I’ve been thinking a lot lately about beating the odds. Not by being lucky or tough or heroic, but by recognizing the errors that we make when calculating odds of success based on assumptions and pattern matching. Assumptions can be wrong. Patterns might not apply. Inevitability is often a wall with cracks in it. Sometimes, if we can see the cracks, we can find our way through to the other side. This is important to remember at so many difficult moments in life, including when we try to change a system like education that seems unchangeable.
A dear friend recently told me about one of Instructure’s secret weapons in beating the seemingly overwhelming odds that they would become just one more failed LMS startup on the giant trash heap of failed LMSs. It was one I hadn’t heard about or noticed before. This surprised me because I have been a student of Instructure’s secret weapons. Like many other students of EdTech, I was convinced that an LMS startup breaking into what was then a static market was virtually impossible. I was taken completely off-guard when they not only survived but eventually knocked Blackboard out of the top spot.
So I spent a lot of time studying Instructure to understand what I had missed. I learned many lessons. The company leadership didn’t have just one silver bullet. They employed many strategies and benefited from the luck of timing. I have written multiple posts about these factors. I thought I had cataloged them all.
I was wrong. I missed a big one. One that I could have seen based on evidence that was accessible to me at the time. As a species, we gain so much value out of our ability to pattern match that we sometimes have trouble seeing beyond our confirmation bias to see that the “odds” we are calculating are based on a pattern that doesn’t apply.
What—or who—is the deciding factor?
At the time that Instructure came on the scene, LMS selection decisions were led by, and often made by, CIOs, who compiled the lists of requirements and had a great deal of latitude to impose their will. At that time, most LMSs were hosted by the campuses and the IT folks could always come up with some technical reason that other stakeholders were not in a position to dispute. LMS sales processes, therefore, were long and expensive affairs that involved a lot of checklists and, often, wining and dining of CIOs. The LMS company that had the best CIO sales force, which was a capital-intensive capability, would win most of the time. Blackboard had the money and a really good sales team that was finely tuned for those sorts of enterprise sales. Therefore, they won a lot of sales. It seemed like an insurmountable barrier, especially for a small start-up that was known for snarky T-shirts and PR stunts whose primary audience was clearly not CIOs.
In retrospect, it should have been obvious. Instructure deliberately ignored the existing LMS selection structure and created an alternative one. They saw that the percentage of faculty who were making meaningful use of the LMS was rising. They saw the frustration among faculty with the teaching tool that had often been chosen by the IT staff. And they understood the power faculty had on campus if only they could be activated.
Instructure fomented faculty rebellion. When they set up a party across the hall from the famously over-the-top BbWorld conference and gave out T-shirts saying, “I cheated on Blackboard with Instructure at BbWorld,” they were targeting faculty who, by and large, had a very different reaction to a conference that felt modeled after Oracle’s than their campus CIOs did.
As Instructure gained traction, evidence mounted that strategy became more than just a PR campaign—if you were open to seeing the signs. While the most common and popular talks at other LMS conferences were variations on the theme of “We migrated to a new LMS and survived the experience,” the most popular sessions at Instructurecon—for years—was how to organize faculty committees that could take more control over the LMS selection process.
I missed that trend, probably because I would have dismissed such talks as sideshows. They didn’t match the pattern I knew. Procurement processes didn’t change. And they certainly didn’t change in response to a few user sessions run by some upstart vendor.
Today, heavy faculty involvement in LMS selection processes is the norm. It’s hard to say how much Instructure created that change, accelerated and shaped change that was ready to happen, or just rode the wave. It was probably a bit of all three.
Regardless, the larger point is that the odds we place on success are heavily shaped by our assumptions. And sometimes those assumptions run so deep that it’s hard to even see them. They’re like the air we breathe. Nor are odds—and common wisdom—meaningless. We thrive as a species partly by employing heuristics to simplify complex and time-sensitive problems. Averages tell us something. They can tell us a failure rate in an easy-to-understand format, for example. They just don’t tell us why the failure rate is what it is. As my father likes to say, if you put your head in the oven and your feet in the freezer, on average you’ll be comfortable. So if you’re going to beat the odds, you have to understand what you’re facing, why the failure rates are so high, and how you can change the equation.
Another challenge with beating the odds is that, whatever you may think about your opportunity, others will be applying the usual heuristic. Others whose support you may need. So even if you’ve figured out a way through the problem, if the people whose help you need don’t believe you, then you’ve got a problem. To maintain sanity and hope, you have to look at that perception problem as just another part of the wall. There are often cracks in people’s judgments and willingness to take a risk. They are not entirely consistent. Look for those inconsistencies. Try to understand them. Often there are deeper rules operating below the surface. Those rules can be the key to finding your way through a wall of “no.” Sometimes it’s direct persuasion. Other times it’s figuring out a way to go around them until you have strong evidence. Instructure didn’t try too hard to directly persuade CIOs. They stirred up a mob of faculty with torches and pitchforks to show up at the door of the CIO’s office. (By the way, torches and pitchforks are often very persuasive tools, especially if you can get somebody else to carry them for you.) Maybe there was no way for Instructure to persuade CIOs to try a new, relatively untested platform with features whose advantages the techies didn’t understand. Fine. Persuade the people who do understand and who can have an influence on your behalf.
Don’t think it will work? Well, that’s probably the biggest wall of all. Self-doubt. Sometimes things are impossible. Sometimes we will fail at solving absolutely critical problems that we throw our whole selves into. But we always fail when we give up. If you see a path and your task is important enough to you, then you have to pursue that path. You have to. Otherwise, what are you here for? What’s your purpose?
Beating the odds, and accomplishing the impossible, often requires us to mercilessly throw out our assumptions and relentlessly hunt for even the tiniest hints of cognitive dissonance that may point the way toward a path that averages and conventional wisdom miss. When facing a dauntingly complex problem, like systemic change in the ways that educational institutions operate, getting out of the tight spot we’re in can seem impossible. The trick is in training ourselves to see the cracks in the wall of inevitability and then believing what we see. Whatever others may tell us.
Fred M Beshears says
Michael,
You should consider writing a book on systems analysis and design. It could focus on the problem of designing (or selecting) a Learning Management System. Also, given your broad background, it could be more philosophically sophisticated than the typical systems analysis and design text book.
Here’s an example of a philosophically sophisticated book on systems analysis and design. Even though it’s rather old, it might interest you. Also, note that it is not focused on LMS design.
Computers in Context: The Philosophy and Practice of Systems Design
by Bo Dahlborn and Lars Mathiassen
copyright 1993
Even though it’s old, its the most philosophically sophisticated book I’ve ever seen on systems analysis and design.
To make a long story short, they discuss systems design work at three levels:
1. Construction – this approach fits with the waterfall method, which involves working a lot on getting “THE RIGHT” specification agreed upon before you start coding. In this case, the importance of having a good group system analysts is seen as being really important. Here, someone interested in getting a good model of reality would be a philosopher of interest. Rene Descartes comes to mind.
2. Evolution – here both analysts and programmers work in small iterative increments, which fits with the agile approach to software development. This also fits with the blind men and the elephant story since it assumes that there are many groups of stakeholders who don’t know the big picture and use different terminology. Further, the stakeholder groups may not know what they want until they start to see the system unfold incrementally. In this case, the importance of having a good group facilitators (i.e. discussion leaders) is seen as being really important. Here, the perspectivalist Fredrick Nietzsche might be a philosopher of interest.
3. Intervention – here groups within the organization may be fundamentally in conflict. There will probably be winners and losers, so groups are going to fight to protect their interests. Here, the organization (or some groups within the organization) are looking for an emancipator to lead serious organizational change. Of course, other groups may be looking for a strong traditionalist to resist organizational change. Here, the revolutionary Karl Marx might be a philosopher of interest.
Allison Wood says
Brilliant, Michael. Thank you for your remarkable insight.
Daniel Gordon says
Thanks for this, Michael. Do you have any thoughts about the extent to which the balance of power shifting from IT to faculty was facilitated by Canvas being cloud-based rather than locally hosted?
Michael Feldstein says
Daniel, they’re related. One of the reasons faculty were increasingly restless is that many IT departments were struggling to keep their LMS running during critical, high use times (like the first few weeks of class or the period of time just before finals). Instructors didn’t care about cloud in and of itself but they did care about the reliability that came with a cloud implementation run by a team with experience to support reliability at scale.
Santee says
I suppose one inevitability Instructure couldn’t avoid was chasing dollars and letting greed win out. Now their product is terrible, and their reviews on Glassdoor are abysmal. It’s a slow sinking ship. Lol, couldn’t have happened to a bunch of finer folks.
John T says
I like this take on the issue, Michael – especially connnecting the concerns of faculty and IT. To faculty, SpeedGrader was a killer app. To IT the cloud hosting (and demand scaling) were critical to success. I think it’s both that were necessary. Solving problems that neither group could well articulate, and a willingness to support small pilots to create advocates, helped them to stand out.
enkerli says
+1 Insightful
Obviously, there’s the “text” about Canvas. Your take is useful for a variety of reasons, including your work with Phil and the mindshare it accrued during that period. The expansion of Instructure’s marketshare is a compelling story (often drowning out some of the other narratives or, certainly, the weak signals).
There’s also a subtext about broader trends in EdTech. Well, ok, maybe not totally subtext. You do call it out. It’s just that… It’s a much deeper story than the Utah gang’s apparent success.
One (weird) way to put it is that it’s about user-driven innovation. Sure, what you describe about Instructure is already a Classic Clay Christensen Case of “disruptive innovation”, which goes well with von Hippel & Schrage. (Didn’t they all teach at the same biz school or am I confusing between parts of the Cambridge intelligentsia?) What makes this trend deeper is that those with pitchforks are also shifting their/our roles.
My personal biases in listening to these stories: I’ve been among teachers with pitchforks about LMS-related decisions and I now work for a nonprofit which manages many instances of a mainstream LMS. Also, I’ve used quite a few of these platforms (from Sakai and Brightspace to WebCT, and from Open edX and Blackboard to Moodle and in-house solutions)… yet I’ve only observed Canvas from the outside. In the abstract, it sounds like an improved user experience over the alternatives… until Moodle 4.x.
Speaking of UX. An interesting trend which is likely to affect platform-happy EdTech is the move from “Instructional Design” to Learning Experience Design. Not too long ago, Downes shared this:
https://www.downes.ca/post/73821
(The original post is in my reading list. I’ll probably push it at the top, right after this one.)
Stephen excerpted this:
“what we’re focusing on is making it a true ‘experience’, not just an instructional event. Ideally, we’d like (it) to be transformational.”
Fully agreed. With most ID experts, I tend to focus on tropes from “design thinking”. That might be a mistake. What Quinn says here is closer to the discussions I’ve had with UX experts, who keep insisting that LXD is no different from designing a website. What they miss, I feel, is that learning itself is the experience. Very different from designing a product or service.
Speaking of which, it does sound like you’re getting close to systems design, which goes well with what Beshears said above. In that case, there’s an embedded theory of complexity through this work. And that goes well with the Open Recognition work by Hilliger et al.