All data lives for a period of time. Some data has a short lifespan (I.e. today’s news) and some has a very long span (i.e. a social security number). Regardless, all data is time sensitive and with a good data life, it will die a noble death due to data apoptosis. However, if your data has not been designed well in the first place and secondly managed well, then it will most likely, at some point become orphaned, marooned or morph into something other than it was supposed to represent. This, is the tragic story of data decay. It is almost certain in weak systems, data dies prematurely due to data necrosis
The decay of data
Similar to the unpleasant thought of tooth decay; copies of the data with conflicting values and a lack of metadata to resolve the variations is the basis of data decay. As the quality of the data dissolves, decay grows and ultimately the data dies a premature death (necrosis). Which record is the right one? The answer is, we don’t really know; the one that seemingly has more truthiness. That is what data decay is.
And we certainly have a lot of data these days. And, we accept that some of that data is decaying. And so I say to you, don’t give any thought about the size of your data set. BigData, tiny data, it’s all data frankly. Focus instead on your high-valued data records/elements and grow a strong data lifespan management program from your high-value collection.
With today’s technological solutions,we can keep watchful eye on the health of our data in a manner once impossible. Several decades back, when disk space was at a premium, we were pretty clever at reusing data fields and data quality was beholden to a programmer. Those days are long gone; the more data the better now! It’s never been a better time to lay out an operational data quality program to keep data in your systems running with a healthy exercise regimen throughout its lifespan. And don’t forget to include a strong metadata program (like Spider8M) with an anticipated data apoptosis date as all data, like living cells at some point cease to exist.
Where to start
Know thy highly valued data first and learn all about what it takes to bring it to bear and its preservation. Next, set out to define an operational data quality framework and get your systems to a picture of full data health.
Click here to further discussions of data lifespan services.