“We are surrounded by data,” exclaims Prof. Michael Kowolenko, of NC State’s Virtual Computing Lab. “Every business, whether a back room collecting financial information or a sales force trying to determine what customers want, is collecting data but that data is rarely easy to access by everyone with a need.”
Kowolenko observes that data may be tailored to a sales force, for instance, but that won’t help an R&D organization trying to understand where demand is coming from, or a financial backroom service trying to understand what’s going on in that R&D organization.
Moreover, he adds, there are many businesses — manufacturing, for instance — where you collect a wealth of information around how your manufacturing process is running but you don’t really use that information. You let it all stockpile while perhaps a few select individuals can access it. Kowolenko’s lab is currently working with CIMS and IBM to solve this problem by “separating the signal from the noise.” Here’s his account:
Structure and Other Obstacles. The difficulty is that so many data sources are structured in a way that makes it hard to extract the necessary information. And often you’re dealing with an IP organization whose charter demands maintaining the data in these sources in a particular structure and format, and restricting access.
Often the available software and tools are cumbersome and difficult to generate reports from. In addition, along with trying to understand what’s happening with data, you also have to have a process that allows you to look at the data, evaluate it and put it in the right context.
You have to understand how to ask questions of the data, but many times people don’t do that in a structured format in the sense of, “What question am I trying to answer? What problem am I trying to solve?” Or perhaps, “Do I have a problem?” or, “I actually have an opportunity here but I’m not taking advantage of it?”
That’s because they’re not looking at data in the appropriate context. Suppose, for example, an R&D organization could do something with a new technology that meets a demand the customers are relaying to a sales force and yet the two don’t talk. If that R&D organization had the ability to query a customer service database it just might get new ideas about products needed or innovations or technologies it could apply to creating greater business value.
The New Information Tools
What we need, therefore, is to develop new tools that can be applied to a wide variety of applications and that many different people can access and use easily. And that’s what CIMS is doing in our partnership with IBM’s jStart Emerging Software Technologies team. The tools we’re focusing on in this particular context are language analysis tools. One is called IBM Content Analyzer and it’s available through the Cognos Business Group. It’s bundled with software called LanguageWare. The other piece of software that’s still in development for massive data processing is called Big Sheets. Much of the data we have is text. When somebody writes something or fills out a form, it’s all text that is kept in a very structured format with predefined key words. Consequently, you are limited in how you can look at that information.
With these new linguistics tools, however, we can look at any sort of data. It can be text, it can be numbers, it doesn’t matter. We can look at it andset our priorities for what we want to see, and the tools are powerful enough to break a sentence down into nouns, verbs, adjectives, and adverbs, and allow us to build relationships within a sentence, a paragraph or a page of information.
People often assume this is analogous to a Google or a Yahoo! type of search. Well, yes and no. When you conduct a search, you’re using the priorities, which that service has established. It has rules about how it posts information and how it evaluates information. However, with these new tools, you determine what the rules are, you determine what’s important, and therefore you’re able to ask questions that are specific to you.
So, while a search using a Google or a Yahoo! is very quick and gives you a million hits, the issue is how you interpret what’s in those million hits. The new linguistics tools take a little longer to use. But they collect information from any unstructured or structured source you point it to. They can read all the information — a webpage, a financial page, whatever —and import it into a format that allows you to ask questions of it.
So instead of having a million hits, perhaps you get only 50 hits but it’s the 50 hits that you’re most concerned with. You don’t have to spend all that time going through the million trying to find that signal in all the noise.
We believe this technology is a gamechanger. It’s disruptive and those companies that are first to adopt it will win the game.
The tools that IBM is developing and CIMS is applying will allow more and more people to access that sort of information without having to pay a service and with the capability of designing the search to meet their own criteria. If you think about innovation as bringing together different ideas to solve a problem, these tools allow you to do a more elaborate sort of siliconbased research: You can ask questions,bring information together and, perhaps, bring parties together who can develop new technologies and innovations.
For example, NC State University has developed a new technology for delivering drugs through inhalation —a “Smart Inhaler.” In the past, the Office of Tech Transfer would have had to contact every drug company it could in order to determine their interest in licensing the discovery. But with these new tools, we were able to go into a database and ask, “Which drug companies have recently had a failure due to a drug delivery device issue and might therefore be interested in our technology?”
So now we are able to conduct a much more directed search that removes all of the noise and gets you a good, strong signal.
Objective Search Strategy
Along with the tools, however, it’s essential to have a process in place for conducting the search; specifically, an objective search strategy that allows you to ask the relevant questions: Do you look at your processes? Do you look at your customer feedback? Do you look at that information and develop strategies based on it, and do you do it in a timely fashion? Can you be predictive, say, about where the market is going or what changes you have to make?
In addition, you need a workforce you can train to ask those sorts of questions. It should be part of their job requirements to ask, “How are we doing and what sort of information do we have to validate whether we are doing well or not so well?” When things start going south on you, it’s generally too late to save your business. So you want everybody constantly looking at this information, building a culture of continuous improvement by working around the data and developing new strategies bottom line: You want to get out of a command-and-control environment and more into that coaching environment where everybody can look at how the plays are running, so to speak, on the field and can make adjustments. These new tools and techniques will give us that ability
CIMS Industry Fellow,
NC State University Virtual Computing Lab