One more post from me about domain values – this time, on how we deal with them at GoldenSource.
As I mentioned before, you really need dedicated data-mastering operations to handle domain values. Without a rigorous data process in place, the domain values you reference can be of questionable accuracy because actual values do fluctuate with some frequency.
GoldenSource acts very prudently about all aspects of data mastering, which extends to the way we handle domain values.
An absolute prerequisite for taking this approach is to be in close partnership with data providers. It is the best way to help ensure we are always on top of upcoming changes to domain values well in advance for our clients.
We have a dedicated team that monitors all those changes as well as reviews every domain value change or addition, performing each review as soon as GoldenSource becomes aware of it to assess the level of its impact. This process also initiates the critical task of identifying the corresponding information from other data providers, and then aligning them to a common Gold Copy value (which of course still preserves all the original information, too, for lineage purposes).
To do this effectively requires a data model and a platform that’s capable of managing any number of external domain value sets. Each data provider has its own data value sets or enumeration codes for the same thing. On the surface, however, they appear very different from one data provider to another, and therein lies the challenge.
So, to compare apples to apples, you need not only to align the data points themselves, but also each descriptor within those data points. That requires a two-dimensional alignment, so if you’re looking at, say, the Corporate Action Status across data providers, you need to cater for each possible value.
Performing this kind of alignment also requires a standardized ‘conveyor belt’ operational process which enables enough lead time to manage any changes. If new domain values are coming in all of a sudden because the data provider’s announcement process had a miss, you need a responsive process that allows you to react right away, within a very condensed timeframe.
Without this process in place, there’s a very risky domino effect: data integrity is compromised, inefficiencies ensue, analysis is not accurate, decisions are misinformed, strategies misfire, opportunities are lost, and reputations are damaged.
But being prepared with data mastering operations, you avoid all those consequences. Think of it as prevention-focused wellness care for the data you rely on.