ProblemBy Gary Bolles
Technology: Content Management
Every company has too much unstructured data to manage.
Content management may be the information equivalent of the Gold Rush: Give workers the tools they need, and they'll find the nugget of knowledge gold buried in a ton of data dirt. The only problem is, in many cases you're the one who's actually helping create the dirt.
Most of the data that companies accumulate isn't created by workers typing names and numbers into database fields. That's structured data, information with clear labels defining exactly what the information is about, making it easier to find and use. It's the unstructured dataunlabeled information with no indication of what it is or how it's to be usedthat creates headaches.
Every time someone writes a Word document, types an e-mail message, creates an Adobe Acrobat file, or generates an audio or video fileusing all those tools that IT providesanother shovelful of unstructured data is tossed onto the company's information mountain. How much? Analysts estimate that fully 85 percent of all the data in an organization is unstructuredand that the amount of unstructured data in the average business doubles every two months.
Why is that a problem? Suppose you're looking for a company's address. If that information is sitting in a structured database, you simply use a query tool to search for the company's record and the address field appears. Simple. But suppose you're searching the uncharted, unstructured vastness of the Web for that same information nugget. Sure, it's probably somewhere on the company's Web site. But unless the word "address" is sitting nearby on the same HTML page containing the data you're after, a keyword search isn't likely to find it.
It may not be a big deal to go to the company's Web site, find the "About" page, and write down the address, but it requires a lot more time and effort than typing a few keywords. In fact, IDC research says the average knowledge worker spends 21/2 hours a day panning for information nuggets in unstructured sources like Web pages and Word fileseven though many of those pages and files may be their own. A year's cost for 1,000 knowledge workers not finding what they need, according to IDC: $6 million.
Welcome to the world of unstructured data.
Ask Your CTO:
How much of our data is unstructured?
Tell Your Business Constituents:
Here's the difference between structured and unstructured dataand here's why getting the information you need is so difficult.
Ask Your Data Guru:
How much of our current efforts are focused on managing information in databases versus helping workers better coordinate and use their unstructured data?
Just imagine what you could gain if you cut that unstructured data mountain down to size.
Suppose you could provide workers with up-to-the-minute business intelligence blending data on customer activity with pertinent internal documents. Or link together the long chain of insurance documents associated with a complicated claim. Or deliver reference manuals to technicians around the world, no matter how poor their Internet connections.
Enter content management. Companies automating business processes ranging from resume management and contracts analysis to patents claims and litigation tracking turn to content management software to ease the creation and coordination of large amounts of unstructured data. U.K. researcher Butler Group makes the persuasive argument that "in an information economy, content, and the processing of that content, must be carefully managed in the same way that physical assets and production processes have been managed in an industrial economy."
According to META Group Inc. Senior Vice President Andy Warzecha, three major business incentives are pushing companies to explore content management:
Getting a handle on all the information the company is producing. If you have the nagging feeling that you're not making the most of the information employees are creating, you're not alone. According to META Group, 60 percent of the Global 2000 have some type of content-handling software in place, ranging from simple document and image storage to complex information creation and dissemination processes.
Reducing cycle times. CEOs are seeing that having the right information at the right time can speed critical business processes. Having access to data that helps workers make better decisions faster is critical for designing new products and solving customer problems.
Getting ahead of compliance issues. Depending on your business, the Sarbanes-Oxley Act, the Health Insurance Portability and Accountability Act (HIPAA), new SEC disclosure regulations and the Food and Drug Administration's 21 CFR Part 11, which mandates electronic records and signature management standards for industries like biopharmaceuticals and medical equipment, will affect how you coordinate your company's content. If, for example, your e-mail records are subpoenaed due to suspicions about violations of "Reg FD" fair-disclosure laws, do you have the technology in place to adequately thread together discussions to prove your company's executives didn't overstep any boundaries?
Add the increasing pressure on companies to maximize their information capital and human resources, and content management begins looking critical for companies trying to provide decision-makers with usable data. "That's the crux of the content dilemma organizations are facing: 'How do I get to the right information that allows us to make better decisions?' " says META's Warzecha.
Ask Your CTO:
What kinds of content-related technology tools do we have today?
Review Your Company's Strategic Goals:
How clearly do our goals define the kinds of information employees need to be successful?
Ask Your Legal Department:
Just how exposed are we if we can't prove who has accessed specific unstructured data within the company?
Thinking about a content management system? Buyer, prepare.
The content management label is fuzzyafter all, what is "content"? Call it "information management," and you'll understand just how amorphous the issue can be. You're going to need a clearer definition if you want to get a handle on your unstructured information.
To figure out what kind of software you'll need, evaluate what kinds of information are central to the company's most critical tasks. Doctors need drug analyses, customer service reps must have reference materials, insurance brokers require complete descriptions of premiums. Next, ask how they access critical information. Is the data they need most useful to them on a Web site or in electronic document storage? Are workers most likely to be sitting at a desktop computer or carrying a PDA? Building profiles of such requirements will give you a clearer understanding of whether you'll need the content management equivalent of a few shovels, helping workers dig data out of a relatively small pile, or an industrial-strength backhoe that sifts through mounds of documents as rapidly as possible.
Typically, organizations in industries with tried-and-true processes for creating content, such as media, financial services and government agencies, will also want to automate the process of creating information. Using workflow software to improve collaboration and content development, teams in such industries can more efficiently develop training manuals or produce magazines. Being able to control the steps required to develop unstructured informationgiving you the opportunity to tag and categorize it at the same timecan increase flexibility when the time comes to distribute it.
But stay focused. Analysts say that "boil the ocean" content management efforts that attempt to automate data access across the enterprise are doomed to fail. Help business units understand they'll get better results by developing and managing their content locally, but within companywide guidelines. That's how Fairmont Hotels & Resorts Inc. tackled the problem. The company used Vignette Corp. software to automate the process of creating and publishing information on Web sites for its 41 independent properties, each of which manages its own promotions and customer bases independently. Each hotel can even create targeted Web sites for every wedding party and meeting, so attendees can get specialized event information. "The process would kill us if it were centralized," says Vineet Gupta, Fairmont's vice president of technology.
Ask Your Business Constituents:
Which employees feel the most pain when trying to manage their unstructured information?
Tell Your Business Constituents:
Let's devise a way to help those departments get what they needthen create a rollout plan to deliver to other groups as well.
Tell Your IT Implementation Team:
We need to offer the best balance between local and companywide management rules.
Choosing content management software is mostly a function of where you have the biggest challenges.
At its most basic, unstructured information can be handled by systems that simply manage the way you catalogue and store documents. At the other end of the complexity scale, content management systems can automate the workflow processes required to develop, store and disseminate information from start to finish. But if you're just beginning to look at the software that's available, prepare to be overwhelmed: At least 150 companies offer such applications. "There are still too many vendors in this space for them all to succeed in their ambitions," says Nick Patience, an analyst at the451. He separates unstructured data-related software into four categories:
Document, content and knowledge management. These systems work like conveyor belts for unstructured data, automating the process of piling documents and other files together. Think of linking the numerous components of complicated insurance claims. Some systems also speed the steps required to create the data in the first place, like building product catalogs or publishing this magazine.
Search and retrieval. Even with all that unstructured data sitting in the same system, workers will still need an easy way to find that one specific document that will make their task at hand easier.
Taxonomy generation and data visualization. Since search by itself often isn't enough, some companies like to add their own labels, or metadata, to describe individual chunks of information as they're pulling documents together. The categories used for those labels, known as a taxonomy, vary significantly: Compare the categories needed to label the data in medical records to those required for an auto-parts catalog. But it isn't enough to categorize unstructured data: It also has to be visualized, using tools that can organize unstructured information into a useful picture, such as building a three-dimensional view of the resumes of job candidates that shows the strongest candidates located closest to you.
XML databases and tools. Increasingly, companies want to label previously unstructured data in industry-standard ways, ideally keeping descriptions in the same files as the data itself. That's how XML (Extensible Markup Language) works. "The Web has made people realize that information can be more cheaply distributedif only we knew where to find it," says Philippe Gélinas, CEO of Ixiasoft, a vendor of XML content management software. Placing documents and Web pages into an XML database can help make unstructured data that much easier to locate. Remember, though, that XML standards for specific industries are constantly changing, so automatically exchanging documents with customers and partners isn't a given.
Analysts say the lines between the four categories above are blurring, and that the larger content management companies are beginning to make the process of integrating their software easier.
Will you be content with your content management software? Users and analysts say it's possible, with tightly focused efforts designed to support continual changes in workflow processes. Says Diane Brassea, manager of client reporting for AdvancePCS Inc., which processes pharmacy claims: "We've been improving business processes and reducing customer-response times. You'd be crazy not to do it." Just don't think that the content management arena is going to settle down any time soon. "Some people think it's a mature market," says IDC analyst Susan Funke, "but it's not."
Ask Your CTO:
In which of the four major content management areas do we need the most help?
Ask Your Data-Modeling Team:
Which data taxonomies are we using today, and are they sufficiently flexible to support our needs tomorrow?
Ask Your Vendors:
Can you help us get started by meeting the needs of a specific department while assuring me that you'll be flexible enough to help us expand later on?
: LPL Financial Services">
Strategic Profile: LPL Financial Services
Even if it's well organized, it's not enough simply to make your unstructured data available to customers and employees on a web site. In some cases, companies can reduce support costs simply by making their online data easier to find. That's the approach taken by lpl financial services, which decided to add a little structure to data searches.
CIO: Esther Stearns, managing director and CIO
Project Manager: Mike Hamm, service center assistant vice president
Problem: LPL calls itself the largest independent brokerage firm in the U.S. It coordinates the efforts of 4,400 representatives, many of whom call regularly for advice on policies and procedures. Accord- ing to Hamm, the company's Vantive/PeopleSoft CRM software reported that 50 percent of its inbound calls from brokers concerned such inquiries, even though many of the answers were available on LPL's Web site. But the site's inefficient search engine forced users to sift through hundreds of Web pages to find what they needed. It was much more efficient for them to call.
Goal: Reduce overall costs for customer service calls.
Strategy: Install a search engine from iPhrase Technologies Inc. that could automatically "spider" the company's Web site at least once a day, automatically indexing pages and pulling them into the search engine to make the content easily available through keyword searches.
Challenges: Initially, call volumes remained high as brokers learned how to find the information they needed.
ROI: The company went from 10 percent to 27 percent monthly call volume growth to flat growth, and did it at tax time, its busiest time of the year. LPL avoided having to hire another two or three customer service reps as a result.
Assessment: Hamm reports that, along with the cost savings, he's seeing significantly increased broker satisfaction. "Now they're getting answers they never thought they could," he says.