Tools to Cope With the Glut of Digital Information Available
Although there are many tools available for us to use in a new model of learning delivery that we suggested in the December/January issue of this magazine, I’m going to cover three that you might find interesting: “Wizard Directories,” “Embedded Training” and “Federated Search Engines.”
One of the newest tools emerging in the 50 percent category is a result of the Internet, social networks and instant messaging. To illustrate, I’ll relate a story that a friend of mine told me about a new employee he had hired as a graphic artist. In only a few short months, the young lady had become of his best employees. She was turning out stunning designs, excellent copy, and her work was receiving praise from a lot of my friend’s customers.
So one morning, he decided to walk over to her work area and congratulate her on her fine work.When he got there, he was surprised to find that she had 10 Instant Messaging (IM) sessions open on her desktop. He looked at the array and said, “Who are these people? Are you talking with your friends at work?” She turned and said, “Actually no. This is my Digital Tribe. I come to work with them every day.”
Totally dumbfounded he pleaded, “Please explain.” She explained that she had come to know them from comments they posted on technical forums or from documents they published on the Internet. Some were good at graphic design, while others were good at copy. Some had an in-depth knowledge of Photoshop and Illustrator, while the rest were expert at other parts of her job. They simply helped each other every day. One supplied design advice, while another helped to untangle and simplify copy. Her own contribution was an in-depth knowledge of Corel Painter. They, in fact, had created a virtual community of practice, or, as Napoleon Hill would have described it, a “master-mind” group that helps one achieve success.
We thought this mode of collaboration was excellent, akin to people going to forums to collaborate with their peers. So we took this notion of a “Digital Tribe” and began talking about how to use it to implement a directory of wizards in our IT community of 5,000 people.
We began by providing a listing of the technologies the community was using across the enterprise.We then asked a pilot group to rate their knowledge in each of those areas on a scale of 0 to 10. (You might want to hide those ratings for a variety of reasons, by the way.) The purpose was to be able to quickly search for people by their level of expertise in various areas, in order to put them in touch with people who needed their expertise. In other words, we were giving our organization the ability to build their internal “digital tribes.” If you’re a small organization, then you will more than likely end up with only one or two people
who are the experts in a given category. However, when someone needs a quick answer about direct deposit, or how to give CPR to a fallen worker, or how to fix a transformer that’s been hit by lightning, an automated Google-like search of the Wizard’s database will quickly identify the people that can best provide an answer.
Our eventual plan was to provide a MySpace or LinkedIn-type page where employees could post other job-related data for co-workers to view. That information might include coding tricks, links to useful information or knowledge, tips, examples of their code, or whatever they felt would be appropriate on their personal page, including their personal credentials and perhaps some fun facts as well. This is highly oversimplified, but you get the idea. If I was looking for an HTML person or I wanted to assemble a team, I could easily search the Wizard database using a simple search engine.
One of the best tools we developed was the “federated search.” Think of this as a “mothership and her fleet.” Basically, we handed the key search term to the mothership, which in turn passed the information on to a fleet of highly specialized search engines. Some of these engines were good at searching books in our digital library provider, while others were tuned to search code libraries or the Internet. We then brought the search results together from all the engines and allowed the user to select from the various sources — the Internet, code repositories, digital libraries, and eventually, wizards who were knowledgeable in the area they were searching. Phase I of that implementation only took a couple of months and was an instant hit.
The real trick was to make key wizards appear in the search results, so here is what we were thinking about and experimenting with in that area. Generally, we had to convert tacit knowledge — the knowledge that we walk around with in our heads — into a searchable commodity. We looked at two ways of accomplishing that seemingly impossible feat.
One was very intrusive, using a similar methodology to that of search engines. Basically, the technology could read through everyone’s e-mails and count the number of times a person used any word or phrase in his or her communications. The more the person referred to a topic, the more likely it was that the person could be assumed to be proficient in that area. The software ranked all of the words they used, throwing out adjectives and the like, and ranked them from most used to least used. Impressive, but our heads were reeling around privacy issues, ownership of a person’s e-mail content, and the other implications of employing such a tool.
Instead,we chose a less intrusive way, which was to let people self-declare their knowledge in a topic area. In the IT world at my former employer,we hadmore than 2,000 technologies in use across the enterprise,with more than 5,000 IT professionals.What we did in a small pilot was to ask the people to declare their expertise in one ormore of the categories with which they were familiar. So if they were familiar with HTML, they might rank themselves at the 5 level on a simplified scale of 0-10. If they had no knowledge in an area, they ranked themselves at 0.
However, we thought that one more data point was needed to help neutralize the over-aggressive self-rating personality from the timid personality. So we asked for the people to express their experience in terms of years. Thus, if a person declared
they had 2.5 years of HTML experience and a 10 level of expertise, we would multiply the two numbers and assign them a ranking of 25. Similarly, a Webmaster
might rank themselves a 9 in the HTML discipline, and show 12 years of experience, for an overall ranking of 108. As a search engine might measure a site’s popularity by the frequency of certain words used on the site as well as the site’s popularity
by measuring how many other sites linked to it, we now had a way to rank our wizards for a specialized search engine in our federated search engine fleet.
However, we had to overcome one more problem: People didn’t perform searches using domain words like “Java” or “HTML.” Instead, they used words or phrases that existed within that primary domain of knowledge. So a person who wanted to make text move across the page might search for “marquee html code,” a sub-domain of knowledge within “HTML.”
The elegantly simple solution to that problem was suggested by one of our vendors. The suggestion was to take the indexes from the books in our digital libraries — such as “The Complete Book of Java”— and strip the page numbers from those indexes to create the possible search terms. The result was a universe of terms we could associate with each primary domain of knowledge — in our case, the 2,000 technologies in use. So basically, when the mothership handed out a phrase like “marquee html code,” the people search engine had to find that phrase within our library of index terms, which in turn was linked to one or more main domains of knowledge — in this case, HTML. In turn, those main domains of knowledge were then linked to the people who possessed that knowledge, ranked from most familiar to least familiar (a self-rating of at least 1). The strongest candidates would appear first, and others would fall under them by their self-ranking in that domain of knowledge. So from most any search term, we were able to link to a wizard.
There were other problems as well. We had to make sure that we didn’t show the
same person all of the time, to avoid distracting that person from the performance of their his or her job. That turned out to be less of a problem, as most people preferred to avoid the most experienced person when they had a simple question, or simply found their answers on the Internet or an internal knowledge repository.
When we finished the pilot, our “federated search” engine took the top 10 results from each of the (fleet) search engines, which now included people, and presented them to the person looking for information. It was up to pursuers of information to choose where they would go for their answers. The engine performed very admirably and compared in response times to typical search engines. And, as you can imagine, the entire IT department raved about the engine, which was nicknamed “The Fed.”
Another useful tool was based on a product that was originally developed by my team at PeopleSoft (now Oracle).We were wrestling with the problem of the amount of time it was taking to train end-users in our various product modules.
The solution we eventually used was to embed the training into the PeopleSoft application itself. Using a team of curriculum developers and a third-party provider, we looked at all of the job roles that worked with each of our software modules. As expected, only a small handful of application screens were used consistently in the performance of each person’s job. So our plan was to focus on only those screens in the classroom part of the training.
We asked our product developers if they would reserve the F1 key for us to use.What we then did was create a simple program that looked at the label on the screen when people clicked the F1 key. If they were looking at the cash application screen, we simply went out to a repository of information that dealt with that function and showed them how to fill out the screen, where the data came from, and where the data was going. Knowing that the applications would be altered by every corporate user, we used a simple Word document to house the standard training. Not only did that make it easy to change, it also meant that we could incorporate audio or video files into the training materials.
These new training modules were eagerly adopted by our user community. Instead of pulling people out of their jobs for weeks, we now could train users in a specific job role in a day or less, and then teach them how to invoke the rest of the training using the F1 key.
So now we have yet a third way to move training out of the classroom and into the workplace. And there are many more to talk about once we make this shift in philosophy to a “working proficiency” and to thinking about how to make people more productive and proficient in their jobs. The technologies for moving training into the workplace — wherever that might be — have never been more bountiful.
The potential for reducing the amount of time that people spend searching for information (one-fourth of their day) provides us with a very attractive ROI to accomplish this shift. This new approach in thinking about proficiencies and curriculum delivery strategies will take several years to implement and will necessitate the cooperation of vendors who need to embed training into their products. However, the value of this approach goes beyond the education and training organization. It most certainly helps to position companies and individuals in a new economy where people are wearing multiple hats because of downsizing, and where organizations continue to struggle to survive.
As with earlier moves away from the classroom, this move will face challenges on many fronts. This model proposes to eliminate the primary dependence on a 2,000+-year-old model of classroom instruction.We expect to have to set goals carefully with departmental employees, industry experts, academia and other influential parties.
The ways we measure success will also have to change dramatically. Our measures will no longer be simply student days or similar criteria, but we also will have to measure our ability to provide information very quickly in order to avoid long interruptions in a person’s productivity. That promises to be a very interesting debate.
Just as it took years to get the Internet to where it was a viable way of doing business, it will take time to move to the notion of dealing with “essential knowledge” and a “working proficiency.” Many discussions will need to take place with the human resources department, corporate information security, legal and other groups that might be impacted by information coming and going from both
inside and outside of the company’s firewall. When we are successful, the rewards
will benefit both individuals and corporations. The gains in productivity for corporations will allow them to move even faster in the marketplace — or, more simply, survive in these trying economic times.
—Joe DiDonato is an independent consultant who has worked with technology providers like Oracle and major corporations like Countrywide Financial. Contact him via e-mail: firstname.lastname@example.org.