This article was published on October 10, 2017

Using complex analytics to create better grading systems for e-learning

Using complex analytics to create better grading systems for e-learning
Andrei Tiburca
Story by

Andrei Tiburca


Andrei is an avid marketer and content creator. You can connect with him on Twitter(@AndreiTiburca) or read his articles on www.webdesignled Andrei is an avid marketer and content creator. You can connect with him on Twitter(@AndreiTiburca) or read his articles on where he's an editor.

Training is often one of the biggest expenses for businesses, and that’s especially true in times of rapid and radical technical change like we are experiencing now.

With business needs changing quickly and technology advancing even faster, companies are constantly reevaluating how they spend their training dollars,” asserts David Wentworth, the principle analyst at Brandon Hall Group. “It also can be difficult to determine just how big the training function should be related to a company’s size and the industry in which they operate.”

Ad budgets, naturally, can vary widely, but for a mid-sized companies, Wentworth’s team estimates an average training budget of around $3.7 million per year.

Better Grading Systems

Yet it’s not the size of the budget that matters, but how effectively it gets used. And for many companies, training fills the skills gap in the current workforce.

This is true whether you’re looking to train new staff or attract new talent with specialized skills. “The hardest positions to recruit for are technical – and that outlook isn’t changing anytime soon,” explains Matthew Bell, a multimedia learning designer who has worked on training solutions for Salesforce and Xerox. “By 2020, the Bureau of Labor Statistics estimates there will be a gap between the number of open technical positions and skilled workers able to fill them of nearly 1.4 million.”

“New hires can be risky, too – they need training, on-boarding, and adjustment periods, and even then they may not end up being a good cultural fit,” Bell continues.

As companies increasingly turn to elearning solutions for technical skill training, ROI becomes especially critical. For many organizations, the answer is to create training – or utilize third-party elearning options that can teach these skills to existing staff. Yet elearning courses are expensive to develop.

“An elearning course ends up being the most expensive asset to create, requiring an average of 106 hours to put together, which results in an average cost of $53,742,” Wentworth notes. It’s not the bottom line cost that should concern company leaders, but the return on investment. Yet proving ROI presents an elusive challenge, for many reasons.

A lot of it comes down to employees not being able to apply the skills they were certified in – or said they were certified in. This, and other factors like the implosion of several major coding academies, have people questioning the value of online learning, and how to quantify it.

“Understanding an individual’s performance in cloud environments has unlocked the ability for us to tailor training to that person’s skillset,” Linux Academy Founder and CEO Anthony James told me. “It also allows us to give organizations insight into employee and potential candidate technical abilities that they’ve never had before.”

The robust yet rocky state of the elearning industry

The elearning industry is gargantuan and growing fast. Data from Global Market Insights pegged the industry at USD 150 billion in 2016, and predicts it will continue to grow by more than 5% a year for at least the next seven years.

Better Grading Systems2

Along with that growth comes consolidation. A few years back, new elearning course marketplaces and institutions – especially those offering technical skills – were launching in droves. Now the industry is settling into a more sustainable place, with mergers, acquisitions, closures and pressure on the better ones to differentiate. Yet the industry as a whole is suffering from some serious PR damage.

The fake universities phenomenon – institutions whose credentials are not legally accepted but which succeed to trade and exploit the desires of people trying to gain qualifications – remains a huge challenge. And the highly public Trump University fraud allegations brought the problem front and center.

Although some of these institutions may offer an element of education, they undermine the validity of genuine qualifications. The solution, according to many in the industry, is to improve elearning grading systems and provide a richer insights into the specific skills students acquire in training. And that’s a massive challenge.

It’s fairly clear that our current mainstream educational grading systems provide a complete view of neither the students themselves, nor what they’ve learned. For Linux Academy Founder and CEO Anthony James, this recognition led to some serious soul searching.

The academy, which offers a range of technical training with a heavy emphasis on Linux coding, articulated the challenge straight on.


“We looked at the toughest challenges organizations face for cloud training and hiring, and then applied our expertise to come up with a solution,” James recalls. “We want to change the face of how professionals train for cloud-based skills and how organizations find, validate, and train talent.”

Using big data analytics to measure knowledge

What do grades really tell us about the students taking them and about an education program? How can we better determine one’s retention of knowledge and skills?

Test scores, which fuel most grading systems, can be misleading. There is a lot of evidence suggesting that standardized testing systems are seriously flawed – and that they can be cheated. On the other hand, many of the world’s most successful people did not receive good grades at school, or have even dropped out so that they could achieve more. What does this tell us about the connection between grades and professional success?

This type of soul searching eventually led to Linux Academy’s rollout of Cloud Assessments, a virtual workplace that allows students to see if they can apply their technology skills in the same setting they would see on the job.

The idea presented a complex data problem – there were a host of factors to track, parameters meant to measure not only if students can solve the problems, but also if they are doing so efficiently, at a professional level. The system also needed be sophisticated enough to take into consideration individual factors like students’ background, employers, time spent on tests and the number of attempts it takes to pass each “Quest.”

All of this, of course, means mountains of data, and the inherent problem of not just analyzing it, but simplifying it to be useful to the student. To meet the challenge of this huge data problem, James partnered with data analytics and business intelligence firm Sisense, to pull the insights from their metrics and gain a clear picture of student performance.

“Their tools were critical in helping us gain a real-time understanding of individual student performance, and they provided actionable analysis in aggregate across student types and cohorts,” James explained.

At Linux Academy, they’ve applied these ideas to the training they offer through their own business, with virtual labs meant to duplicate a real-world work setting. All of that, of course, comes with a massive amount of complex data; both Cloud Assessments and the online training through Linux Academy make use of Sisense Pulse, which generates immediate alerts when certain metrics go beyond normal ranges – if Quest failures spike, for instance, which may indicate a problem with the content.

Unlocking superior e-learning value

This kind of insight has huge value to organizations. The data can be easily shared with – and understood by – employers so they can get a sense of the real skillset of both current employees and potential hires.

It can also help preserve the integrity of training budgets. Many companies provide bonuses and free application fees for employees who pass industry recognized certifications. Evaluations can be used before paying for training, by analyzing the student’s current skill level on a topic. For the students themselves, these appraisals offer a tangible way to explore their own skills and see what further training they might need to meet their career development goals.

By leveraging student training data in this way, student appraisals are poised to make an enormous difference in elearning, both for employers, and for students. In addition to helping employers make better hires and see better ROI from their training budgets, this is a huge benefit to students, helping them understand if what they’ve learned is something they’re actually ready to apply in an operational setting.

As we begin to settle in to the age of big data, we’ll see more companies learning how to truly leverage that data into insights like these. And as this becomes more widespread in the elearning industry, we will begin to see more platforms differentiating themselves in meaningful ways.


Get the TNW newsletter

Get the most important tech news in your inbox each week.

Also tagged with