D2L announced the acquisition of H5P Group on July 9, 2024, just three days ago. Many people have asked me for my opinion. To save myself from repeating the same points over and over again, I’m going to cover the topic extensively in this post, which I can then share as a reference.
Please note that this may be a long post, and I will not provide a TL;DR. Also be aware that this post reflects my opinion, which is inherently biased, and contains speculation that may not be entirely accurate.
H5P Group and H5P
In case you did not know (all these facts): H5P is a versatile software framework that allows you to create interactive content on the web and to integrate it into a number of different platforms. It’s not a Learning Management System (LMS) however, but is only about creating and sharing content. However, H5P can be used in combination with LMSs and heavily is by universities and schools in different ways.
H5P is (or was) developed by the Norwegian company H5P Group (former name: Joubel). H5P Group is (or was) a stock corporation, yet shares were not publicly traded. The software framework H5P is openly licensed and free to use. The initial funding included public money via NDLA, but private funding too. Development is mostly funded by earnings from H5P.com, a commercial closed-source platform which is run by H5P Group and which offers a couple of premium features that are not available in the free version of H5P. Recently, most development of so called content types for H5P was done by me as a contractor, funded by different clients (mostly companies and universities) that agreed to openly license the work in exchange for receiving a discount on my hourly rate.
D2L and Brightspace
Again, in case you did not know: D2L is a publicly traded stock corporation based in Canada. Their primary product is the Learning Management System Brightspace, which is a closed-source software-as-a-service platform. Brightspace is predominantly used in North America, according to my understanding. It’s worth noting that, anecdotally (I could be wrong here), D2L’s main business currently focuses on corporate education and training rather than universities and schools.
Brightspace can be enhanced with various add-ons, such as Creator+, which serves a similar purpose to H5P by enabling the creation of interactive content for teaching and learning. D2L generates revenue from licensing the use of Brightspace and uses these earnings to continue its development. Additionally, D2L offers services related to course development, instructional design, and other educational support.
What happened and what were the reactions?
As mentioned at the beginning, D2L has now acquired the H5P Group, which sent shock waves through the H5P community. This development came as a surprise, even to me. In addition, the communication around the acquisition was poor. While secrecy was required by law to prevent insider trading, as D2L is a publicly traded company, the news was spread through media and from there via social media channels rather than through official press releases with more detailed information. H5P Group’s own status report was released more than 24 hours after the news broke.
So what were the reactions? I contacted several people directly and searched the web to gauge the community’s sentiments. It’s important to note that these reactions occurred before the official press releases were issued, so they reflect the initial surprise and concern.
Nearly everyone feared that the open part of H5P would be off worse soon
Some felt disappointed and even “betrayed”, stating that they are believers in “true open source” without further specifying what that term meant. I can only assume that “true open source” means contributing to the code just for fun or purely to make the world a better place without and to improve as a human being – anything corporate-y stinks. I know that notion. It’s a fine idea. Very Star Trek, I like that a lot.
On the other hand, it’s not the reality today. There are hardly any large open source projects that work that way. We still have financial concerns. Either companies have dedicated people who get paid to spend a significant amount of their time directly improving the code, or companies sponsor “independent” developers to work on their code. Don’t forget about writing documentation, support, … There are many small projects run by enthusiasts in their spare time, but H5P is beyond that. And experience shows that trying to maintain a large project with only volunteers often ends in burnout. I would be amazed if there was a vibrant H5P community that could organize itself and do all this on a volunteer basis, but I haven’t seen any attempt to go in that direction yet, and I doubt it would be feasible.
And on the other hand: At no time has the development of H5P been driven by such a “true open source” initiative. Yes, the people involved believe in the idea of open source, but they also have bills to pay. I’m one of them – I wouldn’t have to give a discount when I’m supposed to create openly licensed code, I wouldn’t have to try to convince clients to release the work to the public, I wouldn’t have to maintain all the contract work I’ve done for free. I may be digressing. Last week, the model was that commercial entities drove and funded the development of H5P. This week the model is that commercial entities are driving and funding the development of H5P. What has changed? Sure, one could argue that other business models could have been chosen or sought – but what does that have to do with the acquisition?
Some argued based on past experience
One person compared this to when Moodle launched Workplace, which didn’t mean that nothing was being done for Moodle, but he said that he felt a certain difference between the platforms.
Another person referred to the openly licensed editor TinyMCE, which was bought by a company with the result that the next version will be proprietary. You have to stick with the current version, which will eventually become obsolete, pay for the new version, or develop something new.
Another drew the same parallel for H5P and said: “Only h5p.com makes money directly, so I think that’s where all the energy will go. In the worst case, the open source part will not be maintained at all. If there really is a H5P version 2, I suspect it will be completely closed source.
In general one is worried about the open side
The gist of all the answers I found was that people were worried about the open side of H5P. You probably have nothing to worry about if you only use H5P.com anyway, duh.
Why is this relevant? You could just say: “It’s not like open source is free, it’s just that others pay for it (with their money or time)”. That’s true, for some or many it is about that. I just touched on that topic before. But it’s also about “free as in free speech, not free as in free beer”. I don’t want to go deep into this topic, but just give some real examples why a closed H5P (or H5P.com) is not an option for some of my clients (companies and universities) – it’s not for the money:
- They cannot run H5P content types that are not hosted on H5P.com, so you cannot use some of the 20 or so content types that are already floating around but stuck in H5P Group’s review queue.
- They cannot fix bugs they find in existing content types, and they cannot add features they need directly. They could contribute the code, and are often willing to do so at their expense, but H5P Group is a huge bottleneck. Their processes take a lot of time, but without a reliable schedule for when contributions might be released, my customers are better off hosting and modifying H5P components themselves.
- They cannot add additional features they need, such as custom code to score responses on the platform, etc.
- For universities in particular, it’s also about a feeling of being in control of data.
Still not convinced? Why should you care what others want? Well, why do you think we have cool things such as Lumi (not the D2L software which is also called Lumi, that’s just another source of confusion that just popped up)? Because the source code is open and people can create new tools to enrich the world of H5P.
These reasons (and others) have nothing to do with a price tag. It’s about empowerment. And people are afraid that this power will be taken away from them. No, nobody would rip the current version of H5P from their servers, but maybe people need a perspective, too.
What also happened?
As mentioned earlier, both D2L and H5P Group issued press releases that hold important information that’s not covered by news items. Let’s go through these first, and I’ll directly quote the FAQ items here.
What the press releases say
“What is happening? D2L’s acquisition of H5P Group will enhance the H5P product and services by integrating additional resources and expertise. Our goal is to maintain H5P’s innovative and user-friendly aspects while leveraging D2L’s robust infrastructure to improve performance, reliability, and support. We can expect continued innovation and improved service delivery.”
That sounds promising. I don’t know about D2L’s infrastructure. But if additional resources and expertise can be used to speed up the development process (there are some interesting things in the pipeline, ranging from improvements to the editor to upgrading accessibility support from WCAG version 2.1 to version 2.2 and more), who wouldn’t want that? If it also means reviewing all the code contributions that I and others have made over the years that are still up in the air – hell, I will not say no to that.
If D2L really does help to introduce automated end-to-end testing to improve the software, as H5P Group’s status report states, then let me just say: I was first 😉
“Will the acquisition benefit the H5P community? D2L is committed to investing in and contributing to the H5P community as it helps more educators deliver engaging and interactive learning experiences, both to D2L users and beyond.”
That’s not regarding development. That’s a commitment to the community. Could be more free webinars about H5P topics. Could eventually be a next H5P conference that people have been asking about but that would need resources for organizing. Could be … What would you like to see?
“Will H5P continue to be open source and continue to be developed? Yes, all parts of H5P that are open source today will continue to be open source and will be maintained and improved at a greater pace than before.”
That’s what people wanted to hear, wasn’t it? Some may dissect that sentence, however. What is that “today” doing in there? What about parts created tomorrow?
On the one hand, this sentence does not say that all parts created from now on will be closed source. On the other hand, how is the situation different to the situation one week ago? Since the launch of H5P.com, H5P Group has been developing new parts, some of which have remained closed, while others have been made public. A recent example is the development of the “Row” library, which allows users to add rows to Column content (or other content types). Its source code is openly available, but a more convenient editor widget remains closed source for paying customers of H5P.com.
“Will H5P continue to maintain integrations for Moodle, Canvas, WordPress etc.? Yes, all existing integrations with other LMS platforms will continue to be maintained. This guarantees that current users will not experience any disruptions and can continue leveraging H5P’s capabilities across various learning management systems.”
That’s also what people wanted to hear, right? We’ll discuss this a little more in just a second.
“Will it still be possible for community members to contribute code? Yes, community contributions will continue to be welcomed and encouraged. This open collaboration is essential for the ongoing innovation and improvement of H5P, fostering a vibrant and active community.”
The process will need to be streamlined for sure. You can’t simply ignore questions of code contributors on github for instance like today. But in general, confirming the willingness to accept contributions is what you’d want for an open source project.
What the CEOs said
On July 10, 2024, one day after the acquisition, Svein-Tore Griff-With (CEO of H5P Group) and John Baker (CEO of D2L) participated in a webinar for so-called H5P Ambassadors. They shed some light on the situation and answered questions. The webinar was recorded, but I am not in a position to share the recording with anyone who is interested. But of course I can quote some things.
Svein-Tore pointed out that H5P Group doesn’t have the capacity to achieve what they want to achieve on their own. For example, they are aware that they have handled pull requests poorly in the past, and that reviewing content-type contributions takes too much time. He didn’t say so, but I suspect he felt that progress was too slow, even given the recent growth of H5P Group and a good balance sheet – there was no need to sell.
Svein-Tore also made it clear that the people will remain the same and the vision will remain the same. The H5P Group will remain a dedicated product group that will essentially continue to work as before, but will be given more resources. To me this sounds a bit like Mojang within Microsoft when Minecraft was sold – Microsoft didn’t kill it, I still love the game.
John then addressed this and reassured me that it is indeed the intention to keep things open as they are. He also emphasized that accessibility plays a big role in interactions within Brightspace (or Creator+, more on that below).
When asked how D2L plans to benefit from keeping H5P open – after all, they paid $25 million for it – John said they’ll continue the good work that’s been done, but they’ll invest even more. H5P will grow at a faster pace and John likes the strategy of being present on other platforms like WordPress or Moodle.
Speaking directly to what I do for a living, John mentioned that there is a large long tail of interactions or content types that need to be built, and that no company in the world could do it alone, so having a solid open infrastructure makes sense.
As for H5P integrations for other platforms, it was said that there was no intention to take them away, and that John would reach out to the maintainers of those platforms. Of course, Brightspace would get the best integration with H5P.
All is well that sounds well?
That didn’t sound too bad, did it? You will probably quickly find reasons why all of the above may not be true, or why things could go down the drain anyway. I have listed some of these arguments above, along with the immediate reactions that followed the announcement. I prefer to be a naive optimist and ask myself what (other) reasons D2L might have for actually keeping H5P open. Here’s what I came up with, though again, this is all in my head – I don’t know the plans of D2L and H5P Group.
They said so
First of all, they explicitly stated this in the press releases. John Baker (CEO of D2L) and Svein-Tore Griff (CEO of H5P Group) also confirmed it verbally at the H5P Ambassador meeting. The H5P Group is said to remain a separate entity creating a distinct product. Changing this later would likely anger many H5P enthusiasts and significantly undermine D2L’s credibility. On the other hand, since often open-source projects get shut down after a takeover, they could gain social capital if they prove doom-sayers wrong.
They probably would not gain that much
Looking at the current situation, one could conclude that there is no need to change anything. H5P.com is already profitable, earning about $1.5 million in 2023, with an upward trend, even while cross-funding the development of openly licensed H5P components. D2L may hope to attract more customers to H5P.com by closing the open side of H5P. However, H5P Group could have gone this route already, but chose not to. It seems unlikely that larger institutions such as European universities that currently use H5P on self-hosted platforms would migrate to H5P.com or Brightspace due to loss of trust, lack of funds, and organizational barriers.
Lower the barriers to migrate to Brightspace
D2L may have strategic reasons for keeping H5P open. Brightspace includes an add-on called Creator+ that allows users to create interactive content. However, from what I hear, Creator+ has not been well received. As a result, many users are said to have chosen H5P.com instead, resulting in a loss of business for D2L.
In addition, because H5P.com easily integrates with other LMSs, it becomes easier for users to switch from Brightspace to a competitor. This increased the risk that D2L would lose even more business.
Other LMS vendors, such as Canvas and Blackboard, which can also be easily connected to H5P.com, may have faced similar challenges. But the situation has changed. Now, it’s worse for them. D2L could try to use this opportunity to more easily migrate Canvas and Blackboard customers to Brightspace.
This strategy doesn’t necessarily depend on the open nature of H5P. If an LMS has a local integration for H5P (via a plugin or direct integration), it can host the content independently without needing H5P.com. D2L might want to eliminate this option for other LMS vendors. However, by keeping this option available, D2L can indirectly control a critical component of its competitors’ platforms, which in itself is a strategic advantage.
In addition, by setting the right incentives, D2L could potentially persuade some users to choose H5P.com. With H5P already integrated, users may find it easier to migrate to H5P.com, and ultimately D2L could convince them to migrate to Brightspace.
Keeping parts of H5P open would support D2L’s strategy of “infiltrating” the competition. It would be beneficial to maintain a clear distinction between the “open H5P”, “H5P.com”, and “H5P in Brightspace” experiences to ensure that users do not get “everything for free”. This approach sounds like something that would, of course, not be well perceived by the “open H5P” community. But the situation is nothing new that arose just now.
A content store, maybe
H5P already has the H5P OER Hub, which can be used not only on H5P.com, but also on Drupal, Moodle, and soon WordPress (the code is ready for review). It could also be included in other H5P integrations like the plugins for ILIAS (a learning management system mainly used in Germany and the Netherlands), Stud.IP (mainly used in Germany), TYPO3 or even Edlib.
The H5P OER Hub allows users to openly share H5P content with a global audience and conveniently reuse it in their workflows. This hub could be expanded into a store for H5P content, similar to an app store for mobile phones. D2L would not have to produce the content, but could allow publishers to sell it and collect royalties. D2L would control the infrastructure to reach instructors directly through various learning management platforms. Both D2L and the publishers would benefit from distributing this infrastructure across as many platforms as possible, including competing solutions such as Moodle. A good way to achieve this is to use openly licensed source code.
To support this idea and encourage publishers to create more content, it’s beneficial for them to have easy access to content types that enable them to create high-quality content. Therefore, existing content types should continue to be improved and new ones should be made available to as many people as possible. This is another reason to embrace openness for H5P.
And before someone wants to lecture me why this would be impossible: Yes, I am aware about technical and legal issues that one would face when pursuing this. You’d need lots of resources to tackle this – something that H5P Group didn’t have …
Finally, the finale
For me, the most relevant sentence that Svein-Tore said during the H5P Ambassador seminar was:
“I think, no matter what we say today, you won’t be completely comfortable by it. But we’re also going to follow this up with actions. You’ll be seing D2L investing into H5P, you’ll be seing us keeping everything open as is, and we also hope that we’ll be able to make things noticeably more open and improve some of the things that we haven’t been able to do so well today since H5P Group is just a small company with 30 employees.”
Let’s judge the new situation by what’s going to happen then, not by what doom and gloom our minds come up with or by what we believe might happen. It is pretty safe to say that D2L is not going to make all parts of H5P software open source 🙂 However, I don’t see why they couldn’t have a strategy that works well with maintaining the current status and even investing in it.
As I mentioned myself, I am biased. I used to work for Joubel for one year, and I still know many of the people who work at H5P Group today in person – and I trust Svein-Tore in particular. That doesn’t rule out that D2L might break their promises and screw H5P, but that’s something out of my control, and I’d cross that bridge when I reach it. Also, my business as an IT freelancer is solely built around H5P and requires it to be open, or otherwise I need to do something else. So I better show some optimism, right? And if I can, maybe you can, too.