One of the biggest frustrations found in most credit unions is the numbers between two or more departments don’t jive. The conversation almost always leans toward trying to figure out which data set is correct, which one to believe.
The ultimate goal for these organizations is to find that elusive “one source of truth.”
So, why don’t our numbers balance between reports?
When we talk with credit unions, we find they have different definitions for common data terms and fields. Anytime the definition varies the count of that field will also vary. Having one consistent definition for these common terms is essential to getting to one source of truth. Establishing one common definition is complicated by the reality that different third-party software systems often use different definitions for these common terms. To get to one common definition will require a process to “normalize” the extracts from each of these systems and this is best done during the implementation and integration process.
To achieve consistent numbers the data must be reporting at the same time and date stamp. Any variation in the time stamp will result in data inconsistencies. This is often complicated when different systems generate updates and append the data at different times. The way to standardize this is to collect the data with the same timestamp. This may be impossible with some systems, so reporting the time stamps of each data source will help management understand the reason for the data inconsistencies.
We all know statistics can be used to tell lies as well as truth. The easiest way to manipulate data is to change the query logic. Often, query logic is a primary cause of report imbalances. Standardizing the queries on common reports can mitigate the risk of queries being used to “abridge” the story. To help management understand the differences in reports, a footnote with any query logic variances should be noted.
Getting to “one source of truth” should be an objective of any data strategy. But, like most strategies, this does not happen overnight and it is not an objective that can be solved with a software program. It takes a dedicated effort to find uniformity in queries, definitions, and report timing.