Skip to Content
Author's profile photo Kurt Holst

SAP Predictive Analysis to help select and place KPI’s in a performance management solution


This blog is intended to discusshow to create SAP Predictive Analysis models on top of a performance management solution, for instance based upon the Balanced Scorecard methodology, using SAP Strategy Management.

Using this approach gives some benefits when leveraging Predictive Analysis to help select the KPI’s in a performance management solution such as placing KPI’s in the right perspectives in a Balanced Scorecard methodology.

In this blog I will show how Predictive Analysis can provide evidence-based selection of KPIs in the four perspectives to the link between KPIs be uncovered, which should give a more accurate picture of the strategy map. The presumption is that you now have an existing balanced scorecard model with data to work with.

One of the greatest benefits of the Scorecard is the way it tells a story with the visualized cause-and-effect between KPIs in different perspectives. Using statistical analytics such as the algorithm as APRIORI as shown in this article can help validating the segmentation of KPIs could make this story telling even more trustworthy.

The gained knowledge using Predictive Analysis statistical functionality can be an important supplement on how one can manage and select KPIs for a Performance Management solution for instance when using Balanced Scorecard. The result from the exercise can for instance be used to validate, redesign and simplify the KPIs in a Balanced Scorecard. Evaluating and actually measuring the hypothesis of cause and effects of KPIs in different perspectives in a Balanced Scorecard could imply a higher possibility of getting the right measures and in essence the ability of being more proactive.


The cause of less than optimum performance management solutions such as Balanced Scorecards has in some cases been
identified as related to design and hypothesis of the KPIs. Moreover these KPIs are usually chosen based on gut feeling and of course also industry specific knowledge.

The obvious question here is how if we could devise a method to continuous evaluated and improve the selection of KPIs to a
performance management approach for instance using the Balanced Scorecard and in specific strategy maps with illustrations of cause and effects between KPIs in the chosen perspectives? This process will also be discussed and illustrated with an example.

Below I have illustrated how the use of statistical algorithms could help in showing the affinity analysis of KPIs and in essence underline the story telling using strategy maps. As shown the expected final result of evidence reasoned based strategy map incl.:

Overall Strategy map with cause and effects.png

Figure: Uncovering the KPI affinity using SAP Predictive Analysis to segmentation of KPIs.

How did I do this:

SAP Predictive Analysis have built in capabilities to use APRIORI algorithm also known as basket analysis to find records into segments with similar output field values. 

Apriori[1] is a classic algorithm for frequent item set mining and association rule learning over transactional databases. It proceeds by identifying the frequent individual items in the database and extending them to larger and larger item sets as long as those item
sets appear sufficiently often in the database. The frequent item sets determined by Apriori can be used to determine association rules which highlight general trends in the database: this has applications in domains such as market basket analysis.

SAP Predictive Analysis v. 1.0.11 is capable of leveraging data stored in the following data sources:

PA Data sources.png

  Where the option “Freehand SQL” can extend the capabilities of accessing further data source – requires local database driver

What algorithms to use to evaluate and chose KPIs for the Balanced Scorecard:



To find the KPIs that are associated we can use the Aprior algoritm.

Presenting the raw KPI data in SAP Predictive Analysis:

KPI APRIORI Results.png

The KPI data in this example consists of only 3 columns:

Questionnaire ID: Unique record – transaction number.

Manager ID: The manager who filled out the performance management / Balanced Scorecard questionnaire

KPI Name: The name of the KPI chosen by the manager.

Using the Algoritm “Association, R-Apriori”:

APRIORI Algorithm.png

Configuring R-Apriori:

APRIORI Algorithm setting up.png

After running the model this is how the “Predicted results” looks like using the Apriori algorithm.

APRIORI Algorithm results.png

As shown below an Apriori tag cloud chart enables you to visualize and find the frequent individual KPIs, based on the association rule. In this visualization chart, the highly prominent rules are the strongest ones. The prominence of the rules varies as per the onfidence and the lift value. Higher the confident value deeper is the colour of rules and higher the lift value bigger is the font of rules. You can change the value of the support, the confidence, and the lift, by adjusting the respective Range Slider:

APRIORI Tagcloud.png

Predictive Analysis is also capable of presenting the validity of the model as shown below.

APRIORI Scoring.png

Understanding the algorithm, Apriori: The first step of Apriori is to count up the number of occurrences, called the support, of each member item separately, by scanning the database a first time. We obtain the following result.


With the use of PA we now have a grouped gross list of KPIs which can be placed in the Strategy Map. For example KPI2 (E2),
KPI7 (E7) and KPI3 (E3) have a strong correlation as shown above and should then be placed in the same perspective in the  strategy map:

Strategy map perspective.png

Next step is to try and find the individual correlation between KPI’s in the different perspective using a bottom up
approach. In short we are trying to create the arrows (cause-and-effects) between KPI’s in different perspectives to heighten the illustrative capabilities of the Strategy Map. I have described this approach in another blog specific dedicated to determining the correlation strenght between KPIs.

What benefits could this method of Predictive Analysis on performance management approach such as a balanced scorecard imply?

• The right segmenting & selection of KPIs in a performance management solution makes the story-telling with the strategy map more trustworthy and in essens the organisation will know what is the overall goal, strategic objectives and how they individually can help the company by fulfilling the specific underlying KPIs.

• Predicting the outcome based on historical KPI’s and especially determining if a number of independent KPI’s change will have an impact months later on other dependant KPI.

• Measuring the correlation strength of KPI’s it could help select the right KPI’s to a Scorecard.

• Using statistical methodology to actual proving the correlation between KPI’s could uncover else unknown information and in specific relations between different KPI’s in a strategy map approach.

• Sanitation of KPI’s in a Scorecard – many times a lot of KPI’s are “for a safe or organisational-reasons-sake” put into a Scorecard. Evaluating if there is actually a need for all these KPI’s could lead to fewer KPI’s to monitor. This could also lead to higher transparency and less need for maintenance and data preparation.

• Increase the trustworthy including evidenced based reasoning of the story told in a strategy map.

This was just one example of how SAP Predictive Analysis with just one algorithm can enhance a performance management solution. Besides providing statistical proven reasoning to the strategymap Predictive Analysis could also assist many other areas which I will explorer in this blog.

Please feel free to comment.

Best regards,

Kurt Holst



The literature on Balanced Scorecard is extensive and there are many different methods for selecting KPIs for each of the perspectives. Professor Per Nikolai Bukh (2004) argues for the use of iterative qualitative depth interviews with the company’s top management to break the strategic objectives of KPIs for each of the four perspectives. An advantage of
this method is that there is created a dialogue about the company’s strategic objectives, and through this
, the strategies could be consolidated further. Strategies that cannot be decomposed will as part of such a process be
reformulated or completely removed. This will require more involvement by management than is outlined in the problem definition with few qualitative interviews. If a company should choose to implement Balanced Scorecard
it will be necessary to involve management in the selection of strategies and degradation, etc. In the present
thesis is the problem definition stated that data from the current performance management method is used to fit
into the
Balanced Scorecard method.

Using the KPI library from the existing performance management solution of from SAP Strategy Management to find the needed KPIs for the Balanced Scorecard and then finding the associations between these KPIs.

Example of KPI definition:

KPI Definition.png

Example of KPI library:

KPI Library.png

To find the gross list of KPIs that was collected in the questionnaires (Appendix last in this blog) the hypothesis is that we would like to find the common denominators of KPIs in each perspective. This approach might realistically need a management overhaul depending on the findings and the companies’ specific business domain – but in this article we focus on the statistical tool approach and the benefits gained by clustering, predicting and data mining KPI data.

Another method is for instance to perform quantitative interviews with the employees who thought of having a knowledge of strategies. This method makes it possible to involve more employees than when use is made of the qualitative interviews. Communication could be done for example through questionnaires sent via. mail or
electronic questionnaire. An example of such a form is prepared below:

Questionare 1.png

The completed questionnaires could subsequently be used to group each KPI in the four perspectives of the Balanced Scorecard
method. As well as a similar questionnaire survey to collect data on cause-and-effect relationship to the strategy map:

KPI “Employee Satisfaction” that belongs in the “Employee, learning and growth perspective” affects the following KPI “Internal process perspective” (the example assumes a linear relationship in the strategic map where each lower perspective points to the next perspective without jumping a perspective over):

Questionare 2.png

The completed questionnaires could subsequently be used to place each relevant KPI and then draw cause-and-effect relationship in a Balanced Scorecard strategy map.

Results of the guided process of questionnaires:

Employee Satisfaction is placed in the perspective “Employee, learning and growth perspective” and is linked as a cause-and-effect to the KPI “Efficiency of core processes” in internal process perspective.

Questionare first draft.png

Figure: Building the initial scorecard using questionnaires or from existing solution



“Strategikort. Balanced Scorecard som strategiværktøj – danske
  erfaringer”. Per Nikolaj Bukh, Heine Kaasgaard Bang og Mikael W.
  Hegaard. Børsens Forlag A/S, 2004. ISBN: 87-7664-017-5


Assigned Tags

      1 Comment
      You must be Logged on to comment or reply to a post.
      Author's profile photo Former Member
      Former Member

      Very good post, well explained and complete.

      Thanks Kurt!