Lessons in content management

Thomson Learning company NETg provides e-learning content, technology and services to 5,000 companies around the world, including Intel, Barclays, Nestlé and Experian.

This consists of multimedia courses in around 2,000 IT and business-related topics; technology books (including 50 e-books); boot camps for accelerated IT certification; and executive education, including an MBA degree organised in association with Cardean University.

However, the bulk of NETg's business is the delivery of courses over the web. These range from tutoring in software applications, such as Microsoft Office and Lotus SmartSuite, to professional courses in team leadership, stress management and business ethics.

Driving all this is an online database currently containing 85,000 learning objects, each of which is indexed and regularly updated so that it can be seamlessly blended and delivered to millions of desktops around the world.

What is a learning object? Jim L'Allier, chief learning officer for NETg, describes a learning object as a "specific and measurable objective".

It contains "a learning activity, generally between three and five minutes in length, designed to allow the learner to achieve that objective, and then an appropriate assessment to determine whether or not the objective has been achieved".

From a technological point of view a learning object will contain multiple components, including video, audio, sound and text files, screen backgrounds and so on, all of which need to be combined into a single package.

Since many components are standardised it makes sense to reuse them wherever possible. The first content management challenge, then, comes during the object creation process: content creators need to find and access useful pre-existing components, rather than recreate them.

"A good example would be a screen background or screen capture, which will be used in many different contexts," explained L'Allier.

So it is important that, once created, all components are effectively indexed, catalogued and described.

As a content provider, NETg's most pressing challenge is to ensure that it delivers the right content to the right user at the right time. To do this it has created proprietary 'XtremeLearning' technology. Once logged on, learners choose courses from a menu of subjects pre-selected by their human resources department.

Before beginning a course, the user will take a pre-assessment questionnaire. "This draws on a question database associated with each learning object, and determines what the learner already knows," said L'Allier. "We can then present them only with the objects in which they have not demonstrated a competency."

Users can also seek out discrete objects for a specific learning purpose using a search function. "If, for example, you are building a spreadsheet and you find yourself needing to know how to do 'net present value' in Excel, you can search through the database of objects and pull out one that deals with that function," explained L'Allier.

In this respect, NETg's model is akin to a traditional online host, such as Dialog, rather than a distributed web-based service. That is, it operates as a data warehouse in which relevant information is located by means of proprietary search technology.

"You don't have to put in Boolean operators, but just type in a natural language question. Searching is done mainly on keywords, using neural network technology based on a Bayesian algorithm," said L'Allier.

In addition, NETg's objects retain and exploit knowledge about how they have been used previously, giving them the intelligence to assess their relevance to users' queries.

"Each object has a history, and a memory of how it was used against a certain class of problem," said L'Allier. "When a learner searches on a particular problem, the object may say to itself: 'I have seen that problem before, and I estimate there is a 95 per cent probability that I am the solution.'"

This concept of building intelligence into the data as well as the search engine is a step beyond the Dialog model, although it could be argued that the classification codes used in Dialog files provide a passive kind of intelligence.

Moreover, unlike Dialog, NETg's clients can use a Visual Basic-type object editor called NLO+ to customise the way in which learning objects are presented. Training managers can, for instance, add a bespoke splash screen to course material with the organisation's logo or colours.

Additionally, human resources can go into NETg's database of learning objects and rearrange, delete or disable individual objects using NLO+.

"If, for instance, your organisation uses Microsoft Outlook, but certain features are not used in the company, the training administrator could go into that course and turn off objects offering training on those features," said L'Allier.

Early this year, Canadian telecoms company Telus began licensing a mixture of management and technology courses from NETg, with the aim of making them available to around half its 30,000 employees.

The management process at the customer's end couldn't be simpler, according to Josh Blair, vice president of learning services at Telus. "All we do is select the courses employees can take, manage the assignment of IDs and passwords, and track their learning," he said.

And since NETg automatically notifies the training department when an employee completes a course, tracking employee activity is a cinch. "We just have to update the employee development record when notified," explained Blair.

He particularly values the ability to use the NLO+ editor to customise courses. "Many of the courses NETg offers take maybe four to six hours to complete," he said.

"Our experience is that there is more value and uptake in e-learning if you can break the learning modules into small subsets, so we build 15- or 30-minute courses.

"We may also customise it with Telus information. If the course is on ADSL, say, we include some information about our ADSL roll-out strategy."

The data warehouse model means that the management at NETg's end is also relatively easy, although as the service develops it is becoming more challenging.

For instance, rather than reaching out over the web and accessing courses from NETg's central server, some customers are asking to host data on their corporate intranet. And, as more customers customise courses, version control is becoming increasingly complex.

"We now have multiple sites creating and editing multiple learning objects all coming back into a repository of generic objects. We have to tease these apart as they go out to the desktops of users, which means keeping track of a constant stream of objects," said L'Allier.

Customers' demands for ever-greater flexibility also make management difficult. "We realised, for instance, that customers want to access learning objects from both NETg and competitors such as SmartForce in such a way that objects can be combined," he said.

Likewise, users want to be able to mix and match their own learning objects with those belonging to NETg. But these developments raise questions about the use of proprietary technology.

Today, for instance, Telus licenses e-learning courses from NETg, SmartForce and SkillSoft, although the latter two have recently merged. In addition, the company creates its own e-learning courses using technology from click2learn.

The problem, according to Blair, is that all these approaches use different technology. "If we built a course on NLO+, and were then to sever the relationship with NETg, we would lose our own courses since they would only work in the NETg system," he said.

As a result, NETg has to think about more open technology solutions. It also anticipates the need for a way of identifying the ownership of individual learning objects. NETg has no solution to this at present, but is working with the Advanced Distributed Learning (ADL) initiative.

Established in 1997 by the US Department of Defense, ADL is now a collaborative effort between government, industry and academia aimed at creating an interoperable distributed learning environment.

This has led to the creation of the Sharable Content Object Reference Model (Scorm) which provides a tagging framework for enabling the interoperability and reusability of web-based learning content.

Among other things, the specification includes a dictionary of tags aimed at describing learning content in various ways, including content description, ownership, price (if applicable), technical requirements and educational purpose.

Effective use of standardised metatags could provide greater flexibility in billing, explained L'Allier. Today, for instance, users have no choice but to license libraries of courses.

"Our clients continually point out that there is an inherent contradiction in our offering a modular object technology that cries out for pay-per-use, but which we treat as a library when selling," he said.

Blair indicated that pay-per-use would be very attractive to Telus. "As with any such relationship, we would need ceilings to protect us, but it would have the attraction that we only paid for true usage, as opposed to potential usage," he said.

However, pay-per-use billing is as much a business model issue as it is a technology one. "It's a question of how we disconnect the current pricing structure tied to libraries, and reconnect to a new one at an object level," said L'Allier.

As content providers like NETg make greater use of meta tags, so the flexibility with which learning objects can be exchanged, combined and sold will grow. For some, the end game is to make learning objects sufficiently autonomous and self-defining that they can be set loose into the great pool of data on the web.

Individual users, or intelligent agents, could seek them out and purchase them, using perhaps credit cards or e-cash, on an ad hoc basis.

Such a scenario, however, is not currently on the horizon for NETg. Before it took that path the company would need to give serious thought to digital rights management (DRM) technology. But, NETg appears surprisingly relaxed about copyright issues.

Today it makes no use of DRM, although with users able to save courses to their laptops for offline viewing, presumably there is little to stop them emailing the content to unlicensed users.

"That's a worry that is always going to be with us," concluded L'Allier. "It was a problem we faced when we were selling courses on CD-Rom, and it's no different when you sell individual learning objects."

Richard Poynder is a freelance business and IT journalist.

Like this story? Spread the news by clicking below:

 del.icio.us     Digg this     reddit!