Rolling Forecast in IBM Cognos TM1

Recently, a client has asked to implement a rolling forecast model to move towards a more dynamic way of forecasting so as the periods move forward so does your forecast so you are always forecasting 12 / 18 months out in to the future.

With a rolling forecast the number of periods remain the same so as each period is traded it drops out of the forecast and another period is added. This is best shown with a diagram:

In TM1, this can be easily achieved using a period slider rule. To enable this functionality  the user simply updates the current month string in a global assumptions cube to start the forecast, this in turn updates a rule attached to a period slider dimension which updates the attribute values for those months within the period dimension. The business rule attached to Cube B then pulls data from “Cube A” Actual’s based on the attribute values for those periods and populates values in Cube B for Current Forecast Periods. Note that Actual’s are against real periods in Cube A ( i.e. Dec 2011 instead of m-1, the rule translates real month Dec 2011 into an “m-x’ month and updates the period description using an alias mask being the real period name ). See Ouput below:

Turning on the Alias for the Period Description below for Cube B below:

 

 

 

 

 

How the slider rule works (translate sliding period i.e Dec 2011 into sliding period i.e m-1 ). See embedded pseudo code in rule below:

Output in Period_slider attribute cube:

Now that you have an idea of the workings, the only real challenge that remains is to encourage/influence management to think outside the box and adopt a different way of thinking when approaching the forecast, one that I imagine is a mere walk in the park.

Enjoy! See attached for full download of above including cubes and rules.

Data PeriodSlider.zip

If you like this post, please spread the love…

One comment on “Rolling Forecast in IBM Cognos TM1

  1. Desi French says:

    Thanks. This was helpful 🙂

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.