Evaluating analytics tools: Don't judge a tool by its label |
 |
By Hannah Smalltree, News Writer
07 Dec 2006 | SearchDataManagement.com |
|
|
When the government is using high-tech analytics tools to try and catch bad guys, they don't really care what the system was originally designed for -- they just want it to work.
Companies can also benefit from keeping an open mind and casting a wide net when evaluating analytics tools, according to consultants at Eastport Analytics. The Arlington, Va.-based firm's clients include the Internal Revenue Service, the Department of Homeland Security and private companies. Eastport uses commercially available analytics software and tailors systems and processes to clients' needs. Their business relies on having a broad knowledge of analytics tools, and they maintain a database of more than 400 products, which grows weekly, according to Mark Williams, chief technology officer. But though Eastport takes a global view, it has found that users tend to think much more narrowly about analytics, considering only common tools geared for their specific vertical industry or horizontal problem.
"Within certain communities of interest, there are analytics capabilities that other communities could take advantage of, but they don't know about them," Williams said. "We may find a tool in one industry that's applicable to the others, but they'd never know about it because they don't go to the same conferences or read the same publications."
For example, the legal profession has many tools that help manage evidence and extract concepts from reports. There are similar tools in the pharmaceutical industry, for evaluating clinical reports, and in law enforcement and intelligence. But many of these kinds of organizations aren't aware of tools used by other industries that may have similar -- or better -- capabilities, Williams said.
That said, there are some general-purpose analytics tools that are deployed across very different industries. For example, Somerville, Mass.-based Spotfire Inc. has clients in the pharmaceutical, energy, financial, intelligence and government markets that use the same analytics products for different purposes, according to Dylan Cotter, Spotfire's business partner marketing manager. In fact, Eastport Analytics recently created a system for the IRS to analyze banking data about individuals with offshore credit card accounts. The system used Spotfire and Microsoft SQL database -- corporate world tools deployed for a government application.
Don't let the names fool you
Analytics tools sometimes appear to be different or inappropriate for certain applications simply because of nomenclature, Cotter said. For example, corporate marketers use the word "segmentation" to describe the process of analyzing a customer base for patterns and splitting it up into distinct sub-markets. This is similar to what the pharmaceutical companies call "clustering" -- looking for patterns in large sets of clinical data. On the back end, the processes could use the same analytics engine, though the different terms could make them sound dissimilar to a business user.
A company can also customize a tool by modifying the interface based on the context or audience, according to Jonathan Adams, a principal at Eastport. One Eastport project transformed a standard rules engine into a tool that business users could understand. The Eastport consultants created a new skin for the rules engine, changing it from a technical-looking system to a more user-friendly analytical tool. The back-end engine was exactly the same, but the new front-end interface made all the difference to the business analysts -- and to the results that they could produce with the new tool.
In general, companies solving analytical problems may be well served by going beyond their industries' short lists and looking at the wide array of available tools, Adams said. There's no shortage of options out there and companies shouldn't limit themselves, he said. As always, it's critical to figure out the basic requirements of any analytics problem and not to be swayed by a persuasive vendor or particularly attractive user interface.
There's another good reason for companies to think big when it comes to evaluating analytics tools, Adams said. "You can end up with a much cheaper solution if you know what's out there."
|
|
|
|
