The Annals of the Dark and Dreadful Instructure Wars of 2019

as told by

Matt, the Great FUD Warrior, Breaker of Keyboards, Smacker of Thine Own Head, Asker of Questions That Should Never Be Asked

If you were lucky, you were spared the heartache that came out of nowhere over the announcement that Instructure will possibly be sold to Thoma Bravo, a private equity firm.

Well, it should be stated that the concern from those that always have concerns over these sales announcements was expected. The quick “shush shush – nothing to worry about here” that came in response from people that usually understand the concerns over past Private Equity sales in Ed-Tech (Blackboard being the typical example) was the surprising part.

For the record, the first skirmish actually started when Jon Becker asked what possible outcomes there could be of the sale, Audrey Watters responded with her thoughts on that, and someone made a sexist attack on Audrey’s knowledge of private equity. Also for the record, they initially did not disagree with her points that prices would go up or that Instructure would be broken up and sold off. They stated this was impossible because there is nothing about Instructure that could be broken into parts. Many of us pointed out the sexists problems with the way he expressed his opinion (not his underlying opinion about PE), but he dug his heels in. We also pointed out that there were actually many things within Instructure that could be broken apart, but that was apparently grounds for fighting (even though it is true there are many parts of Instructure that could be broken off and sold, and just because Thoma Bravo has a history of buy-and-build strategy, there is nothing stopping them from still selling off parts they don’t fit their strategy. “Buy-and-build strategy” and “selling off parts” are not mutually exclusive). Within that argument, the idea that the all of the data that Instructure has been bragging about for a few years could either be sold or monetized (more on the important difference there later).

Sometime within the next day, Jesse Stommel made the tweet that really kicked off the main war (I don’t know if he was replying to comments about the value of data in the earlier arguments or it was a coincidence). This is going to be a long post, so I am trying not to embed Tweets here like I usually do. In what was an obvious reference to Instructure bragging that their data was core to their value as a company, Jesse made the comment that we can now know that this value they were bragging about has a price tag of $2 Billion dollars.

Now, can I just say here – I don’t think in any way that Jesse thought that “Instructure data actually cost $2 Billion.” I’m pretty certain he knows that personnel, assets, code, customer payments, etc all are part of that value. Its just that there was a lot of bragging about data being core to the company value, and a huge gap between the market value at the time) and $2 Billion dollars, and that his professional analysis was that data contributed to that in a big way.

Then there was some debate over the value of data in an Ed-Tech company. This was followed by some shooshing and tone policing towards anyone that thought there should be concern over the lack of transparency about data that Instructure has become known for recently, as well as concerns over what could change with new owners. This led to people retreating to their own corners to express their side without having to be interrupted with constant tangential arguments (and there is nothing wrong with this retreating).

Audrey Watters has written her account of the ordeal, which I recommend reading in its entirety first. I am tempted to quote the whole thing here, so really go read it. I’ll wait.

Okay, first I want to clarify something. In my mind, there is a difference between “selling data” and “monetizing data” even though there are obvious overlaps:

  • “Selling Data” is taking a specific set of data (like from a SQL data dump) and selling that to companies that will turn around and sell it to others (which does happen with educational data – more on that later). When someone says “what good is someone knowing that I submitted Quiz 2 back in 2016?”, they are referring to data as a set archive of database rows from a set date. It is kind of looking at data as a crop of apples that were harvested at a specific time. There was concern over this as a possibility, and we will look at that later.
  • “Monetizing Data” is any form of making money directly from creating, manipulating, transporting, etc data. This happens a lot in every day life, and not all instances of it are bad. The core business of most for-profit LMS companies is the monetization of data – nothing in an LMS works without data. Grade Books need data to work. Discussion forums are empty without data. Analytics dashboards show nothing without data. This is kind of looking at data like a the fruits of a field of apple trees that are constantly growing once picked. You could wipe an LMS database of all past data (well, assuming you could find a way to do so without shutting down functionality), but as soon as you turn it on it the code starts generating a massive set of new data. For many, the main concern with the monetization of data is who controls the data and what will they use it for in the LMS? Will I get manipulated by my own data being compared to past learner’s data without either of us knowing about it?

Now, to be fair, most responses were fairly nuanced between the two “sides” of the war. For the record, my “side” in the great Instructure War of 2019 is that “data has the potential to be used in ways that users may not want, which could include monetization, and both Instructure and their potential buyer are not saying enough about what their plans are.” I think that is close to what many others thought as well, but our position was mainly reduced to “all data bad!” While the other side was reduced to “data has no value so stop worrying!”, I do want to examine the idea of whether educational data can have value (to be sold or monetized) outside of either side.

Instructure’s View of Their Own Data

First, I think it is prudent to start with Instructure’s own view on their data. While it would be hard to reference they amount of bragging they have been doing about the value of data at conferences and sales calls, we only have to look at their own words on their own website to see how they view data.

First there is Project Dig. They start off by proclaiming that they have become “passionate about leveraging the growing Canvas data to further improve teaching and learning.” That passion “became our priority, and over the years we’ve provided greater access to more data and designed new, easy-to-understand (and act on) course analytics.” How can a priority of the company not be a huge factor in what they are worth? This is all under the banner of “we’ve been focused on delivering technology that makes teaching and learning easier for everyone.” Obviously, as an LMS, that focus is their main revenue maker as well. And now data is the priority for that focus.

FYI – the target launch date for their tools that will “identify and engage at-risk students, improve online instruction, and measure the impact of teaching with technology.” is…. 2020. Conveniently after the proposed sales date it seems. Again, how could this priority focus of the company that is improving what they offer to customers not be a huge factor in the current sales price?

But they do recognize that there are some problems with digging into data. What word gets mentioned A LOT in the FAQs about potential issues (hint: it involves a word combined with data that starts with “s” and rhymes with “felling”).

Some key highlights from the FAQs:

  • “Will your practices be consistent with your data privacy or security policies?”They say they “are not selling or sharing institutions’ data” – but only because they choose not to. It important to note that the question about selling is there because they feel they can do just that if they want. But they assure us they won’t. Of course, new owners can change that.
  • “Is this really just my data, monetized?”Basically, they say it is not an example of monetizing your data just because… they choose not to, not because they can’t. The implication still remains that the possibility is real and it is there. Then give examples of how they could. Again, new owners are not limited by this choice of the old owners.
  • “What can I say to people at my institution who are asking for an “opt-out” for use of their data?”This is the core problem many have with monetization of data: feel free to do it, just give me the option to opt out at least. They say a lot, but don’t really answer that question (which is very concerning).Important to note that they say “Institutions who have access to data about individuals are responsible to not misuse, sell, or lose the data.” Then they say they “hold themselves to that same standard.” Nothing says this couldn’t change with new owners. But how do you “sell” data that is worthless? They seem to think selling data is as possible as misusing it or losing it. It certainly would be a lot easier to say “no one is out there buying or selling educational value.”
  • (While they do say a lot about openness and transparency, many customers have expressed frustration at some lack in those areas.)

(an important side note in support of Watters point earlier is that, in addition to the main products of the company that each could be broken off and sold or things like assets, employees, etc, these data projects and services represent even more parts that could easily be broken off and sold if a PE firm so chooses to at any point)

Then I give you – Canvas Data. The doc for this service is really a whole page of ideas for how to monetize Canvas data, along with the existing tools to do it. Which is really the goal of the project: “customers can combine their Canvas Data with data from other trusted institutions.” What it doesn’t quite clarify here is that these institution include companies that sometimes charge money to create manipulate, and transport student data inside and outside of Canvas. Many people trust Canvas to vet these companies, but sometimes these arrangements are obscure.

I will give one example from an organization that I think is pretty trustworthy – H5P. H5P does integrate for free with Canvas for free. However, some activities designed in H5P generate grades (which is student data). If you want to transport that data back into the Canvas grade book, you need a paid account with H5P. This is just one example of how a company can monetize Canvas student data, even down to one small data point.

Now, while I see no reason to distrust H5P, I can’t force students to trust an organization they don’t know. What if they didn’t want grade generated on all these websites (because H5P is not the only company to do this)? What if they were not comfortable with a company profiting on moving around their grades? Or what if they were concerned about what the change in LMS ownership meant for all of this?

Anyways, all of this information is up on their website because Instructure believes that their data is very valuable, and that it can be sold. Why would they not point out that data is worthless by itself? Why would they talk about all of this if people weren’t asking these questions, if entities weren’t asking to buy data?

(And Instructure is not the only LMS doing this. Even Blackboard’s Ultra is already ready to do more with data, to monetize it today: “We’re not just handing you data. We’re surfacing data that matters when it matters most— to foster more personalized interactions and drive student success.”  In other words, they are not just doing what other LMS have always done and managed (handled) your data to monetize it, they are adding value by surfacing and doing more. They “drive success,” because success sells. If you have ever been to an LMS sales pitch, you know analytics, personalization, success, all these terms used on the pages I shared are key sales terms to convince organizations to sign the contract.)

The Marketplace for Student Data

One of the contentions of the “LMS data has little to no value” side is that no one is buying or selling student data, as in dumps of past records. It seems that there are existing marketplaces for student data, to the point that someone wrote a journal article on the whole thing: Transparency and the Marketplace for Student Data. “The study uncovered and documents an overall lack of transparency in the student information commercial marketplace and an absence of law to protect student information.” Sounds like a pretty good justification for concern over any student data out there, whether currently under consideration for sale or not. Why it that?

Taking the list of student data brokers Fordham CLIP was able to identify, Fordham CLIP sought to determine what data about students these brokers offer for sale and how they package student data in the commercial marketplace. There are numerous student lists and selects for sale for purposes wholly unrelated to education or military service. Also, in addition to basic student information like name, birth date, and zip code, data brokers advertise questionable lists of students, and debatable selects within student lists, profiling students on the basis of ethnicity, religion, economic factors, and even gawkiness.

That is not all.

Under the Radar Data Brokers

Get ready for this one: there is no evidence that educational data is even staying specifically within a dedicated student data marketplace. This article on under the radar data brokers compiled a list of “121 data brokers operating in the U.S. It’s a rare, rough glimpse into a bustling economy that operates largely in the shadows, and often with few rules.” Most of the entries on the list don’t get into the specific data they collect, so the fact that “education” appears three times on the list for the few that do is concerning:

  • BLACKBAUD INC.
    A “supplier of software and services specifically designed for nonprofit organizations. Its products focus on fundraising, website management, CRM, analytics, financial management, ticketing, and education administration.” (Wikipedia)
  • MCH INC. DBA MCH STRATEGIC DATA
    MCH “provides the highest quality education, healthcare, government, and church data.”
  • RUF STRATEGIC SOLUTIONS
    A marketing firm owned by consumer identity management company Infutor with a focus on travel, tourism, insurance, e-commerce, and education.

These are places already in business, already buying and selling data. If you look at the chart of the attributes (types of data) that Acxiom collects, “education” is one. Its not hard to believe that – if they don’t already have it – they would be very interested to add “I completed quiz 2 on such-and-such date” to that massive collection on each person.

So Why $2 Billion for Instructure?

The only concrete answer we have now is “nobody outside those privy to the details knows.” There are many speculations out there – some of which started the Instructure Wars. One of the main ones I haven’t touched on, that probably summarizes one side of the Instructure Wars, is that the data adds little to nothing to the value of Instructure (despite their own claims to the contrary), but it is “simple math” getting from the current market value to $2 Billion.

I think there is a point to be made in the “simple math” argument (although I would be careful calling it “simple” or claiming “people just don’t understand” if they don’t agree). I would say that even basic Math has to account for the value of data (both at the price it can be sold and the value it can add through monetization). Autumm Caines made the comparison that data is the engine to the LMS car, and you don’t really buy one without the other. In fact, those that are claiming that the data have no value are accusing Instructure of being the shadiest used car sales people in the world: “If you will buy this new fancy car, I will throw in the engine for free!”

However, it seems that different Instructure investors are now disagreeing with the $2 Billion price tag, some thinking it is too low. In fact, they think it “significantly” under values the company. I would assume these investors have access to details about the price of the sale, and if the $2 Billion was simple math, I don’t know if there would be much room to disagree. The cost of the code would be tied to revenue it generates, and therefore would be static and easy to calculate. Various aspects like assets, personnel, and the value of the income from investors are all relatively fixed. Even the future revenue is based on various predictive factors that would be hard to argue.

Seeing that the data is the newest priority of the company, and its value is difficult to calculate, might that be the best candidate for the source of this disagreement? Maybe, but the only for really, really complex data that leads to complex calculations that could easily be off. And LMS data is pretty straight forward… right?

Well, not so much. Kin Lane took a look at what the public APIs of Canvas reveal about the underlying data, and its a dozy. That is another article that I could quote the entire thing, so please take time to read it. I know the page looks long, but that is because he lists 1666 data points (!!!) in just the public APIs alone (while pointing out there are many private ones that probably have many more). He also points out how this structure and the value that it brings easily accounts for the $2 Billion price tag and more, especially when combined with the costs of code and people and customers and so on.

Now, of course, I am willing to bet that there are multiple factors that are causing the investors to fight over price. It could be that they think Canvas’ movement into the corporate training space is about to take off. It is probably a combination of many factors, some I have not even touched on here.

But it is just not far-fetched to think that there is a real possibility that the data is driving the prices, either a commodity to be sold, and/or a service to be monetized.

As I finished this post, Jame Luke published a blog post based on the economics side of the issue. While it does expose that many of us (especially me) are using the wrong terms, the basic idea that the PE will not have education’s best interest at heart and that the data is driving the economics here. He touches on some of the reasons why about data I do here, goes into a lot more depth, and shares several plausible scenarios for the future goals of Thoma Bravo, including one that makes the monetization of data very central to the future sales value of the company. A detailed but necessary read as well. I don’t have time or energy to go back and correct what I got wrong based on this post, so feel free to blast me in the comments if you so desire.

The Ferocity of the Battle

Right now, my direct messages on Twitter are booming with multiple people that are all flabbergasted as to why this is so controversial. We get that there would be disagreement, but the level of ferocity that one side has had in this battle is surprising. Especially since we all thought these were people that agreed that data does bring significant monetary value to a company.

“Data is the new oil!” we were told. But now we are told it was only… black paint all along? Something that will be there in every painting, but doesn’t cost much to be there?

Many people have offered thoughts on why some are some determined to fight this fight. If they get shared publicly I will probably come back and add them here. For my part, I just don’t know. People that want to protect student’s from misuse of data I get, but they have really gone to extra levels of fight over this (beyond what they usually do, that is). The real surprise is the shear irritation from the Learning Analytics community. I know – how did that happen? None of this Instructure kerfuffle says anything bad or good about Learning Analytics, yet they are in the thick of the battle at times.

Still, why is the basic message of “we should be vigilant to make sure that a company that has been a bit opaque with data issues recently gets sold to another company that may or may not be more open, because their data has the possibility to be exploited” so controversial right now? Why must so many people be proven right on the exact price of data? I don’t know.

For me, there could be a news report tomorrow that has Instructure stating “yep, the price was all about the data,” and I would just respond with “okay, thought so.” I get the feeling I am going to be buried in a barrage of snide Tweets if the opposite narrative goes in the news.

Which, let’s be honest, that will be the narrative from Canvas. They have to say there isn’t much value to the data no matter what the truth is. If they let on that it has actual, real value, every single school, teacher, and student will immediately want to sue for their share. Even if there is no lawsuit, the public relations nightmare would cause untold damage as people get mad their data had direct value in the massive sale.

Of course, the reality is that it does not matter what comes out. Canvas already bragged about the value of the data they are monetizing. They already are using it in ways that people don’t want. People have a good reason to be upset about the monetization of their data because it is already happening.

Thus ends the accounting of the never-ending Instructure Wars, as best can be summarized near the end of the dread year 2019.

As the wars drag on and alliances are strained, many began to wonder….

Will this ever end….

Leave a Reply

Your email address will not be published.