Model Open Analysis vs Model Performance Analysis
Hi!
I reckon there are two reports we can ask Anaplan to run for the models to gauge its performance. Wanted to know a bit more about Model Open Analysis and Performance Analysis reports - what do they entail and when should we go for one over the other. Thanks!
Best Answer
-
The model open analysis is a service that Anaplan Support can offer for anyone having modelling performance issues - contact support@anaplan.com
Other types of analysis (including concurrency testing) can be done through Anaplan Professional Service teams and you can contact your Anaplan customer account person, or work through a partner to set this up.
The third level of model analysis is available to all Hypercare customers, speak to your dedicated Hypercare advisor to set up.
In terms of what they offer, the model open analysis is the most basic. It does show all the calculations in the model, done in full, when the model is loaded. Some of these may be relevant to the particular performance issue you are chasing. It shows the total processing time, rather than duration.
Other model analysis can be more targeted towards specific user actions, showing only what is triggered when a user change is made. This allows for very targeted optimizations to be done.
The model open is a hammer, the targeted analysis is a scalpel!1
Answers
-
There are 3 types of performance analyses/services you can work to get from the Anaplan Customer Success team:
- Model Open Analysis
- Model Concurrency Testing
- Model Optimization Recommendations
Model Open Analysis
I can't find any examples of the model open analysis, but there is a brief description of model load times in this best practice article: . There is a further note in the comments:
Anaplan Support can provide a breakdown of the line items calculations to show the largest calculations. Often there are 1-10 line items that make up 80-90% of the calculation time. Reviewing and optimising those will often have a dramatic effect on Model Open time
My understanding is that when the model initially opens, it runs through all the calculations in the model. The model open analysis helps to guide areas of the model that can take a long time to calculate. This is not the sole driver of the end user experience in the model though.
Model Concurrency Testing
There is a good overview of this process here:
For concurrency testing, you will need to provide one or more end user "journeys" through the model, and the Anaplan team will write a script to simulate that behavior in the model as if many users were doing it simultaneously. This will highlight areas of the model that do not perform well when a large number of users in the model at the same time.
Model Optimization Recommendations
You can request a "Model Optimization", where the team looks at a specific process or input (that you specify), determine the drivers of poor performance, and provide recommendations on how to improve performance. The closest resource I can find on this is in a subsection of the concurrency testing articles on support:
Note that none of the services do any automated tuning or changes in your model to improve performance. You will still need to make these changes on your end based on the information or recommendations provided. Proactively focusing on keeping your model in line with Planual best practices is the best way to lead to strong performance.
5
