Data Discovery was introduced in the December 2022 patch update for Oracle FCC (Financial Consolidation and Close), and enhancements were added in the March 2023 and July 2023 patch updates. Data Discovery helps analyze issues by generating grids, and it displays which calculation the data came from. Oracle Support was using this functionality before it was available to everyone. It's nice that we can use it now!
I regularly get the question from clients, "This number doesn't add up, where did this come from?" Finding the answer to this question is similar in most cases: zoom in on all dimensions to see where the data came from, determine if the metadata has special characteristics, check related calculations, and check ownership. Going through all these steps can be time-consuming and require knowledge of the application setup. With Data Discovery, some of this work is done for you by right-clicking on the cell in a form you are analyzing.
"This number doesn't make sense, where did this number come from?"
An example is shown below where Data Discovery refers to the cause of the problem (incorrect ownership). It is also very interesting to find out which calculations affected the data, but that would make this blog too extensive. However, there is excellent documentation available from Oracle:
Oracle also dedicated an event to it:
The basic idea of data discovery is that you drill down from high-level cells to low-level cells, and Oracle does this for you in the relevant directions (=dimensions).
Now let's take an example of how to use data discovery. I intentionally used an incorrect consolidation %.
This is the "normal" Parent-child hierarchy, where Almere has a 70% stake and a minority stake of 30%:
In the parent/legal entity structure ("flat ownership structure"), I intentionally created an inconsistency in the Dutch consolidated structure (don't do this at home if you don't need to! However, it can be very useful if you have 70% ownership and 30% minority interest with a lower parent, but the remaining 30% is owned at another level and the minority should be reversed). The 70% ownership is the same as the "normal" ownership hierarchy (see previous screenshot), but in the flat structure I set the consolidation method to proportional instead of Subsidiary. No more minority interest % is automatically calculated:
In a standard form, I see an "unexpected" elimination of minority interest. This is unexpected because 30% minority interest must be calculated and consolidated to the top of the entity hierarchy. However, it is not unexpected if we consider the flat ownership structure where I set Almere to proportional, with no calculated minority interest.
After simply adding the Data Discovery functionality to the form, when I right-click on this unwanted value, I have the Run Data Discovery option:
This option creates a pair of grids, for example this one, with elimination data:
The symbol in the upper right corner of the screen means there is an attachment:
In some cases, the attachment already answers the problem. Not in this case, although it tells me, for example, that I have three metadata validation errors.
I drill down further by right-clicking on the cell with the number -22,500 and choosing Run Data Discovery again. Then more grids are generated, for example this Ownership Grid:
This gives me more information about the cause of the unexpected data at Dutch Consolidated level. At Dutch Consolidated level, you would expect a minority percentage of 30%, just like at Amsterdam Consolidated. But something else seems to be going on, because at Europe Consolidated both Percent Consolidation Input and Percent Min Input are 0% instead of 100% and 30% respectively.
This second inconsistency in the ownership data is confirmed in the "flat ownership" of E021 - Europe Consolidated:
In this case, I was able to resolve the inconsistencies in the data with Data Discovery, which pointed me to ownership data. Data Discovery can be very useful in solving data problems. Some practice with Data Discovery and knowledge of FCC is required before you can reap the benefits of Data Discovery. But with Data Discovery, you let Oracle search for you in several directions at once, which is much faster than going through all the options one by one (datagrids, ownership, metadata).
