This blog can be considered as a continuation to the earlier blog that I had put up regarding the various ways of using Multiple Measures in Performance Point. This blog deals more with the monitoring part of of the performance point trio and walks through the process that one could possibly use to create a KPI that consumes data from a model that handles more than one measure. The focus here is on the first approach, "Option One: Structure does the trick", i.e. using a dimension which is solely there to hold the various measures in the model (cube). The other approach of using the assumption model is pretty straight forward as one has to just include the appropriate assumption model in a particular model and s\he can use it right away in the dashboard designer.
So continuing with the previous scenario, we are interested in using two measures say the PowerCount & the ChairCount for all the rooms and want to show that in a KPI, which could later on be a part of a scorecard. These KPIs help us find how well we were able to match the number of member count in a room with the present number of power output or chair counts. This is just an example to put the concept across and shouldn’t be related to anything in real life. The trick here is to understand the basics of the KPI and leverage it to come up with a business solution.
Let’s start from the basics…when one creates a KPI he typically uses two metrics, the actual and the target. After he mentions the two metrics, he has to do the data mapping for each. Step one would be to create KPIs for both the measures and then step two would be to use them in a scorecard along with other dimensions; mostly a time dimension.
Creating the KPI…
The process for creating the two KPIs would be the same except a slight change as mentioned below. Individual steps follow…
1. Create a blank KPI.
2. Use the published cube as a data source via the data source mapping for the actual.
3. Select the default measure, "value".
4. In the "Select a dimension" section select the "measure" dimension.
5. Change the member selection from "default" to either ChairCount Or PowerCount. This along with step# 4 does the trick and gets the slice out of the cube that we are interested in.
6. Repeat the same steps for target. One could possibly use a different data source & dimension combination for this.
7. Repeat the said sequence for the other KPI as well; the difference would be in the step# 5. One would have to choose the other member in the dimension than that he choose earlier.
Creating the ScoreCard…
Once the KPIs have been created, one could use them in a scorecard and also add another dimension(s) in the columns, typically time or something else as per requirement to generate an appropriate context in the scorecard.
In the following figure, I have used a day-wise hierarchy to display the two KPIs for a couple of days. The target value that I have used is 10 with the scoring pattern being, closer to target is better; also, all of the days don’t have data to show. One could alter the values displayed in the blank cells by changing couple of properties.
Once this is in place the other dimensions can be added as per requirement.
Stay tuned some more stuff to follow…:-)