What’s yours is ‘mine’

Data mining met Wall Street in mid-May.

The Bellevue, Wash.-based firm digiMine Inc. announced that Dow Jones & Co. purchased its digiMine Enterprise Analytics product.

Dow selected the product to manage, mine and report on Web site usage data for its Internet publications including The Wall Street Journal Online and Barron’s Online.

The Wall Street Journal site alone serves more than 640,000 subscribers.

The digiMine product reportedly provides data mining that enables Web publishers to discover distinct subscriber segments and to better deliver personalized news and information content.

In addition, business managers at The Wall Street Journal Online will receive needed analysis to optimize online marketing, site design and customer relationships, said Todd Larsen, the online Journal’s general manager.

Data Mining also goes on in the Tampa Bay area’s business back yard.

Call center companies, credit card service centers and marketing firms monitor technical performance, consumer buying patterns and trends, perform risk management and collect demographic data.

Data mining uses include service improvements, new consumer campaign development, targeted cross-selling and product-bundling opportunities.

Sykes Enterprises Inc. collects consumer technical performance data, protects and manages it, and mines it in order to extract information sets for Sykes’ clients.

Gerry Rogers, senior vice president and chief information officer at Sykes Enterprises oversees data retrieval, virtual warehouse storage and mining for the international e-commerce and customer relationship management call center company.

Sykes specifically focuses on customer care and technical support performance criteria and scenarios that lead to service solutions for its clients, Rogers said.

Sykes, publicly held, has more than 40 offices on four continents. Sales for the fiscal year ending Dec. 31, 2001 were $496.7 million. The company’s stock currently trades at around $9 per share with a 52-week high of $13.47.

Sykes’ customer list, held confidential by company officials including Rogers, has major financial firms and their credit card operations, personal computer manufacturers, software publishers, hardware developers and telecommunication companies, news releases about Sykes have shown.

Personal computer powerhouse Gateway Inc. was revealed as a Sykes customer in April when the Poway, Calif.-based company ended its relationship with Sykes and switched to Service Zone Inc. in Homosassa Springs.

Data mining provides unique customer support, said Rogers.

“As an outsourcing company, we use fairly sophisticated tools to data mine those elements that our customers specify,” he said. “As part of our support, we have requirements to extract pretty unique data elements on behalf of our customers. That’s the gift of data mining. It allows you to be unique.”

Sykes doesn’t extract specific marketing data for customers because the client company often holds the data within a proprietary customer relationship platform, Rogers said.

Instead, Sykes concentrates on real-time performance data based on consumer service level agreements and client company service criteria.

“The primary task is that we get `X’ number of calls on behalf of a customer and handle those calls based on a set of service level agreements,” Rogers said. “Part of that task is to report back to our client the performance criteria for the end-user customers.”

Additional data includes telephone call lengths, abandonment rates — the length of time a caller waits before hanging up — and other criteria differentiated by individual client customers.

Sykes uses a virtual data warehouse where pulled data is displayed in a real-time format to support customers.

“I think there’s been a lot of hype around data warehousing,” Rogers said. “I put it that way because with the rapidly falling prices of hardware and software, to manage operations people tended to keep everything because they thought they might need it sometime. That’s worn off, and people are putting a lot more forethought into information.”

The real issue is mining those data elements and turning the elements into useful information, said Rogers.

“In markets people are in today, they have to move quickly and that information has to be provided them on a real-time basis so business decisions can also be made in real time,” he said.


Rogers brings 36 years of varied telecommunications experiences to Sykes. He served as general manager for business growth markets at AT&T and has lectured on international marketing and technology at the Georgia Institute of Technology, Emory University and Texas A&M University.

Credentials count when it comes to data mining, said Alexander Linden, director of emerging trends and technologies practice for international information technology consulting firm Gartner Inc.

Linden, based in Frankfurt, Germany, leads Gartner’s coverage on data mining, innovative information management and intelligent systems. He holds a doctorate in computer science and was project manager for General Electric Co.’s strategic enterprise technology division.

Statistics, such as those engineers are re-learning in Six Sigma classes, and calculation of variances are vital data mining tools, Linden said.

He sees many companies gathering customer feedback but not possessing the necessary human resource skill sets to analyze product complaints and create overviews.

Data mining expansion lagged because of this, Linden said.

“The major reason is that it takes certain skills to do it and people that have learned statistics,” he said. “It takes a tremendous amount of qualitative analysis to predict customer behavior. This is a very sophisticated thing that’s not for the masses. This explains why data mining is not a rapidly moving field. It’s a complex area, and not many people can pick it up over a weekend.”

Customer relationship management operations stand to gain the most benefit from data mining in a data-rich environment, said Linden.

He cited trends affecting what he calls “capable enterprises” and “average enterprises.”

The first is the “business as usual” syndrome as businesses struggle to keep up with front office changes. Data mining remains simplistic until enterprises start appreciating embedded data mining in databases utilizing software platforms from Oracle Corp., IBM Corp., NCR Corp., Microsoft Corp., and Blue Martini Software Inc., Linden said.

Meanwhile, Sykes utilizes a proprietary branded overlay system for its data warehouse called “Sykes Business Insight.” It takes raw data components and puts them into logical customer data segments.

Linden said that data mining becomes more sophisticated as companies increase interoperability, text mining and model combinations, which are able to account better for multiple and sometimes conflicting goals including customer satisfaction versus costs and productivity churn.

While Linden sees the entire market for data mining as a procedure increasing, he sees it embedded in other applications.

“Largely, it’s going to be embedded into enterprise applications ranging from customer relationship management suites over to supply chain management suites, over to databases, and also into knowledge-content document packages as well,” he said.


Sue Bross, vice president and technology testing manager at JP Morgan Chase and Co. in Tampa, sees data mining gaining new importance.

“It goes back to the basics. Are you are in business to provide a service to your customers, and what is it they want?” she said. “There’s an old saying, `The devil is in the details.’ The details are in your data warehouse. The answer is in data mining details.”

Bross previously was chief information officer at Catalina Marketing Corp. in St. Petersburg.

Catalina focused on what new ideas could be brought to the customer, she said. Cost estimates were put together for programs to help retailers understand what their frequent shoppers were doing.

Data mining att JP Morgan goes on at the corporate level, especially in credit card services, said Bross.

“It’s focused on, `How do we put together more programs that capture more of what our customers do?'” Bross said.

“Data mining is about finding all the `Ah ha’s.’ You can do a query against a data warehouse and get a lot of information back. Data mining means having the wherewithal to analyze what the data is telling you. That’s the great thing about data mining. You are looking for things you wouldn’t know without it.”

By: Pamela Griner Leavy

Source: Tampa Bay Business Journal / PDF

Leave a Reply