By Dave Hannon
Many organizations today are investigating the best ways to extract value from the incredible volumes of unstructured data being produced in their enterprise. And frankly, most of what I’ve read about this made me think there is a LOT of work involved in finding the useful needles within the unstructured haystacks – possibly more work than it’s worth for some organizations.
But this week, I read a fascinating research briefing from MIT’s Center for Information Systems Research that put forward a boiled-down strategy for making the most of unstructured data, which has me thinking it’s not only possible, but valuable. And the the takeaway for the IT organization: this one’s on the business, not you.
To dramatically oversimplify it, the strategy in the CISR briefing goes like this: for data to add value it must be used. To be used, data must be searchable. To be searchable, data must or should have metadata. And unstructured data, by definition, does not have metadata. So to gain value from unstructured data – or to “structure” it – metadata must be added.
Makes sense, right? But here’s where it gets interesting from a who-does-what perspective. To effectively structure the data it must first have a business purpose. In other words, you must have at least a vague idea of how this data will be used and by whom before you can add metadata to it.
“Structuring data may appear to be an IT
-led activity,” the CISR brief says. “But because there are many ways data can be structured, doing so in a way that makes it easy to use in a business context requires a clear business purpose…Without a specific business purpose, structuring is open-ended, because it is not obvious from the documents or images what the right tags or search terms should be.”
And identifying that business purpose and developing the right workflows for structuring that data must be done by the business, not the IT organization.
For the IT organization, I’m not sure if this is good news or bad news (you tell me). Obviously, IT must “nurture” (MIT’s wording choice, not mine) the business efforts, but business must take the lead here. There’s just no way around it.
“Without business purpose and business ownership, managing unstructured data is an exercise in frustration,” the CISR briefing concludes. “With it, your unstructured data can become a new source of long-term value.”
Hopefully, this helps those organizations that might be working under the "data is an IT project-period" mentality. As always, I welcome your thoughts.
(Blogger’s note: the MIT CISR Research Briefing, “Leveraging Unstructured Data to Capture Business Value” was written by Anne Quaadgras, research scientist at CISR, and Cynthia M. Beath, professor emeritus at University of Texas, Austin. It includes a case study on Chevron's efforts in this area and is available to registered users of the CISR site. But if they let me re
gister, they’re obviously not too picky).