Vendor

Report Inconsistencies are Frustrating

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?

 

Definitions

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. 

Timing

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.

Query Logic

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.

E-mail me when people leave their comments –

You need to be a member of Credit Union Big Data Analytics Community and Summit - CULytics to add comments!

Join Credit Union Big Data Analytics Community and Summit - CULytics

Comments

  • CU Employee Community Chair

    Sometimes these “single version of the truth” discussions force credit unions to admit nuances to core concepts they've bandied about too carelessly for years. The example I like to cite is how our credit union wrangled over the seemingly simple question, “how members belong to our credit union?”

    It forced us to recognize that, for very legitimate reasons, we need to maintain multiple definitions of who does and doesn’t count as a member. From a financial/accounting perspective, we count as a member anyone who has a balance — for the simple reason that every penny has to be owned by someone and member counts need to foot to the balances associated with those account ownerships. In this context, deceased members are still members — at least until there's a final disposition of their balances. But joint members on accounts? In a purely financial sense, they’re NOT counted in this context, because you’d end up double counting the dollars and cents — not at all good.

    Contrast that with a marketing or service concept of who is and isn’t a member. Here, members who’ve passed away shouldn’t be counted for obvious reasons — they’re no longer folks who’ll be visiting our branches or reading our newsletters. But joint members SHOULD count. They are part of our financial conversations. They often are accorded voting rights in our board elections. When we want to express the broadest measure of our impact on our community, it’s a truer measure.

    These are just a few of the member class and status considerations that go into who does and doesn’t count to either definition.

    In whatever case, we do ask reportwriters, when they reference a count of members, to be specific — and note WHICH definition of members they’ve engaged. Both general definitions have very practical merit, in their respective contexts. It’s just very useful to not confuse the one for the other — and spelling out these particulars in report footnotes, as Mr. Jones suggests, is a best practice we’ve encouraged.

This reply was deleted.

Related Post