with Data management import
✓ Replace Conditional aggregations with Restricted measures in stories