Two easy ways to measure system performance
When you are implementing SAP Cloud for Customer or are running SAP Cloud for Customer for some time, you may or may not run into complaints of business users. Users might complain about slow performance that impacts their daily activities. Many factors impact the system performance, from local client configuration to slow internet connections and even specific configuration choices in SAP Cloud for Customer.
Rather than passively waiting until users complain, you can use standard functionality to constantly measure SAP Cloud for Customer performance and take proactive actions to troubleshoot performance issues.
In this blog, I describe two methods to measure system performance. The first is making use of the APDEX score. The second method uses response time kpi’s. Both methods are based on what users perceive as acceptable performance.
The APDEX score
The “Application Performance Index” (APDEX) is an open industry standard that measures user satisfaction regarding the application response times. The APDEX value can be any value between 0 and 1. The higher the APDEX value, the better the user experience.
The purpose of the APDEX is to measure if the actual response times meet the user’s expectations. You can ask your users what they expect to be a satisfactory response time. For example when opening a list of customers or opening a sales quote. Similarly, you can ask the users what they think is tolerable and what they consider as unsatisfactory or frustrating.
Next, we categorize these into 3 categories (satisfied, tolerated, and unsatisfied). A satisfied response time is considered to weigh more (1 point) than a tolerated response time (0,5 points). An unsatisfied response gets 0 points. This is expressed in the APDEX formula:
The response times are available in Datasource COD_RESP_TIME_ANA in SAP Cloud for Customer. Using this data source we can build reports and KPI’s that give actionable insight into the average response times and APDEX scores for each object and screen in SAP Cloud for Customer.
When you execute a transaction in SAP Cloud for Customer the response times that are measured can also be retrieved in the browser. In Google Chrome hit F12. On the Elements tab you can scroll down to the Performance Statistics where you find the actual end-to-end time and the specific SAP Cloud for Customer Workcenter and WorkCenter view
Although the data source contains a standard key figure (Average Response time), this key figure cannot be used in the formulas we need, so we have to create a custom key figure for the average response time as follows, AVG = Response time / COUNT
Calculated Key figures:
Using custom calculated key figures we created four (calculated) key figures
- Satisfied response time get value 1 when the response time is under 3 seconds,
- An unsatisfied response time gets value 1 when the response time is over 3,1 seconds
- Between 3 and 3,1 seconds the response is tolerated and also gets value 1
- The APDEX calculated key figure itself is calculated according to the APDEX formula: [satisfied] + [tolerated]/2)
For each measured response time, the value 0 or 1 is assigned so for each transaction i.e. opening a ticket or opportunity by a user, an APDEX score is calculated. Based on this report you can create KPI’s that shows a real-time APDEX score for a business object (Visit, Ticket, Sales Quote, etc.). The below image shows the calculated APDEX score for opportunities. In 24 cases there were 20 satisfied responses, 2 tolerated and 2 frustrated. According to the formula the APDEX is calculated: [(20 + 2/2) ] / 24 = 0,875.
Performance KPI tiles
In the second method we use KPI’s that are based on the average response times. Again by asking business users what they expect to be acceptable response times we can use these as threshold values in the KPI. This is the same as the APDEX method but the KPI value now shows the average response time and not the APDEX value.
So based on the reports we have built we get something like the image above. Instead of opportunities it shows visits. These reflect the response times for the visit search results screen to load and the time it took to open a visit. For simplicity, we total these times but of course, you can also make a split and hence create two KPI’s.
In the Visits KPI, we use the visit Performance report and the required report view and selection. On the values and threshold tab, we enter the response times that business users consider acceptable. (For simplicity we use the same treshold as used for opportunites).
Once the performance KPI is defined we can add the KPI to the home screen. The Visits KPI shows 2,09 seconds which is an acceptable response (satisfied) time for business users and falls in the green zone. The Accounts KPI shows 2,90 seconds. Was it just a glitch or should you keep a close watch on this?
By using the APDEX method you can also compare this with the APDEX of other applications. The big advantage is that the APDEX is a clear and unambiguous value for business, IT, and management on which they can act.
Troubleshooting performance issues
So now we have actionable insight into the performance of business transactions in SAP Cloud for Customer. In case the KPI indicators frequently fall in the red zone. It’s time for further analysis. Be aware that many factors impact response times including customizations, internet connections, or custom developments. If you are interested in how you can further analyze and improve your SAP Cloud for Customer performance, I recommend this SAP blog.