BI Stumbling Blocks Outlined at IDC Forum

By Lisa Vaas  |  Posted 2005-06-08

BI Stumbling Blocks Outlined at IDC Forum

CAMBRIDGE, Mass.—Theres this map of downtown San Diego, see, and its got big green dots designating businesses, and you can draw new fiber-optic lines in purple on the map, and presto!

You can see how many of those businesses those new fiber lines might be near; you can drill down into Dun & Bradstreet data to get the dirt on how much money those businesses have (and how likely they are to spend it on Cox Communications); and you can even get an estimate on what the fiber will cost.

So cool.

If you can use MapQuest, pointed out Coxs Mark Snow, director of field marketing, you can use it.

"It" being Coxs CMD (Cox Marketing Datamart), a data mart that sits downstream from Oracle Corp.s Data Warehousing and that integrates BI tools from MicroStrategy Inc., with MapInfo Corp.s location information technology mixed in.

It was only one of many demonstrations of next-generation analytics/BI technology now on the market and on display at IDCs Business Intelligence and Business Process Forum East: Impacting Business Performance by Applying BI to Repeatable Operational Decisions, here on Tuesday.

Yech, what a long and bo-o-o-o-ring title. But honestly, this stuff is finally getting useful to people on the front lines—useful in the midst of the work people do, as opposed to useful in some painful-to-read BI "report," as has been the traditional product of BI in the past.

Any of Coxs field reps can use the CMD, for example.

Or, in another demonstration, this time of Business Objects SA-enabled dashboard technology, Emergency Medical Associates Director of Data Management Jonathan Rothman showed how the managing doctors in his company—all of whom have attention-deficit-hyperactivity, as befits emergency room doctors, Rothman said—can just look at a dashboard to find out whats flashing red.

Too many patients walking out of emergency rooms because they have to wait too long?

Danger, Will Robinson, youre in the red zone, and profits are walking out the door.

For its part, Avid Technology Inc. is using SAP AGs NetWeaver to construct cube views from three areas: CRM (customer relationship management), marketing and campaigns, and finance and customer service. That information is pulled into dashboards to give a 365-degree view of the customer, available to Avids 150 worldwide BI users, said Chad Wright, manager of BI and CRM for the digital editing software company.

Wright pointed to an ROI study the company commissioned from Peppers & Rodgers that put the three-year rate of return at 68 percent.

Not bad, thanks to faster lead assignment and distribution, better segmentation of target audiences, and better analysis of marketing campaigns, using the hierarchical structure of SAPs CRM Market Planner.

Next Page: If this BI stuff is so good, why arent more people using it?

If This BI Stuff

Is So Good, Why Arent More People Using It?">

I could go on and on. There were plenty of nice case studies like these presented, but whats really worth looking into are the factors that are blocking the adoption of these advanced data analytics tools. As EMAs Rothman said, hes always surprised at the low level of BI adoption when he asks audiences for a show of hands.

Whats holding people up? For one, Drew Irving, director of Information Engagement at Diageo N.A., said that standardizing APIs has been a challenge at the beverages company.

"We have a lot of opportunity in sales and marketing, but we dont have the business relationships," he said. "We do have [the relationship] on the CFO side. She looks at it as standardizing and simplifying the business. Shes gotten the CEO on board with her. … The senior-level sponsorship will drive it."

Starting at the C level means the CFO and the CEO get to choose the dashboard metrics they want to look at. That will help the company shift from the hundreds of versions of monthly marketing reports theyve suffered with.

Thus, people will finally stop spending all their time questioning the metrics—i.e., "Where do these numbers come from?"—and can instead spend time on real business-enhancing decision-making, Irving said.

Thats cool, because customers are hung up on three things, and consensus is one of them, said Kirby Lunger, an analyst for Painted Word Inc.

Lunger told me that her customers are focused on data cleansing, data centralization, and getting management buy-in and user uptake for these new technologies.

Training was touted by many as being the way to get around user and management resistance. Specifically, many presenters advised hand-picking BI champions in divisions or departments.

EMAs Rothman got more specific: His advice was to hit em in the wallet.

He found a doctor at a hospital who was dying to do BI.

The project helped the doctor improve how much money the doctor was making, by targeting the time/days that patients walk out the door, increasing the number of nurses on staff during those crucial times, and thus decreasing the amount of patients that walk out the door.

More patients seen, more money in the doctors pocket.

"You talk about money with these folks. Were all human beings," Rothman said. "If you could show all of us how to make money by just paying attention … and thats what we did."

Next Page: "Why isnt anybody talking about security?"

Why Isnt Anybody Talking

About Security?">

But what really struck me was a conversation I had with Dietrich Falkenthal, a computer scientist from the Boston area.

The reason he was at the forum was to get a sense of the state of the commercial industry, particularly in terms of security vulnerabilities associated with storing and processing of critical data in BI applications.

Specifically, why doesnt anybody talk about security around all this extremely sensitive information in BI?

Whereas the primarily commercial enterprises represented at this BI forum were concerned with data that would give them, say, better pricing, whats at the heart of all this BI is sensitive competitive information that could be used in nasty ways by companies competitors.

"BI is shared among companies in supply chains," he said. "If you get in there and manipulate it, you could cause an adversarial company to make some wrong decisions," Falkenthal said.

"Business intelligence applications are a likely target for malicious attack from hackers, disgruntled employees or others. Such attacks probably occur, but it may not be in a companys interest to let us know about it because news about potential tampering with data utilized for business decisions could have material effects on the firms."

Falkenthal was also interested in visualization technology.

The key is not the amount of information that can be collected from sensors, user inputs or other data sources, but how to make it useful, especially in tactical environments.

Visualization technology is important to medical services, law enforcement and the military, for example, because they have a limited time to make decisions.

What cant be done with current technology, as far as Falkenthal could discern, is to come up with automated tools to intelligently handle complex real-time data.

"Now, even when people are presented with visual data, if its the wrong data presented, we can reach the wrong conclusion," he said.

"Tools are needed to process a lot of data and take some burden off users. Essentially, to do a smart push of important data that the user doesnt yet know he or she needs. For the most part, its still garbage in, garbage out, but visualization tools may help."

Falkenthal wasnt talking about data cleansing, per se, where records are combed through to eliminate name spelling variants, for example. He was talking about incorrect data correlations.

He didnt give me any specific examples, but I can imagine plenty of scenarios where you dont want police officers or airport security personnel to jump to the wrong conclusions because of incorrect data correlation.

"An interdisciplinary approach involving technology, economics, organizational and policy perspectives is needed to correlate data of this nature, because these systems are too complex to analyze or design using traditional systems engineering approaches," he said.

Research in this area is new, but he pointed me to universities such as MITs Engineering Systems Division or to companies and research labs that are really thinking about the future.

"Embedded BI, visualization, decision-centric tools, real-time decision-making and business process automation are all necessary to empower the end user in what we call the extended collaborative enterprise," he said. "This is all about pushing out power to the edge—to the end-user."

Interesting. I didnt have the opportunity to follow up with IDC analysts about this security issue, but theres always tomorrow.

Check out eWEEK.coms for the latest database news, reviews and analysis.

Rocket Fuel