Suppose we want to build a “Preferred Vendor” column with a BI tool. We’ll need to add a column to the dataset, but this isn't Excel (or Spendata), so we can't. Instead, we’ll want to create an “index file” that relates Vendor Parent to a notion of “Preferred”, “Bypass”, “Approved”, and so on.

Next, we’ll link the index file into our dataset by key, in this case Vendor Parent. We’ll then derive a new column (“dimension”) across the link, referencing the related node from the index file. Now we can filter on the new column.

What happens when Vendor Parent names are changed? The linkage will be lost in this case, so we’ll need to remember to manually edit the linked file when that happens. We also need to update the index file when we add data to the system, and that must be done offline and manually as well.

For every custom segmentation we want to add we will have to add a new index file, external to the system, that we must maintain externally. Furthermore, in order to accomplish this magic, we will need someone familiar with database concepts who can set up the link, build the new dimensions into the BI tool, create new views of the dimension inside the BI tool, and so on. This person doesn’t need to be a database expert, but the task does require a degree of skill and diligence.

More complex questions may also need to be answered — for example, what happens when there is more than one match in the index file? When we link across to the index file, do we get both items? The first item? The second? A combination of the two? The answers to these questions may require even more manual attention to the index file.

The bottom line is that the level of expertise and diligence required to build, maintain, and link external index files is too high to be reasonable. Consequently, analysts don't do this, and they're not going to do this in practice.

With Spendata, the same task takes almost no time and can be accomplished by any business analyst.