Simulating Data with SAP Analytics Compass

Compass is a native SAP Analytics Cloud feature which enables the simulation of probable impact brought about by driver uncertainties. It utilizes the relationship defined between the driver and target within the SAP Analytics Cloud model. With compass, you are able to perform scenario modeling of different assumptions and compare the probable outcomes.

Create simulations from a story table or the compass start page to answer business questions that you may have.

It is important to remember that a compass simulation is not the same thing as a time series forecast prediction.

Time series forecasts project the data trend from historical values into the future. The kind of business questions leading to a time series prediction center around what happens if things develop at more or less the same rate. For example, What is the predicted operating income if the trend for COGS development in the past 4 years persists?

SAP Analytics Cloud compass simulations use the Monte Carlo simulation, a mathematical simulation method delivering a range of probable outcomes as result. It does not analyze data trends, but instead, relies on repeated calculation using random inputs. This also means that a formula or definition of relation between the impacted KPI and the drivers is required before the simulation can be executed. As this method explores possible outcomes brought about by random inputs, the business questions leading to a compass simulation should center around what happens if things change. For example, What are the probable results of operating income if the COGS is between 10 to 20 million dollars?

You can use time series forecasts and compass simulations to enhance the attainable insights. For example, you could generate a time series forecast for an observation of what would happen in the future if the current trend persists and then perform a compass simulation on top of the predicted version to understand the risk context if a few key drivers are unexpectedly impacted.

Working with simulations and their private scenarios requires the setup of the necessary role permission and data access rights by the SAP Analytics Cloud administrator.

License: In order to use compass, you will require an SAP Analytics Cloud for Planning, professional or standard license.

Permissions: To work with simulations, you'll need to be assigned a role with permissions for the Compass Simulation object. You also need permissions for the models involved, as well as access to the model data. See the SAP Help Portal Roles, Permissions, and Data Access Control for more information.

You don't need to worry about setting up the simulation and defining the relationship between the target and drivers, because this is handled by the system.

What does this mean? You run a simulation on a specific target, which is the KPI that you want to simulate in the business question, and the drivers are automatically created for you.

You can create a compass simulation in two ways, from the compass start page or directly from a story.

If you use this option, then you will need to complete the simulation configuration in the Simulation Settings dialog by selecting the required model, target, and filters.

In the following example, operating income in US dollars is the target for the simulation. The filters applied in the Simulation Settings dialog are called output filters and are part of the metadata of a simulation. All scenarios within a simulation share the same set of output filters.

For more information, please visit How to Create a Simulation from Compass Start Page in the SAP Help Portal.

If you choose this option, the context (i.e. model, target, and filters) of the selected data cell is carried over into the compass simulation configuration.

Let's start by exploring the SAP Analytics Cloud compass interface and how the different panels work together when running a simulation.

Scenarios are used as experiments to answer specific business questions.

A private scenario is defined by input filters, the specific version it is based on, driver settings, case settings, comparison scenarios (if any) and simulation results. In this lesson, we'll explore these items in more detail as we go. You can run a private scenario for a specific number of Monte Carlo simulation iterations, depending on the chosen precision mode. Private scenarios are accessible only by the person who created it.

A public scenario is created when a private scenario is published. It contains all the settings and information of the source private scenario, however, all values in the driver list are in read-only state. Public scenarios are available to all users with access to the simulation.

More information can be found in Create Public Scenarios on the SAP Help Portal.

Drivers can be leaf account members or base drivers that contribute to the calculation of your selected target KPI, which, in our example, is the operating income.

Filters that apply to all drivers used in a scenario are called input filters. By default, they match the output filters set in the Simulation Settings dialog.

The value configuration allows definition of the range of uncertainty, while the distribution type defines how the system will go about choosing a random value from the given range. Choosing the uniform distribution, for example, will mean that every value within the range should share the same likelihood of being chosen, while choosing a normal distribution will mean that the values in the middle should have a higher tendency of being chosen. Per default, a compass simulation performs random sample with the normal distribution.

The values in Value Configuration and the Distribution selection combine to determine how random sampling is performed on a driver. All that is required is for you to input the value configuration minimums and maximums and, if required, change the distribution. If you leave the Value Configuration on a driver blank, deactivate it, or deselect it, then no random sampling is performed. Deactivating a driver (in the driver list settings) allows you to temporarily exclude effects of uncertainty modeled via the value configuration to quickly run a simulation without performing random sampling for that particular driver in the calculation. Deselecting a driver removes the driver from the driver list and is helpful when you want to focus on specific drivers in a simulation. Note that removing a driver from the driver list does not exclude it from the calculation of the target KPI, it simply indicates that no effect of uncertainty should be simulated for that driver.

In the following example, Distribution has been added to using the scenario's settings, however, the Distribution column is not visible in the driver list by default, but it is still taken into account during the calculation.

Restricted drivers allow you to refine your simulation further by restricting the uncertainty effect to only a certain filter data slice, allowing multidimensional simulations. For example, if you have gross sales data for both the US and EMEA, you can create restricted drivers to add specific value configuration by region. You can simulate, for example, a scenario where there is some churn in the US business but a new contract in EMEA has been signed.

You can choose between three modes of running, each with a preconfigured number of Monte Carlo simulation iterations.

1. Preview: The random sampling of the Monte Carlo simulation will be done 1,000 times. Or put it another way, the Monte Carlo simulation on a private scenario will be run with 1,000 iterations. This is the fastest mode but has the lowest accuracy.
2. Medium Precision: The random sampling of the Monte Carlo simulation will be done 10,000 times. This is a mode striking a balance between speed and accuracy.
3. High Precision: The random sampling of the Monte Carlo simulation will be done 100,000 times. This is the slowest mode but has the highest accuracy.

The probable results are displayed in a graph and divided into pessimistic, realistic, and optimistic cases.

Cases are a collection of simulated target values that fall within a specific range. Simulation results can be divided into three cases. In the graph showing simulation results, a case is visualized by the area under the probability distribution curve within specific boundaries.

Case Settings: You can change settings such as Name, Percentage, and Display Color for the cases in the Case Settings for Scenario dialog.

Now that we have simulated data in an SAP Analytics Cloud compass simulation and explored the various settings, let's take a closer look at comparing results in a scenario.

Once you have saved the compass simulation, you are then able to create multiple scenarios to simulate various combinations of driver value configurations and compare the results to each other.

There are two ways to create additional private scenarios:

1. Create a new private scenario. No restricted drivers or value configurations are included in this new version.
2. Create a copy of an existing scenario, including any restricted drivers that were created and value configurations that were added.

Whether you create a whole new scenario from scratch or create a copy, you can use a different version or new value configuration.

You are able to compare multiple scenarios that have been created in the simulation.

In this lesson, we explored:

• Creating private scenarios in a compass simulation in order to answer a business question.
• Adding minimum and maximum value configuration and probability distribution types to define a driver's randomness in a private scenario.
• Adding a restricted driver and define which data slice you want to set randomness on by configuring its filters.
• Running simulations for the private scenarios with different precision modes.
• Analyzing the probability distribution of the simulated outcomes in the graph.
• Explore the chances of achieving below or above a specific target value using the slider.
• Adding comparison scenarios to gain additional insight into the simulation results.