News Stay informed about the latest enterprise technology news and product updates.

A data management strategy without policies spells doom

There is a need for speed when it comes to data retrieval, but a successful data management strategy begins with data classification, retention and storage policies.

Nevada's Clark County is the fastest-growing county in the country, issuing building permits for the Vegas strip about as often as marriage licenses -- which is a lot, considering that 10,000 people a month say "I do" there. Yet as recently as four years ago, the county didn't have a formal data management strategy in place.

Also experiencing significant data growth is the merchandising unit for all of Macy's private brands -- including INC, The Cellar and Greendog. The unit saves as images such detailed specifications as pants measurements or dishware dimensions, a procedure that has pushed its data storage growth rate to 100% to 150% as it moves into 2011, compared with a 20% growth rate this year.

Organizations usually would throw more storage at such problems -- an approach that is neither cost-effective nor helpful to the development of a long-term data retention plan. Both Clark County and Macy's, however, are tackling the issue by instituting data management policies to ensure that data is easily retrievable, as well as classified and retained properly in the first place.

'Policies first' is the best data management strategy

The first task Laura Fucci was asked to complete on becoming CIO of Clark County four years ago was to purchase an enterprise content management system to manage data collected from the county's 38 departments. She instead went to the records department to talk about establishing a data management strategy -- including a county-wide data retention schedule.

Technology was not on her short list of priorities.

"I've done records retention projects in the past, and it's very easy to get distracted with technology as you put together the records retention schedule," Fucci said. "You can get so enamored with the [technology] functionality that you lose sight of the program itself."

Fucci's first order of business was to establish a steering committee to oversee the records management and data retention program. In the end, there were 20 people in the group, representing the interests of the county's departments and government agencies.

With the help of a subcommittee of the steering committee, Fucci then developed a request for proposal to hire a data retention expert -- in this case, a vendor. One or more liaisons in each department -- 200 in all -- worked with the vendor to identify categories of records and develop a data retention schedule at the local, state and federal levels.

About 80% of Nevada's population and in turn, the records needed by the state for everything from births to air quality , are located in Clark County -- hence the need for such a far-reaching program that involved so many stakeholders. "It was the first time that the county took part in the record process at the state and federal level," Fucci said. "A lot of things being done at the county level were not on the state's schedule, and a data retention schedule was not being captured at the local government level."

In the end, there would be no single, county-wide policy for data retention -- it would be impossible when some records are stored for only six months, while others are not to be destroyed, ever. Instead, the county focused on identifying the data retention schedule needed for each record classification. It did that, in part, with the only piece of technology bought thus far for the project.

The data retention vendor installed a tool that tracked a retention schedule across all records categories: what the schedule was, which laws were associated with it, and which records were destroyed when.

As for data storage, no new technology has been purchased; but a quality assurance policy is now in place to make sure that documents are properly scanned, stored in a format that makes them easy to retrieve, and indexed appropriately. "Every department is doing their own document imaging and storing, but they are all following the same process to make sure that if a record is needed 20 years later, it's where it's supposed to be when someone goes to find it," Fucci said.

Fucci's approach to identifying stakeholders is a critical first step to any data management strategy, said Gwen Thomas, founder of the Data Governance Institute LLC. She recommends identifying stakeholders by classifying data based on how it will be used, with that use typically falling into one of three areas: operational, analytical, and compliance or regulations.

"Each of those areas has their own sets of policies, so there will be inherent conflicts, but those policies need to be aligned," Thomas said. "That means it's no longer a simple matter of managing the storage. Before you get to that, you have to have data governance -- identify the stakeholders, their goals, their policies and standards, then decide how to align them."

Expect resistance to aligning policies and standards, given that each data-use classification has stakeholders who follow not only their own policies but also their own data storage methods. It is best to approach an enterprise data management strategy one domain -- and one project -- at a time, she said.

I've done records retention projects in the past, and it's very easy to get distracted with technology as you put together the records retention schedule.

Laura Fucci, CIO, Clark County, Nev.

Thomas also had advice for what not to do: "The worst approach is to allow the business sponsor of this project, the one setting the strategy, to be someone focused on operations," she said. "They may be aware of some analytical needs for the data, but they certainly won't know about compliance needs."

If Thomas can stress one more point, it's the need for metadata. "Too often, metadata is not collected because those in operations, for example, know exactly what type of data is in a PeopleSoft HR system," she said. "But once that data is moved out of its native environment, into other environments [and for other data classification purposes], it loses its context and needs to be replaced with other contexts." Often, a group lack the incentive to put in metadata because it incurs the cost of doing so, which is why enterprise-wide policies and incentives for data management are so important, she added.

"Data storage strategies are changing because the responsibility for them has migrated from purely being technology driven, and driven upon a small set of policies," Thomas said. "Now enterprises are suffering the costs associated with e-discovery and compliance or search. Once they suffer through problems of not being able to find information, they move from a data management strategy driven by technology to a governance strategy."

A data management strategy driven by the need for speed

Jon Nam, director of technology at Macy's Merchandising Group Inc. (MMG) in New York, is on a quest for advanced data storage technology, but a metadata policy was his first order of business.

MMG is responsible for creating the designs for Macy's private labels, and MMG's IT department keeps track of everything from product quantity and type of material to sourcing and shipping to the stores. This task includes managing all the product designs' images for manufacturing, packaging and marketing purposes. That data is being stored, not just as images but also as Excel files and financial reports.

"All our labels are put on [storage] volumes, so our meta tags have to be accurate and updated easily … there needs to be a transparency to the process of finding the data," Nam said.

Now ready to tackle the back end of his data management strategy, Nam is reviewing a short list of products from data storage vendors: SAN boxes from Cisco Systems Inc., high-end network-attached storage units from EMC Corp., and Sun Flash storage technology from Oracle Corp.

For now, he's leaning toward the Flash technology. "Flash is so much faster, and everything needs to be faster and our capacity needs to grow to match network connection speeds, which are at a gigabit," he said. "That network speed is faster than spinning hard drives, so we need to look at Flash."

Terabytes of data might be small compared to the petabytes of data passing through Macy's overall corporate network, but buying advanced data storage technology sets MMG up better for the future, Nam said.

Technology may help bring it all together in the end, but too often enterprises overlook key policies for data classification and retention, and governance overall. Without these basic polices in place, finding data becomes an exercise in futility, the Data Governance Institute's Thomas said.

Let us know what you think about the story; email Christina Torode, News Director.

Dig Deeper on IT governance

Start the conversation

Send me notifications when other members comment.

Please create a username to comment.