Hyperion Planning underwent a rash of changes in the 188.8.131.52 version release last spring. Not only did the entire UI get a face lift, but Oracle added a lot of new functionality to the tool as well. One of the features that has enticed clients to upgrade to the new version is the Predictive Planning functionality within Smart View. With Predictive Planning, Planning now has the ability to analyze historical data to come up with future predictions. Predictive Planning also creates graphs and data sets that can be used to analyze the past trending against future predictions.
Now that you know a little bit about the background of the new version, let’s walk through how to run your first prediction with Predictive Planning.
To use this functionality on your computer, you will need:
- An 184.108.40.206 Planning environment
- Smart View 220.127.116.11 installed on your computer
- The Predictive Planning application installed
- To be provisioned as Planner and Ad-hoc security roles
Once these requirements are met, you should be able to run predictions for any intersection you have access to.
Your First Prediction
Predictive Planning works by examining the current Planning data form and using those time periods intersections as a target set for what it is going to predict. In order for a form to be recognized as a valid form, it needs to the time dimension in either axis. Once the form is set up correctly, opening it in Smart View will display the Predict option from the Planning ribbon.
The Predict ribbon is arranged into four different sections. Run encompasses all options before the execution, along the actual execution itself. The view section allows you to compare multiple scenarios for trending and analysis purposes. All data options including exporting, filtering and pasting onto the data forms are contained within the analyze section. There is also a help section which includes the help documents as well as a few display options.
This example is going to focus mainly on the Run section to demonstrate the functionality of the new Predictive Planning module. The form I am using has three different versions in the rows; Best Case, Base Case, and Worst Case. It is also focused on only one sales account (4100 – Computer Sales) for demo purposes.
From this screen, select “Set Up Data Form” from the Run menu. We are going to focus on the Map Names section as that is where most of the important setup options are located. In this menu you are able to select which data is the basis for prediction, which scenarios you’d like to compare data to, and where you want the predictions to be pasted on the form. Predictive Planning will provide you with a Base Case, which is the predicted mean calculated using which ever regression method provides the best fit for the sample data set. Predictive Planning will also calculate a Best Case and Worst Case interval based on the prediction interval chosen in the setup options.
Predictive Planning works by comparing intersections of version and scenario. You can see in the screenshot below, we are using Actual (Scenario) – Final (Version) as the source of our analysis. We are going to populate this data in the Forecast scenario, with the versions Best Case, Base Case, and Worst Case, which were setup to match Predictive Planning. We are also adding Budget – Final so we can compare our numbers against that version after the prediction has run. Finally, we are selecting the option that will automatically paste the numbers in the data form after the prediction.
After the setup options are set, we can run the prediction by clicking the Predict button in the Run section of the ribbon. A confirmation screen will pop up showing the accounts you are about to predict, which months you are about to predict and which data set you will be using as your sample set. If you do not have a large enough sample set to predict the number of periods you are trying to predict, you may see a warning like this:
After running the prediction, the Predictive Planning panel will appear and show a graph of the prediction, information about the prediction range as well as statistical information about the result set and methods used to generate the predictions.
One important distinction to make is that none of this data has been written to the underlying cube. If you selected the checkbox indicating you wanted values pasted into predicted scenarios, the form has had these values pasted in, but you will need to click submit to commit them to the underlying cube.
Congratulations – you’ve just run your first prediction with Predictive Planning! This blog post has just scratched the surface of what you can do with this new tool. If you would like to discuss some of the more in depth functionality of the tool, considerations for implementing it or just have any questions in general – please feel free to reach out.