Having reluctantly weighed in on Instructure’s proposed acquisition by Thoma Bravo, I would like to turn back toward positive rather than negative possible futures by describing a different potential vector for the company. Despite the criticisms of my previous post, I do not believe that Instructure is in a deep hole, inevitably heading deeper, with no way out in the foreseeable future. If there are two takeaways from that post, they are these: First, it is easy for an EdTech company to have a sudden and dramatic reversal of fortunes and of customer perceptions, especially in the downward direction. Instructure is at an inflection point where the risk of such a turn is particularly high. (I warned about this over a year ago, after Instructurecon 2018.) Second, the seemingly obvious or proximal causes of a company’s success can be misleading. When we misunderstand causality in this way, it heightens the risk. Instructure has not been a well-understood company in general, and my concern is that the current board of directors and executive management may not have a good understanding of the causes of the company’s historic success. So that post wasn’t about what will happen. It was about what could happen.
In this post, I want to explore the company’s actual current competitive strengths as the basis for the kind of growth in value generation that both customers and shareholders would like to see. In the process, I’m going to have to spend a little more time analyzing their missteps. But beating up on Instructure’s management is not my goal. Rather, I want to offer up a case study for how value can be created in EdTech in the 2020s. In the process, I am going to spend some time on the promise and perils of data-driven affordances.
Reminder: D2L, one of Instructure’s major competitors, recently became the first foundational sponsor of e-Literate’s Empirical Educator Project (EEP). Given that big decisions are in the process of being made about Instructure’s future right now, I am obliged to call special attention to the appearance of a conflict of interest.
Instructure’s remaining competitive strengths
As I highlighted in my previous post, Instructure entered the market with competitive strengths that were difficult and time-consuming for their competitors to duplicate. One was that Canvas was built from the ground up to be cloud-native. From their customers’ perspective, that’s a critical ingredient to the product’s reliability. From their competitors’ perspective, it is a very hard technical task to retrofit a cloud architecture onto a mature product that was designed to run on individual customers’ own servers. Another competitive advantage was that Canvas was built from the beginning to delight end-users, whereas older platforms were at least partly built to delight the IT managers who had made purchasing decisions in the early years of LMS sales, and who often had different concerns than the end users. This too is very hard to retrofit onto an older application. Arguably, Blackboard’s biggest mistake with Ultra has been to set customer expectations too high regarding the pace of progress. The rearchitecture and redesign they have undertaken are genuinely, seriously hard.
But given time and motivation, these challenges can be overcome. And sure enough, Instructure’s competitors have been steadily narrowing these gaps over the years. The race is now competitive in these areas where Instructure had the field to itself for a long time. Likewise, their competitors have improved their customer service and focus on improving the end-users’ interactions with the companies.
So what’s left?
First, while Instructure’s brand has taken a couple of hits lately, it’s still strong enough, and customer memory of their positive experiences with the company is long enough, that lost ground can be regained. There are some customers who have had bad experiences, or growing concerns. The CEO has said potentially alarming things in public and failed to correct his mistakes. Nevertheless, I get the sense that the customer base is still rooting for the company to find its footing again.
Second, there’s the culture. Most of the front-line people and important mid-level to senior managers who grew up in Instructure’s culture of excellence are still there. They still have the same knowledge, skills, experience, and convictions. And they have a collective muscle memory of how to work together in ways that were successful for customers in the past. Until we start seeing either talent flight or layoffs that cut deeply into the wrong parts of the company, then the heart of the original company is still beating.
As for the core platform, while it’s no longer dramatically differentiated, it’s still solid. All things considered, nobody enjoys migrating to a new LMS. Instructure is still on a glide path to at least maintain its market share in the US and grow abroad. They would have to actively screw that up for the situation to change. And while that is easier than it may seem, it would take a while.
Finally, while the market has caught on to the value of Canvas being designed as a cloud-based application with a focus on end-users, there is one other novel decision the company made early on that is still underappreciated. Canvas was designed to be a platform, not an application. To grasp the difference, think about the difference between a Blackberry (if you are old enough to remember those) and an iPhone. The Blackberry was a good phone that had a great keyboard. So, in a world where texting and email were becoming at least as important as voice communications, adding text communication features to a phone was a good idea. Yes, there were a few “apps,” for Blackberry. But there weren’t many, and most of them were bad. The iPhone, in contrast, was built for apps. Instead of making the primary interface a great physical keyboard, Apple made it a pane of glass. The interface was anything that could be created for a touch screen. Including interfaces and applications that Apple hadn’t even dreamed of. The iPhone is a platform. It is infrastructure that is designed for other people to write applications that run on it and through it.
Instructure made a series of principled decisions not to develop certain capabilities in Canvas. Instead, they built both the technical infrastructure in Canvas and the relationships with integration partners to have a very strong ecosystem of EdTech companies that have deep relationships with the company and deep integrations with the platform. Rather than building a great keyboard and email app, they built a metaphorical pane of glass and app store. They created a technology platform and ecosystem that attracted other people to extend Canvas with new capabilities.
This too was a pioneering strategy. Before Instructure, the model for LMS partner relationships had been set by Blackboard in the Chasen/Small era. They recognized that the vast majority of digital education products would have to integrate with the LMS. So they decided to monetize that by charging every integration partner a (significant) fee. In those early days, there was some business justification for that strategy. After all, Blackboard had invested in developing integration APIs—Blackboard Building Blocks—and, since they were by far the largest player in the market, integration with their platform offered substantial business opportunities in the form of access to Blackboard’s customer base. This became a significant revenue source for Blackboard that its competitors envied and, to varying degrees, tried to emulate.
But just as Instructure was lucky to come along right when the nature of the LMS customer was shifting, and right when cloud-based development had matured enough to make building a cloud-based LMS feasible, they were also lucky to come to market right when the IMS LTI interoperability standard was starting to take off and EdTech venture funding was really taking off. The LTI technical integration standard made charging a toll harder to justify, while the proliferation of EdTech startups transformed the locus of value for an LMS from a Swiss army knife into the hub of an ecosystem. Or, as my friend Kelvin Thompson from UCF has memorably characterized it, the potato part of Mr. Potato Head. On the one hand, there is no Mr. Potato Head without the potato. On the other, the potato itself is not where the personality (and personalization) come from. Instructure was less focused on charging partners and more interested in enticing them into adding character to Canvas in new and imaginative ways.
Their platform differentiator is subtle in the LMS world, but one that is critical to their current situation.
In fact, it is essential to the first of two paths Instructure has for developing a product portfolio that I’m going to describe in the remainder of this post.
Acquisitions
There is now a whole galaxy of small EdTech companies that have developed new products or services that touch the LMS in one way or another. Instructure arguably is in the best position of any of the players in the market to identify good companies with products that already integrate well with Canvas and snatch them up. It’s not an exceptionally strong or durable advantage, but it’s one that they could be utilizing aggressively right now.
Former Instructure CEO Josh Coates had a philosophy about acquisitions. Specifically, he was against them most of the time. I heard him give a talk about it once. His strategy was thoughtful and well-reasoned. And it would have been justified had the company succeeded in organically developing a portfolio of compelling offerings. But that didn’t happen. So, at least for the short term, the company is going to need to be more acquisitive in order to become financially healthy. But for that strategy to work, Instructure will need to rely on the knowledge of its employees who have worked on the platform aspects of Instructure’s business to identify truly good companies. What I mean by “truly good companies” is ones that have earned customer loyalty by solving some important problem well. Instructure needs to find compelling products, which is not easy in EdTech. The company leaders will also need to clearly and consistently articulate the reasons why they are proud to have acquired those products and how the acquisitions will better serve their Canvas customers.
Instructure has made one high-profile acquisition under CEO Dan Goldsmith: Portfolium. Here’s the relevant bit from Instructure’s press release announcing the acquisition:
Portfolium was created to help every person realize their full potential by connecting their learning with opportunity. The company helps institutions inspire, assess, and showcase student achievements via its powerful ePortfolio network, student-centered assessment, job matching capabilities, and academic and co-curricular pathways.
“Working with Portfolium advances our mission since it enables us to help people move from the classroom to the workplace,” said Dan Goldsmith, CEO of Instructure. “Portfolium has been a great partner of ours. With their team, and by adding their student success capabilities built on the leading learner network, we will, together, provide more value to both current and new customers.”
Instructure Enters Into Agreement to Acquire Student Success Network Portfolium
Honestly, I don’t know what that means. Portfolium is an ePortfolio. That blurb mostly describes what an ePortfolio does, but adding some pleasant adjectives. Shorter version:
Instructure is acquiring Portfolium, which is an ePortfolio company. Portfolium does things that ePortfolios do. We like it and think you will too.
Honest press release
The only bit in there that isn’t completely generic is the sentence fragment about job matching. But they don’t do anything with it.
In fairness, most people who haven’t written a press release don’t appreciate how hard it is to write a meaningful one. I’m a pretty good writer, but I will readily admit to having struggled with that particular genre at times. That’s exactly why smart companies don’t rely on the press release to carry all the weight. They go out and repeat and elaborate on the message. Relentlessly.
I haven’t heard any such message about Portfolium.
Again, in fairness, I have not been following the LMS market as closely as I used to. But the thing is, if Instructure were really doing this right, it shouldn’t have been possible for me to miss this. I should have read it in the articles in outlets like EdSurge, Inside Higher Ed, and The Chronicle. In preparing to write this post, I did a search to see if I missed anything in the coverage by these or any other outlets.
Nope.
No quote from the CEO beyond what was in the press release. In fact, no quote that I could find from anyone that wasn’t already in the press release. There was an interview with the Portfolium founder in the San Diego Tribune. But it was clearly a local business story, in a local outlet—Portfolium was San Diego-based—and not anything aimed at Instructure’s customers. There’s nothing. Nada. Zip.
When I ask Instructure customers about the deal, they tell me they haven’t heard anything either. And when I have run into Instructure employees at conferences—specifically, ones who are in a position to know about Portfolium—none of them have brought it up with me. In the old days, they would have. These are the same humans. There has been no invasion of the body snatchers. So it appears that the corporate communication strategy has changed. Where it used to be the case that you could get any random Instructure employee driving a golf cart at Instructurecon ((Golf carts driven by employees at Instructurecon is an actual thing. Or at least, it used to be.)) to talk about just about anything, it is now difficult to get even senior Instructure managers to talk about…well…just about anything.
So there is at least one and possibly two serious but very fixable problems here. First, there is definitely a communication problem. Specifically, it would be good to have meaningful communication directly from the people within the company who understand why acquisitions are being made. Instructure employees have been some of the best brand ambassadors in the sector. Circumstantial evidence strongly suggests that they are no longer empowered to speak for the company. By muzzling them, Instructure is voluntarily throwing one of its remaining competitive strengths in the garbage can. And the only senior manager who has been allowed to speak on the record about the Goldsmith era’s first important, high-profile acquisition—to anybody, apparently—is the CEO. The one who has the most to prove about actually knowing the sector and caring about it. And his only statement that I can find amounts to a spoonful of bland pap in a press release.
I don’t know much about sports, but I know what an unforced error is.
While it would be easy to lay much of this at the feet of the marketing department, there has been a pattern of communication in the Goldsmith era that started immediately after he became CEO and has been consistent despite some churn in marketing department senior personnel. In fact, that churn may be an indicator in itself. The leader sets the tone.
The second possible problem is harder to assess precisely because of the effects of the first one. I can’t tell if Portfolium is a good acquisition or not. A lot depends on what problems Instructure’s managers think it solves for customers and how they intend to make it more useful. Which they’re not really talking about. I can’t tell if this acquisition was a paint-by-numbers decision or if there is a real effort to increase value for the customers in a mindful, meaningful way. Instructure has people who know which partners are good potential acquisitions and why. They know their partners and customers really well. I can name a handful of them off the top of my head. In general, they are not being heard externally, which is unfortunate. If they are also not being heard internally—and again, I can’t tell one way or another—that would be a lot worse.
I am confident that Instructure has the right ingredients and the right chefs to cook up an effective acquisition strategy. But the proof of the pudding is in the eating. If Instructure doesn’t both prioritize the acquisition of companies that will serve their customers well and communicate the reasoning and intentions behind the acquisitions, that could easily mean the difference between customers who are hungry for more and ones who are left with a bad taste in their mouths.
Insights
The other area for growth is in providing educators and students with better insights that support student success. It is impossible to overstate both how important and how fraught a topic this is.
On the one hand, colleges and universities fail huge swathes of students all too often. There are students who want to go to college, get admitted to college, but haven’t had anybody to teach them how to succeed in college. Students who never make it to the first day of class. Or who get partway to a degree and drop out, due to lack of skills and support or tough personal circumstances. Students who need to come back to school and reskill while they are holding down full-time jobs and raising families.
Many colleges and universities are not very good at serving some or all of these groups and routinely fail them. This used to be “only” a moral failing. Now it is an existential one, because the most obvious path to long-term sustainability for an increasing number of colleges and universities is to serve more students in their area more effectively for 20 or 40 years rather than for two or four. Institutions of higher education need to learn to perform better. And they need to learn it urgently. To do that, they need new insights. To get new insights, they need better information—that is, better data—and better ways of analyzing it.
On the other hand, we live in an era when people have good reason to be concerned about the misuse of data in a seemingly ever-increasing variety of deeply troubling ways. Educators are responsible for their students. Many of them take this responsibility very seriously indeed. And particularly when it comes to the experimental use of data—even for the best of purposes—that responsibility is hard-wired into institutional processes in very particular ways.
In academia, educational research falls into a larger bucket of “human subjects research.” That category also includes research on topics like how to perform open-heart surgery, how to deal with intractable clinical depression, and how people can be manipulated or fooled by social media. Think about the possible unintended consequences of poorly designed experiments in any of those three areas. We have canonical examples from bygone eras. The Milgram experiment. The Stanford Prison Experiment. Today, any academic research conducted in the United States that involves human subjects must submit its experimental design and protocols for experimental subjects’ informed consent to a rigorous peer review and approval process before the experiment can be undertaken. Educational research must be submitted to the same approval process by the same oversight board that would approve life-and-death surgery or psychological experiments.
But unlike in medicine, an EdTech company that wants to conduct research using student data for product development purposes is required to do…nothing at all. They don’t even have to inform the students that they are being experimented on.
In fact, EdTech companies conduct experiments on a weekly basis that would require a lengthy approval process in some universities. ((Different universities interpret the rules around such approvals differently, in part because a university that has a world-class medical school will likely understand their responsibilities differently than one with no medical school at all but a significant psychology research program, for example.)) Suppose, for example, that product developers want to test which design of a button is more likely to raise user awareness of a feature. They conduct what is known in the industry as an “A/B” test. They show some users one design, other users the other design, and track which design gets the most clicks. This is an absolutely routine software development practice. It is a foundational strategy that developers use to learn how to make their software more useful and usable.
But let’s also suppose that the feature the button activates has a significant impact on improving student outcomes. Using the feature helps to improve learning. The developers, with the best of intentions, are trying to figure out which version of the button will get more students to use the feature that will help them. But in this experiment, the button that turns out to be worse may negatively impact the learning outcomes of students using that version of the software.
This is exactly the kind of contingency that might trigger a requirement for an experimental design review process inside a university.
Imagine that you’re an academic who feels responsible for your students and who lives in that kind of a culture surrounding anything that remotely looks like it could be human experimentation. Imagine that you read the following statement by the CEO of a company whose EdTech product you, personally, require your students to use extensively every single day:
We’ve been working on the scaffolding for [DIG] for well over a year now. I mentioned in our remarks that we already have product validation towards out there in the market. We have instructors and students consuming output from some of the initial experiments with DIG. And we anticipate later this year obviously to make more announcements around specific products and offerings and how we bring them into the market. DIG ultimately is a platform first and foremost based upon machine learning and artificial intelligence. I believe that any multi tenant SaaS company born in the cloud has the opportunity once they hit a certain market share. And in fact, it may even be incumbent upon those organizations to partner with the industry and evolve that industry with new insights and predictive modeling using AI and ML. That’s what DIG is at its heart.
We already have analytical capabilities in our Canvas platform. I want to be really clear and delineate the difference between an analytics and reporting capability, and a machine learning and AI platform. [snip]
We have the most comprehensive database on the educational experience in the globe. So given that information that we have, no one else has those data assets at their fingertips to be able to develop those algorithms and predictive models.
Instructure CEO Dan Goldsmith
Heads would explode. Heads did explode. Heads are still exploding.
Even so, as I said in my previous post, that was a very fixable problem. Dan was still new. He easily could have played the “new guy” card. A mea culpa, a couple of comments to reporters, and a brief but earnest listening tour likely would have blunted the worst of it. The company could have reset and been in a position to have productive conversations with customers about this thorny set of challenges that they need to face together. Instead, Goldsmith said nothing, and the problem festered.
In July, Instructure VP of Higher Education Jared Stein wrote a blog post on DIG trying to settle things down and dispell some of the concerns. I trust Jared and, more importantly, Instructure customers trust Jared. He did a decent job in that post, as far as it went. But one blog post by a senior employee, three months later, is not going to undo the damage of such a bad faux pas by the CEO. A CEO trumps a VP. Therefore, a CEO’s misstatement can only be credibly corrected by the CEO. Furthermore, Pandora’s Box had been opened. Because Instructure didn’t jump on the lid the moment they saw it crack open, all the deeply difficult questions about uses of student data in EdTech have come flooding out. The company didn’t lose everybody’s trust, but they lost the trust of enough vocal customers that now they have a persistent problem.
This too is fixable, but it must be fixed. Some damage has been done to Instructure’s reputation. Enough that a more concerted and sustained effort must now be made which includes actions and not just gestures. But the imbroglio has not yet permanently damaged the company. People have long memories of their experiences with Instructure and personal relationships with employees who still work there.
Update: Jeff Young just published a piece out about the data concerns in EdSurge. It includes quotes from Instructure’s chief spokesperson Cory Edwards, who I don’t know well but have found so far to be a good actor, and Melissa Loble, a long-time Instructure vet and current SVP for Customer Success, who is exactly the kind of person we should be hearing from directly more often. There’s even a quote from a Thoma Bravo representative. All were responding directly to the data use uproar. So this is progress. But still nothing from CEO Dan Goldsmith. Why EdSurge was able to get Instructure’s prospective PE owner on record but not its CEO, about a problem that was set off by a comment made by that CEO…it’s just mystifying.
Education needs insight-providing, data-driven tools to help educators better serve a wider range of students who could succeed if only we were offering them the right kind of help. Furthermore, educational institutions need productive partnerships with the private sector to get these solutions out to as many students as possible as quickly as is possible and responsible. But this partnership can only take place in a high-trust environment. Instructure has had the necessary level of trust from their clients to do this kind of work. They have damaged that trust in this particular area. But not beyond repair. There is still a sound foundation, and any cracks still can be repointed.
This brings me to Instructure’s most compelling product. It is not Canvas. Canvas has only been Instructure’s second-most compelling product.
Brand, brand, brand
Most people interpret “brand” as shorthand for a banal series of tactics employed by marketing departments. Academics, in particular, are inclined to load the word with distasteful connotations of shallow window dressing at best and obnoxious disingenuousness at worst.
Nothing could be further from the truth. “Brand” is another word for reputation. It is who people think you are. It’s how much they trust that you are who you say you are. How much they trust you, period. Brand isn’t a series of tactics. It is the outward manifestation of the character of an organization or individual, as understood by people who have come to know them based on their actions over an extended period of time. Far from being a small set of eye-rolling marketing gimmicks executed by a small set of individuals, an organization’s brand is the gestalt impression that people get from every single interaction they have with every single member of that organization and every single product, website, or other touchpoint. That gestalt is “monetizable” to the extent that people trust the organization to understand their needs and have their interests at heart. They will give you their money if and when they believe they can trust you with it.
Instructure’s brand has, until now, been its primary and best product. It is still one of the best in the sector, even if it is getting a little ragged around the edges. Because the brand is still good, the company can still build the relationships it needs to make good acquisitions, evangelize those acquisitions to its customers, and work with its customers on even the most sensitive (and important) product research and development efforts.
That is what I hope for Instructure and why I have expended so much energy writing these last two blog posts. I want to live in a world where EdTech vendors are successful because their customers and partners believe in them. Instructure has been that kind of company. And it still could be.
But back to the proof of the pudding. In this 21st-Century economy, as the saying now goes, if you’re not at the table, you’re on the menu. Instructure can recover, succeed, and thrive to the degree that the company’s leadership can reignite their customers’ faith that they have a seat at the table. Every decision they make, including financial transactions, should be judged by how it helps or hinders them from doing so.
