Instructure's Acquisition is Not What I'm Worried About

It's interesting how quickly and how slowly the world of edtech moves. I left Instructure at the beginning of 2015 to work on speech-generating software for people with disabilities. Some people tell me I shouldn't care anymore, or that I don't deserve to have an opinion anymore. That's cool, I get that, I just can't seem to stop worrying about what happens to this Little Company that Could. So here we are. There are probably, like, five of you reading this. I hope you had a pleasant week :-).

Growing Up Edu

I was injected with a healthy dose of edu idealism when Devlin Daley and I, fresh out of college, met with CIOs and instructional technologists from around 20 universities and community colleges, and I like to think it's done a lot to shape my mentality around technology. If nothing else, it's been incorporated into this weird mashup of I-don't-know-much and here's-how-awesome-it-should-be that realtors, contractors and consultants all stopped bothering trying to wrap their heads around eventually. I'm kind of a mess :-).

I believe that technologists should find and implement interesting and valuable solutions to existing problems, and that there should be compensation available for their work. It's pretty simple and naive, but it's how I think things should work. Not all good ideas get compensated, and not all problems are well-defined, but there you have it.

I believe that that's what we shot for when we started Instructure. There were clear problems with the LMS, and we had the right timing, background and lack of baggage to introduce a compelling solution. People were willing to pay for that solution because they believed it would improve their effectiveness or productivity.

Spinning Plates Don't Stay Up on Their Own

Neither Devlin or I wanted to actually run the company, so we brought in others at different levels who did, which was undoubtedly a good idea. Some of our new business-heads had different motivations, which is fine, but it meant that as the company grew we had people in leadership roles who never really understood what it was that got people excited about Canvas. They knew people thought it was a game-changer, but they didn't understand or didn't care about the intricacies of the edtech market. Software-company-as-a-generic-machine is a mentality that starts with investors and trickles its way down into all the seams. Sales is sales, support is support, you plug in the right pieces and out comes the revenue.

For us, it seemed like the only place where this broke down was around product strategy. We couldn't seem to empower the product team enough to drive a strong strategy and ended up chasing revenue or headlines. We won some great deals (and headlines) early on, so it's not like it was a complete loss, and it'd be easy to blame our shifting tides of product focus things on the edtech market and its idealism or lack of savvy or privacy-mindedness or corporate sensitivity, but I was over product in those days so, you know, my bad.

I believe that this product problem was part of an issue that's still showing up today with Instructure, and it affects the company strategy in general. It has to do with bolt-on acquisitions and not-invented-here syndrome. It has to do with chasing the dollar instead of earning the dollar. It has to do with forgetting to strategically ignore ROI. It has to do with how nobody gets fired for following the Silicon Valley playbook.

Finding the Right Playbook

The hard truth is that when you bring on an institutional investor, they don't really care what you're already doing, they're only looking at upside because they're there to optimize their return going forward. Some investors do this hyper-short-term, and mess up their companies down the line. Some apply healthy pressure and are more patient as long as they're still confident they can beat the return they'd get from other investment strategies. Either way, they see each company as a revenue opportunity, not a feel-good story to tell their grandchildren.

This isn't inherently bad. They get their money after the company makes money, and the company makes money when people are excited enough about it to give it new money. So investors look for patterns of behavior that worked well in other situations and apply them once again. Accelerate revenue by entering a new market or increasing walletshare for existing customers. Improve margins by paring down R&D and shifting resources away from established product lines. Bring in an executive layer that already has experience handling issues at the upcoming stage of the company's growth. It's a sort of standard operating procedure you can start from to make sure you're not accidentally veering off in a weird direction. Sometimes you may want to veer, but if not then it's the "safe" route because it's so well-trodden. Once companies get big enough you don't need investors suggesting this stuff anymore because the corporate executive team has already been trained in how it all works.

Except it doesn't all just work. Personally I'm not a huge fan of any of the above strategies, but that's ok, they're really just business decisions. They can succeed or not based on what happens underneath and around them. That's where I think Instructure has been struggling, though. It's not that the broad strokes are all wrong-headed, but they won't succeed just because they're what you're supposed to do. At a high level things at Instructure don't seem particularly deliberate or concerted so much as just, what you're supposed to be doing, and I think people are starting to notice.

Unfortunately, I think this has been going on a lot longer than people realize. Before I left there was already hand-wringing around "why can't we just build the next SpeedGrader?" ironically mixed with "let's shift resources over to new projects" and "we have too many bugs so we're going to pause new development for a while and address stability". It's hard not to fall into a routine of medium-term hype-chasing and putting out fires, which I think comes from that business-for-business'-sake mentality. I'm way too much of a product person so I never actually think a product is "done" (it's a personal problem :-)), but it's especially hard for me to see untapped potential in a product -- even one that doesn't have justifiable revenue upside.

I believe that Instructure worked so well in the early days because we accidentally instilled in everyone an urgency and empowerment to think of ways to make things better, even if it was less efficient. Some people come by it naturally, some focus on their own little domain, but the outliers were the ones there to just "do my job". We drove home how much we needed top to bottom ears to the ground, and we absorbed and obsessed over all that feedback so we could throw all sorts of unique ideas against the wall and build the ones that stuck. I've watched that kind of whittle away over time as people got roasted for sticking their necks out, and deadlines and parity and we-can-build-and-sell-one-of-those started trumping that's-interesting-how-could-we-improve-that.

Keep Canvas Weird

Look, none of this is particularly ground-breaking. "Founder says company needs to stop being all corporatey and focus on what made them great in the good old days. Nothing to see here, move along." You could say I'm just drawing from the playbook of what founders did in the past, we all play our part and the pattern repeats itself. I won't get my feelings hurt, I'll just chalk this post up to catharsis or something. I just can't not say something so long as I think my saying something has a chance to turn at least a couple heads back toward the idea of solving interesting problems.

It's not like nothing good has happened. The Canvas API has solidified, mastery-based courses can grow into something cool, and I'm interested to see how the new Analytics views develop. But Canvas can do a lot more. Did you know the original version of Canvas has a feed reader where students could register their personal blogs and then submit homework to their classes just by posting with the right hashtag? Or that we originally had an embed code you could drop in your HTML to allow access to grades, homework submissions and discussion boards from an arbitrary web site? Or that we mostly-completed a heuristic-based tool that would highlight outlier students based on activity and participation in that specific class? Or that we half-wrote a tool for auto-assigning students to groups based on same- or different-mindedness based on a teacher-created survey?

I can tell you exactly why none of those projects stuck, and maybe they all sound like terrible ideas to you. But as Instructure has become more streamlined and optimized, it's harder for me to envision where in the company those types of ideas can even germinate. When business pressures are applied, everything naturally moves toward the well-trodden path, where it will see less resistance. This is a big part of why most big companies just copy each other's features, or end up buying their way to innovation, there's literally nowhere left for the interesting solutions to get enough oxygen.

Typically innovation still happens at large companies only when someone high up has gained enough social capital to just do what they want, or when the company makes a deliberate decision to shield some groups from the business scrutiny of optimization. I'm not confident that Instructure has been set up to accomplish that kind of innovation for quite a while. Even back when I was wrapping up my time, the teams that had enough freedom to do anything more than incremental development and refactoring were pretty minimal. I would be surprised if things have improved on the Canvas side.

What Now?

I don't believe Instructure has lost its soul. I think people still get jazzed about working there. I think customers are still generally happy with Canvas. I think Canvas is a valuable contribution to education. But I don't think customers are looking to Instructure anymore for the Next Cool Thing. And I do worry the company's open floor plan is shifting from "I can walk info wherever it needs to go" to "I have a great view outside but it's kind of noisy in here". I worry that Canvas is about as good as it's going to get, which bums me out since we never managed to tweak things enough to convert the edupunks (personal problem again). There are a lot of forces tugging on the company anymore, and it's hard for me to see how adding another return-focused investor is going to do anything to improve the existing product (indeed, Michael Feldstein argues that Instructure's problem is that it needs more breadth, not more Canvas Cowbell). But those are primarily business decisions, they can end up good or bad depending on how you handle them. For me (and, I would argue, for edtech customers), I care much less about percent ownerships or the size of Instructure's product portfolio than I do about the current (not historical) impact of the products in that portfolio.

Again, I'm not here proclaiming all is lost, woe is my old company. I think there's a course correction in order, but nothing terribly drastic. There are still great people at Instructure trying to make good things happen for educators and learners, and there are great customers who want to share their insights to keep improving on a strong foundation. But nobody asked for SpeedGrader, it showed up because we had the latitude to throw it against the wall. If there were a way to shift back to that old style, where business-heads make sure things stay smooth and maintained, but while also letting enough people get all weird solving compelling problems for educators and learners, I think you'd see a "next SpeedGrader" show up of its own accord. And maybe the portfolio would grow too, who knows? But it wouldn't be expansion for revenue's sake, it would be building interesting things and figuring out how the industry is willing to compensate for them. My druthers would be allocating resources based on social impact as opposed to revenue, but I understand if that's a (ahem) bridge too far.

It's true that growth is slowing for Canvas, so from a business perspective that means there's less value to be had investing there. Unless, of course, people become dissatisfied or disillusioned and move to a different product -- but frankly there's not anyone doing drastically better in the space so the current risk is small. Canvas wouldn't be the first LMS to fester all fat and happy, though, and you can't ramp down and up R&D and expect it to remain effective. There are interesting problems I wish Canvas already had solutions implemented for. Some of them have been sitting in the Feature Request queue for years. There are going to be new problems that Canvas could grow to solve. I wish I was more confident people inside the company will be empowered to create the solutions.

I hope Instructure can find a way to keep solving interesting problems in education, I still don't see anybody else in a better position to provide a major influence on educational technology. I'm confident that the industry will continue to compensate them for their work as long as they do.