!-- Primary Navigation Mobile include start -->


In my colleague Anita Delahay’s recent excellent post, she outlines how Argos thinks about the failings of courseware and what we intend to do differently. That said, the title of her post is “Courseware 3.0” and the first sentence asks, “Why are we still talking about courseware?”

It’s a vexing question. For starters, it’s one of the worst product category names in the history of bad product category names. Worse even than “Learning Management System.” “Courseware” sounds like “Tupperware.” It does not exactly evoke a quality learning experience. And yet, roughly 30 years after the first use of the term that I can find, we don’t have anything better.

We could go with “digital textbook replacements.” That has a nice ring to it. Right? Don’t you want one of those?

For a while, there was a strong, highly misguided, and unintentionally revealing push for “personalized learning” as a supplemental or alternative product category name. Before we get to the substance, let’s stick to the branding for a bit. Is “personalized learning” a thing you can buy? And if it is, would you expect to get it from a product called “courseware”?

After three decades, we still don’t have a better name for a clearly defined thing with a clearly defined educational value proposition. That should tell us that our thinking about it is stuck. We’re in a cul-de-sac. Our collective idea of the value of this product category is fundamentally off in a way that is preventing it from moving forward. Specifically, the framing of courseware is as software that teaches the course. It is implicitly sold as automation that replaces the judgment of the educator.

This is the wrong way to think about the product. First, there is scant evidence that software can do the educator’s job in this way. Second, the product is not selling. While I wouldn’t go so far as to call courseware a failed product category, many of the stand-alone courseware platform companies have folded while sales of courseware from incumbent textbook publishers have been lackluster. They have not been strong enough to turn around the flagging fortunes of these companies. Worst of all, the “courseware” framing fails to recognize educators as knowledge workers. It squanders a precious resource—teachers—while insulting the very people who are charged with selecting and implementing the product.

At Argos, we do not consider ourselves to be a “courseware platform” company. The product category we are creating doesn’t have a name yet. In this post, I’d like to explain why we reject “courseware” and intend to kill that word, even if we are forced to use it in the short term.

The origins of courseware

As far as I can tell, the first use of the term “courseware” was by MIT in 2000 when they launched OpenCourseWare (OCW). In 2001, The New York Times ran an article with a headline that blared, “Auditing Classes at M.I.T., on the Web and Free“. The truth is that OCW at the time was a collection of class websites that mostly contained lecture videos, syllabi, and handouts. The audacious aspect of it had nothing to do with teaching and learning. “On the web and free” was the part that grabbed the attention.

A decade later, Carnegie Mellon University (CMU) launched its response to OCW in the form of the Open Learning Initiative (OLI). OLI was founded on the notion that teaching is more than broadcasting information. It is interactive and can be based on the sciences of learning. So the OLI team went about building a platform for such experiences as well as a catalog of titles that could run on it. Importantly, they continually tested and improved those titles. The platform was designed to enable testing for its effectiveness. That was—and still is—Carnegie Mellon’s signature approach to education: applied learning science (or “learning engineering”). While there was a certain amount of “educate the world for free” competitiveness with OCW, the part that CMU was most excited about opening up was the learning science. Note the (undoubtedly deliberate) contrast in brands. Where OpenCourseWare was about “courses,” the Open Learning Intiative was about “learning.” Where MIT was focusing on the “ware”—the products—CMU was focused on the “initiative”—the collaborative process.

So from its inception 20 years ago, OLI, upon whose platform Argos is built, positioned itself against courseware. OLI has always been about the process of learning to become better educators together. Yes, the OLI platform and curricular products look a lot like what we call “courseware.” But if you focus on the artifacts rather than the process, then you will miss the point of OLI.

Selling little boxes

Despite the contrast with OCW, it’s also partly fair to say that the new element of OLI was “for free, on the web.” Academics, including but certainly not limited to those at CMU, had already been experimenting with software that creates learning experiences for some time and using those experiments to learn about learning. Over time, this research began breaking out into the commercial world either as direct spinoffs or as products that were inspired by the trend and that could displace the paper textbook with something better. A partial list includes Carnegie Learning (a spinoff from CMU) in 1997, Cerego in 2000, Cogbooks in 2006, RealizeIT in 2007, Knewton in 2008, Smart Sparrow in 2010, Zybooks in 2012, and Acrobatiq (another CMU spinoff) in 2013. The big textbook publishers started getting into the act around 2010. When I was at Cengage, I worked on one of the early entrants into this category, MindTap. Wiley+, McGraw-Hill Education Connect, and Pearson Revel are also part of this wave.

These platforms varied wildly in their approach to course design and occupied various points on a spectrum from empty boxes to black boxes. On one end, the empty boxes were simply platforms for building learning experiences according to the principles that the platform developers thought were “best.” On the other end, the black boxes were rigid publisher content. While they might have had some good learning experiences in them and even been “adaptive” in one sense or another, they might as well have been been a bound paper textbook. Educators had little control or even visibility into what the software was doing and why it was doing it.

Regardless of where the products fell on that spectrum, they all pushed increasingly toward prescribing one “best way” to teach. This problem was ironically made worse by the efficacy movement that was started by Pearson and quickly gained traction with both product companies and non-profit funders. Since the products were designed for teaching in a particular way, the way many efficacy advocates began thinking about measuring product efficacy was based on whether educators followed the “implementation model” for which the product was designed. That’s a reasonable way to think about testing the conditions under which a product will be effective but a terrible way to think about the ways in which an educator might choose to use the product in a real classroom with real, human individuals as students.

Unfortunately, the language of research has infected the product thinking. I have been in conversations at multiple publishers in which they lamented the difficulty of getting educators to follow the “proper implementation model” for their product. That’s what teaching has come to be thought of in this paradigm. An implementation model for effective use of a product.

This is a slippery slope toward an Industrial Revolution assembly line vision of teaching. The educator strictly follows a prescribed formula while the machine provides the “personalized learning.” The most infamous articulation of this view came from a former courseware platform provider who described his product as “a robot tutor in the sky that can semi-read your mind and figure out what your strengths and weaknesses are, down to the percentile.” But I’ve heard grant managers at foundations talk about “teacher-proofing” education. This, ultimately, is what “courseware” is imagined to be. It’s software that teaches the course. It’s a robot that displaces a human worker.

Modern Times (1936) Directed by Charles Chaplin Shown: Charles Chaplin (as a factory worker)

Is it any wonder that the educators who actually get to decide whether they will adopt these products are not exactly stampeding to do so? Or that, when they do so, they don’t eagerly follow the “implementation model” under which the product is supposed to be effective? Or that in 30 years, nobody has come up with a product category name that might be more appealing to educators?

There is a better way.

Knowledge work

Ironically, even as courseware and its proponents have been pushing education toward the Taylorism management philosophy of early 20th-Century industrial production, industry has long been moving away from it. As early as 1948, Toyota began developing the Toyota Production System (TPS), a new approach to manufacturing management that enabled it to overtake General Motors as the largest car manufacturer in the world. While TPS is most widely known for developing the just-in-time supply chain, it also was revolutionary in that it devolved management and empowered workers on the factory floor to creatively solve problems on the spot. In fact, many of the principles of TPS line up very well with those of Agile software development. At Toyota, factory workers are knowledge workers.

Today, that view is practically a no-brainer as repetitive, unskilled work is being performed by robots and semi-skilled work is increasingly being performed by AI. The future of work is knowledge work.

So why on earth would we prefer to entrust the teaching of future knowledge workers to robot tutors over educational knowledge workers?

Even that phrase—educational knowledge worker—is redundant. Education is knowledge work. And please don’t tell me that high school graduates on Toyota’s factory floor can be knowledge workers but adjunct instructors with advanced degrees can’t.

“Courseware” as a tool for enabling classroom knowledge workers

Argos’s mission is to empower educators to become more effective knowledge workers in the classroom. What would that require? What do industries do to empower their knowledge workers?

  • Rely on their judgment: The value of knowledge workers lies in their human judgment. If we are not constantly cultivating it and trusting it, then we are not getting value from our knowledge workers. Any curricular product model that does not actively cultivate and empower educators to make educational judgments is moving in the wrong direction.
  • Cultivate a culture and vision of common goals: Organizations that excel through knowledge work invest energy in goal-setting so that all members of that organization can apply their individual talents and judgment toward achieving those common goals. Any curricular product model that treats educators as isolated individuals by failing to facilitate dialog and cross-pollenation is moving in the wrong direction.
  • Encourage empirical experimentation: We cannot improve without both trying new approaches and measuring how well our experiments worked in achieving our goals. Any curricular product model that fails to support both classroom experimentation and testing of results is moving in the wrong direction.
  • Motivate sharing of new knowledge: Improvement and innovation should be shared, as should newly identified problems. This is how collective improvement (or improvement “at scale”) works in an effective 21st-Century organization. Any curricular product model that fails to both enable and reward the sharing of teaching craft is moving in the wrong direction.
  • Provide performance support: Time spent on mindless tasks is a waste of a knowledge worker’s time. In every other industry, we optimize workflows of valuable employees to get dumb, time-wasting distractions out of their way. In tech-enabled teaching, we waste a lot of educator time because we’re stuck with rigid product category boundaries based on relics of analog teaching models. Any curricular product model that isn’t focused on reducing time-wasting tasks—including fiddling with and integrating technologies—is moving in the wrong direction.
  • Provide decision support: Software can bolster rather than compete with human judgment. It can provide valuable and timely information and, increasingly, make useful suggestions based on patterns that the knowledge worker missed or doesn’t know how to respond to. Any curricular product model that fails to provide educators with insights that they find useful for making educational decisions is moving in the wrong direction.

Argos rejects robot tutors in the sky. We reject fetishizing the software. We vehemently reject the notion that educators should be afforded less respect for their time and skill than knowledge workers in every other industry. And we reject the antiquated and counterproductive notions that are embodied in the word “courseware.”

It’s time for something new.