News and Features
Managing the Data Deluge, Part 3
Putting Data to Work Through Tools and Technology
By Pam George
A good report details where a company and its employees have succeeded and where they fell short. A very good report also lets a company make decisions about the future, says Anatoliy Chudnovskiy, Business Intelligence & Data Warehouse Practice Leader for Prime Technology Group in King of Prussia, Pa.
In the old days, searching for information was painstakingly accomplished by either manually flipping through reports or requesting specific data from the IT department. Reading annual and quarterly reports, however, can make even a CEO’s eyes glaze over, and securing the data from the busy IT department might take days.
Now savvy decision-makers use business intelligence tools to search stored data and produce valuable information. “You need to see some kind of trend,” Chudnovskiy says. “At the end of the day, it all comes down to retrieving and viewing records and analyzing data.”
According to Chudnovskiy, good business intelligence tools:
- Make sense of the business
- Measure performance and project the future
- Improve relationships with stakeholders
- Create profit opportunities
Lay the groundwork
Before jumping on the latest, greatest software, companies must understand their data, says Michael Ferreri, President of The Judge Consulting Group, a division of The Judge Group in Conshohocken, Pa.
Define the parameters. What is a customer, and where in the system is the customer information stored? Do the same for partners, vendors, products and services, geographic locations — whatever factors influence the business’s productivity and sales.
Even the best business intelligence system won’t answer all your questions, but data organized in an intelligent manner helps boost the odds, Ferreri says.
Brainstorm about the information you’d like the system to produce. What color dress sells well in January? What about March? Does product A sell better if it’s packaged with product B? Does product C sell more than A and B combined? Pinpoint the information that will help the company be “nimble and sensitive to customer demands,” Ferreri notes.
Information can come from a multitude of sources, including structured and unstructured data, such as emails, blogs and Wikis. You need to know where to look to collect the most useful information.
Like-minded data is often dumped into what is often called a “silo” or a “practice area.” Accounting, customer information, sales figures and human resource data might reside in four separate silos. When companies merge, they have silos within silos. Often, these silos exist in isolation because of the sensitive nature of the information.
Problems arise when you can’t access data in different silos to create comprehensive reports. And the myriad of disparate software platforms — created and supported by different vendors — often aren’t designed to teach each other.
“Companies are so happy they just have data under control and they’re not relying on paper,” Chudnovskiy says. “But the data is not being used to its fullest potential.”
Putting in all in one place
Despite its storage-oriented name, a data warehouse taps the collective information to support business analysis and decision-making, Chudnovskiy says. “It should provide a comprehensive view of the organization.”
An effective data warehouse might act as a central repository, thereby eliminating the silos. All data is stored in one accessible place. Or, the data is integrated from multiple locations.
“When a company builds a data warehouse, they have to keep in mind that they need to bring data from different sources: legacy data, multiple platforms, different physical locations,” Chudnovskiy says. “It needs to be what I call a single, consistent version of the truth.”
Using what’s already there
In these lean times, companies want to get the most out of the software that they’ve already invested in. “The way a company grows may make it challenging to get on a single data platform,” says Tom Connolly, President of BizTech, an IT service provider based in King of Prussia, Pa.
Companies are reluctant to replace applications — even if they’re silo-oriented. Enter service-oriented architecture or SOA for short. An SOA infrastructure allows different applications to exchange data with each other as part of business processes. The data can live anywhere.
Ferreri says “It is a way of saying, ‘OK, we give up. How do we get to the data we need without changing the place where it's currently stored?’”
The advantages of SOA are clear. If an employee is suing the company for wrongful dismissal, for instance, the company can extend a search for related data beyond the human-resources practice area, Connolly explains.
IBM, Oracle, SAP and Microsoft are among the big-name vendors rushing these business intelligence tools to market. Most the products are designed to sit atop those companies’ enterprise resource management systems. IBM, for instance, last year acquired Cognos, a business intelligence and performance-management platform. In 2007, Oracle acquired Hyperion Solutions Corp., a creator of performance-management solutions.
An effective SOA system will allow companies to do thematic searches, recognize keywords and catalogue metadata, Connolly says. It can also flag inappropriate or inconsistent data.
Meeting the need
Local companies are also creating configurable business intelligence solutions to meet their clientsą requirements. Portico Systems in Blue Bell, Pa., for instance, enables health plans to automate and manage all of their provider processes and interactions.
“It’s important for health plans to keep track of all their doctors,” says Sam Muppalla, Chief Operating Officer of Portico Systems. “What was the service provided and to whom? Where is the contract? What fee schedule should be applied for this line of business? We think about how to take their data and turn it into useful, actionable information health plans can use for intelligent processes to collaborate with their external constituents.”
All of the various forms of communication from the providers are kept in the data warehouse, even emails. The technology, which runs on an Oracle platform, features an auditing functionality that keeps track of every change made to a data element, regardless of whether it was changed by a human or by the system. “You can bring disparate information together into one repository and you can enforce the compliance/accountability via the use of business rules throughout the processes,” Muppalla says.
Coldlight Solutions in Fort Washington, Pa., has designed a business intelligence platform capable of identifying patterns in the data. For certain industries, such as the life sciences, Coldlight has established prepackaged libraries of keywords. Users can add to these words, if they wish.
The system goes beyond just searching for words, however. “The idea is to automate the process of discovering patterns in the data and presenting that back to decision-makers,” says Ryan Caplan, President and CEO of Coldlight.
The context matters. “You could say, ‘The team bombed at the Super Bowl,’ or you could say, ‘There is a threat of a bomb at the Super Bowl,’” Caplan says. “Patterns inside a data set are indicative of something that is about to happen. That is where the prediction comes in.”
The tool could therefore alert a financial services company before insider trading takes place — not after, he notes. Prevention is important in other industries. By merging together healthcare customer data with disease identifiers, an insurer can pinpoint the customers with risk factors — along with recommendations of how to reduce that risk. Users rate the presented patterns on their meaningfulness, which “teaches” the system to lean toward patterns with high marks.
The new standard
Depending on the size of the company, the cost of a business intelligence system could run well into the millions, and it may need staff who can harness its power. “It’s not something people typically have built into the budget,” Ferreri acknowledges. “Companies build in purchasing, implementation and training, but not optimization.”
But, that is changing. “The only way to compete today is to be more productive,” he says. “You have to have better information.”
Organizing the data
Business intelligence tools are the next wave of analyzing data and making business decisions based on that information. But for some companies, even organizing data is an issue. Enter Bryn Mawr, Pa.-based Hot Neuron, the maker of Clustify.
The software organizes documents by grouping them into clusters with similar documents. To do so, it analyzes the text and identifies structures within that text that appear naturally. No keywords are necessary. Then it labels the clusters for the user. The software can supplement a traditional search engine, or it can categorize the documents. It is also helpful for finding near-duplicates.
“We designed it to handle millions of documents — the more documents you have, the more critical our software is,” says Bill Dimm, CEO of Hot Neuron. “You don’t tell it what to look for. It looks for what is there. We like to cluster Wikipedia for fun.”
Targeted audiences include firms doing e-discovery and clients analyzing defect reports. The caveat: The person running the program must have access to the files he or she wants to cluster.
In the next issue
Part 4: Storage and Security
Storage has long been associated with backup. Today, it’s also about archiving. The flood of data has made a secure, organized archive a necessary part of records management. Many experts recommend a three-tiered system that addresses critical and non-critical information and the speed with which it must be retrieved
Photo-Illustration by Renée Guie
/ Hollister Creative
|
