In many corporate environments, unstructured data ownership is often taken for granted. Users and departments put data into their designated (and sometimes undesignated) shares never to really care about it again and ultimately passing the responsibility for maintaining that data onto the IT department. IT, of course, has no idea what data is important, what data is more important, what data can safely be deleted, and what fits somewhere in between. The ramifications of this are ever growing central file stores maintained by IT, increased backup times, increased restore times, increased archival requirments. All of these things cost time and money: whose time and money? IT’s time and money.
There is this assumption by the business that IT just takes cares of things like this just because its “on a computer” and that’s what IT does. This was highlighted perfectly by a recent submission to the Shark Tank where a product based company’s European division was selling products at 10% under company list. When confronted, the VP of Operations for the European division said that she didn’t update the prices in the system because IT did that. She was out shortly after that. While this example makes perfect sense to most people because it involves structured, business data, most people don’t see that this applies to unstructured data as well.
What can be done about this rampant abuse of IT? How can users be forced to be good stewards and step up to take responsibility for their data? First, I think that is an important thing to stress: this is their, the user’s data that we are talking about. It should be treated no different than the paper files in their desk or filing cabinets. Do they expect facilities to clean out their desk? I hope not. And if they do, they’ll probaly get laughed at.
In my opinion, there are two clear ways to force the responsibility back to where it should be. The first is metrics along with quotas. The metrics should be provided to users and departments so that they know how much they’re using and the growth of that usage. The metrics can also include oldest files, largest files, least recently accessed files, and duplicate files. Without the quota in place though, users won’t care about any of this. As soon as they hit the quota the first time and can no longer work, then they’ll care. Because they’ve been provided regular metrics and information on what could be cleaned up, they won’t be able to claim ignornace.
The second way is chargeback. I’m a huge fan of charge-back. Why? Because when it comes down to it, if someone needs more space because they say that their critical work depnds on it, who is IT to say no. If that user’s boss gets a bill for all that extra space at the end of the month though, that user may have some explaining to do. An addittional beneift of charge-back is simply the money to maintain and expand IT operational capabilities based upon actual usage and not some arbitray budget.
Additionally, the actual data owner should be asked to define their expectations for their data; this includes retention times, backup schedules, availability, growth expectations, and accessibility. Each of these things, if not managed, can become financial burdens on IT and drain away valuable time and resources.
Ultimately, unstructured data ownership is a partnership between IT and the actual data owner. This should be clearly communicated to the business and all users. A formal SLA is not really needed, just cooperation.
Posted by Jason Sandys 
Posted by Jason Sandys 