Learning experiences at scale, a case study presented by Sarah O’Keefe
Ready to learn how robust content operations keep pace with evolving and complex learning demands? In this webinar, Sarah O’Keefe, the founder and CEO of Scriptorium, describes the successful implementation of a component content management system (CCMS). This project was for a major organization that supports technology professionals with training and certifications.
The level of interest and commitment that we had from the client’s team was a big deal. They now have structured learning content. They have the scalability and reuse they needed and could not get any other way. We aligned their content ops with their business goals of scalability, reuse, and time to market.
—Sarah O’Keefe
Resources
- Looking for more information about this webinar? Read our case study.
- Get monthly insights from Scriptorium by subscribing to our newsletter.
- Check out other webinars about content operations, structured learning content, and more in this YouTube playlist.
Transcript:
Christine Cuellar: Hey, everyone. Thank you so much for being here today. Our webinar today is Learning experiences at scale, and this is a case study that’s being presented by Sarah O’Keefe, the founder and CEO of Scriptorium.
Sarah O’Keefe: My name’s Sarah O’Keefe. I am the CEO of Scriptorium. I’ve been doing this, as some of the younger people in my life like to say, since the 1900s. I’d like to point out that it was the late 1990s, but apparently that doesn’t make it any better. These days, we’ve been through a number of labels for what it is we do. We’ve always talked about publishing and automation. These days, we talk about content operations. How do we automate? How do we fix your processes? How do we get your content into a usable, sustainable, governable kind of situation so that your organization, whatever it may be, can deliver on enabling content?
So we do big projects for very, very large companies typically. Certainly, we’ve done some work with smaller ones, but our generalized customer is a very large company with a lot of technical debt and content processes.
And so today, what I want to do is talk to you about a project that we did for a particular organization which came to us with a set of requirements that were related to learning content specifically. And they had some really, really interesting problems that may resonate with those of you that work in learning content specifically. So we’re talking here largely about technical learning content. How do I use this piece of software, and what does the training or the e-learning or the classroom training look like for that kind of thing?
So our client is a global provider of vendor-neutral training and also of certification related to the training. The project that we did specifically was not on the certification side. So we are not talking about certification exams with all the attendance security issues, but rather the test prep side of the world. So if I’m going to take this exam, what do I need to know? What are my learning objectives? And how do I, as a certifying organization, provide you with learning resources to understand this content and these topics so that you can pass the certification exam?
Now, one of the interesting things about this to me is that that means that for this organization, unlike, I think, probably many of you on this call, content is actually the product, because as an organization that provides test support, test prep, testing services, the actual product of that organization that they sell is content, in this case, learning, e-learning, and other kinds of training.
Additionally, one of the other things that made this pretty unique was that they do not have technical documentation. In many cases, when we talk about corporate environments, it’s, well, first we create the tech-docs, and then from that we create the training content.
Now, in this case, they only have learning content. There’s no underlying technical documentation, because, again, when we’re documenting a product, you sort of write the documentation for the product, and then, potentially, you create the e-learning sourced off of the technical documentation. So there’s a decent chunk of overlap between technical content and learning content when you’re documenting, particularly a software product.
However, in this case, no technical docs, which meant there’s no opportunity to collaborate with the technical documentation people, because they don’t exist, and there’s not that sort of concept of, “Oh, the learning is downstream.”
Now, we can have a long and entertaining, at least for me, discussion about whether learning content should be downstream or whether that should be more of an in-parallel, collaborative, in-sync kind of question. But for these guys specifically, learning content only. And what they ran into was that they needed scalability. They had huge issues with the content itself, like creating the content, keeping track of the content, managing the content, putting it all out there. They had huge issues with localization, getting it translated, because, in fact, this is being delivered worldwide. And then also with output and deliverables, all the different formats that they needed, which, big picture, e-learning content, but also classroom training. And we’ll talk about a little bit more of that as we get into it.
So this was the sort of original workflow, and the details kind of don’t matter. The takeaway here is that everything was fragmented. And you see all those yellow diamonds, those are all learning assets sitting in lots and lots of different places. And there were all these cases where things needed to be built and then moved and then copied and then refactored and this, that, and the other thing, and it was not fun. So that was kind of what it looked like.
And the reality is that, and many of you will find this familiar, that wasn’t what the actual workflow looked like. This was the actual workflow. So it is just … I like to call it spaghetti. I mean, this is actually a yarn tangle, but yarn spaghetti. Just an enormous, very, very difficult to untangle kind of situation where it’s very, very hard to trace back, where did this originally come from? Where was the source? Where did it start? You might think of it as a river system and you’re looking at a drop of water and trying to figure out, well, which creek did this come from? We’re down in New Orleans, and we’re arguing about which river in Minnesota, that kind of thing.
Okay. So these were the requirements. I already mentioned scalability. Just broadly, we need more. We need to do more content, more formats, more deliverables, more languages, and how. So that just, big picture, we’re barely hanging on by our fingernails in the current situation, and we can’t just keep adding and adding and adding and adding people. So how do we scale? How can we expand what we’re doing?
Additional content and content types. There was an acquisition, so there were new workflows and new content from the acquisition and some questions around how to combine that. There was a requirement to do more localization, more languages. Then there were some really interesting problems around licensing, because it turns out that these courses are, in some cases, delivered to educators.
So for example, a community college might license the course material to use in one of their classes, which means that as the creator, I now have to package up this course material and license it downstream to the community college and keep track of that because it’s still my intellectual property. So how do we deal with IP?
We had reuse across courses. So the example I always use here is if you are a company that makes database stuff, then somewhere you have a topic that’s like, what is a database? What is a relational database, and how is that different from other kinds of databases? And probably most of your courses start with that foundational you need to know these things about databases before we explain to you how to use our specific database. And then so they had reuse and some really problematic reuse, but especially needing concepts to travel. And there was an issue with time to market, wanting to get stuff to market faster.
Now, I don’t know about you all, but that one almost feels … time to market and scalability, actually, really many of these issues are just every project we do has these. Nobody ever says, “Oh, we’re fine, we’re good. We can take as much time as we want. And the amount of content we’re doing is decreasing.” That is not a thing. Everybody’s content load is increasing. It’s just a lot. So this was kind of where we started. And then additionally, this was all for learning content, specifically as opposed to other things.
Now, the interesting issue here is that structured authoring, which is what we’re going to talk about today, is actually very rare in learning content. Most of the learning content environments are unstructured, by which I mean you’re building content that is locked into a particular format, and there’s a lot of freedom and flexibility in terms of how that gets built and formatted. You have a template in a PowerPoint or something like that, but it’s just like a light suggestion, not a requirement.
And ultimately, the big difference between structured and unstructured is that structured authoring provides a template and requires you to follow it. There’s enforcement baked into the software. Okay. So I want to talk about why, but before I do, Christine, do we want to talk about polling results off this first one? What kind of people do I have on the call?
CC: Well, first off, I was just going to actually jump in to remind people we have the poll open, so we don’t have enough poll results yet. I’d like to get a little bit more feedback from the audience first.
One question I do have, though, regarding what you were talking about when you talked about the requirements is that are you able to share a little bit about who internally at the organization was starting to recognize … You mentioned an acquisition, you mentioned localization requirements, you mentioned reuse. Are you able to share how the organization identified these requirements and problems, and how that was kind of advocated for to kick off this journey?
SO: By the time they got to us and reached out to us and said, “Hey, we have this problem,” they had actually done a ton of work already. They had done a bunch of research, they had looked into a bunch of different options and had really done a lot of the legwork to understand the problem set that they were facing and what they were running into. This went up to… I don’t know what the exact job titles are or were, because, again, content is the product, it was sort of the chief product person who was responsible for, “This is the product we put out. How are we going to do it?” They also had somebody responsible for content at a very high level, at sort of a VP level, which is unusual, again, because usually what happens is that if you are a content person, on the technical side, you probably report up into something like engineering or if you’re on the marketing side, into the CMO content marketing person, and learning content, it really, really depends, but very often it’s some sort of a customer success person. In this case, the people at the top of this sort of executive food chain were content people, because that’s what they do, because that’s what the organization does. And I think that it is fair to say that that was very, very helpful in getting this done.
So why is structured authoring so rare for learning content? And for us, coming from largely product and enabling content, we tend to look at this through a techcomm lens. And so if we talk about technical communication, it and learning content, or if you look at this project, had a lot of the same goals. You want to manage your content, publish lots of outputs. You’ve got localization, you’ve got variance, you’ve got reuse, you need to do faster updates, you need to go vroom, you have greater consistency and maintain the common core and improve the quality. I mean, this is just a laundry list of standard requirements.
But what happens is that when you get into learning content, the thing that really differentiates learning content from technical communication, so tech-docs, is the focus on the learning experience, the question of what is the learning experience for the learner? When I deliver this downstream to you, Christine, and you’re taking a course, what does it look like to take that course? What does it look like to do that e-learning? And if I can’t give you something that is compelling, then you’re not going to learn anything, and then I failed.
With techcomm, typically, the focus is on delivering the information, but the responsibility of consuming that information is put on the end user. I sent it to you, I optimized the search, I did the keywords, I gave you a procedure, it is well written. Is it the most beautiful thing you’ve ever seen? No, it is not. Do I care as a tech writer? I’m supposed to care and I should care, but ultimately, techcomm has been very focused on the efficiency and the back-end side of things, and learning content has been focused more on the delivery and the experience side of things.
Now, I would argue that we could probably learn some things from each other, meet in the middle, but that’s kind of the issue. And because of that difference in focus, we have a gap in the software offerings because everything techcomm is efficiency-focused and everything learning, to overstate it, everything learning is delivery-focused. So what we run in-
CC: And Sarah? I’d love to jump in actually with some poll results as well. Speaking of everyone that’s on the call, it looks like we have about 45% of our folks today that are in the learning and development space, then about 55% in the techcomm space. And we have a few marketing and support people. So good to know that we have a good base for who’s viewing this, especially in regards to what we have for software options.
SO: Great, thank you. And I think we allowed multi-select, so there’s probably some overlap, which would be actually very interesting because that is a rare unicorn of an outcome.
So the learning experience tends to be tied to outcomes, which I mean, that’s a good thing, we’re for that. Additionally, training is often tied to revenue. Now, for our client, it was very directly tied to revenue because their training gets sold and results in revenue, but even in more of a software-hardware product organization, very often in the training is considered a revenue source. It gets sold to the customers. It’s not something that comes with the product necessarily. Sometimes it does, but there’s much more of a, “Oh, if people are doing training, we’re getting revenue,” kind of focus.
Learning experience, again, generally focused on quality and not efficiency. There’s a requirement for learner records, this issue that I need to keep track of whether you have done your mandatory annual compliance training. I can’t just deliver a document and say, “Here you go, Christine. Here’s your document.” You have to actually read it and then pass a quiz that says, “Yes, I understand that I will not do these bad things.” And I have to keep track of that either for compliance purposes or maybe certification or continuing education or just, generally, our company has a policy that you have to do this compliance training every year.
So learner records are a thing that it’s not really analytics; it’s actually a record of, did you consume and understand this content? Did you pass the test? And that is something that we don’t really do in techcomm, it’s rare. And so it’s a requirement of some of the learning software to capture that, and we don’t really have that. If you think about education, whether K-12 or higher education, a requirement for a grade book, did you do your assignment, and did you turn it in, that type of thing. There’s none of that on the techcomm side, or if there is, it’s very, very rare. And then this last bullet, well, it’s not the last bullet, but it’s the last bullet I want to talk about is this relatively late adoption of digital deliveries.
Now, in techcomm, in the mid ’90s, in the 1900s, we went from printed books that got delivered with the product and started adopting digital content. Initially, that was literally, “Here’s a PDF version of the book. And hey, you customer, you can print it if it makes you happy, but we, the vendor, are no longer going to print the 600-page book and ship it to you. We’re going to make that a you problem.” And that was a … I mean, I was around for this. This was a very literal, “We don’t want to have to pay. If you want a printed version, you’re going to have to pay for it.” And a couple of companies tried, “Here’s the PDF. We’ll get you a printed copy, but you have to pay extra.” That didn’t really work. So the actual option for print kind of fell by the wayside very early, with the exception of people who are required to deliver print. So if you have, again, a compliance reason to ship print, then you do. But what you would see is that the required printed thing got skinnier and skinnier and skinnier and smaller and smaller. And now it’s like four-point type, 28 languages, that’s nothing but warnings, and the actual useful information is online somewhere, because the only thing that’s getting shipped is the mandatory thing.
Now, again, I’m oversimplifying because there are compliance issues here and this is different in the US versus Europe. So Europe tends to ship more print relative to the US because of compliance issues. But that’s kind of the state in techcomm. PDF comes in, the idea of online help, online portals, this and that, that’s like early to mid 90s. And now having people ship a lot of print or print as a primary thing is, it’s nearly always part of a bigger strategy. We’re shipping some print, but not a ton.
Okay. Now, let’s talk about training and learning and the learning environment. Relative to that, the adoption of digital is or was relatively later. And there’s a couple of different reasons for that, but at the end of the day, and this is what fascinates me, if you think about a printed book versus an online portal with HTML in it and all your content, there are certainly ways in which the book is better than digital. You can carry it around with you if the power goes out, you don’t need those kinds of things. You don’t need electricity to consume it, but there are advantages to the digital delivery. Like if I’m a field service tech, I can carry all the books on a tablet. I don’t have to carry this many books. Aircraft carriers used to measure their documentation in shelf feet.
CC: Wow.
SO: That was their unit of measurement. And when you’re on a ship, even a big one, it matters. So having it on a CD or obviously in the cloud, but you want to be careful about that, because what if you can’t get to the cloud because you’re in the middle of the ocean and your internet went down? Okay. So you have a CD. Well, versus 18 shelf feet of documentation. So maybe you have one paper copy and CDs and electronic versions. In learning content, the advantages, the pros and cons of e-learning versus classroom learning, that is a much more complex conversation. And I would argue that at the end of the day, putting a group of learners into a classroom with a really great instructor is always going to be the best learning environment.
Now, it’s not always the most feasible learning environment, for reasons that we’ll get to, but ultimately, classroom training with a great teacher, all of us remember our great teachers. Does anybody remember some e-learning experience that was fundamentally life-changing? Yeah, maybe not. I mean, there’s some good stuff out there, but.
So fundamentally, classroom training is better, but there are constraints. People have to travel and travel can be a real problem. And you may have missed it, but we had this huge issue called COVID where travel was not a thing, which forced people into online learning of some sort. Now, it could be online instructor-led or it could be just flat out e-learning. But what happened was that when COVID happened and everybody got shut down and sent home and isolated, we didn’t really have the option of doing classroom training for a couple of years. And now, even as people are pushing return to work, return to … not return to work, return to office, that the departments have gotten very fragmented. You see all these stories about people that are told come to the office, but they sit on a Zoom call all day because all of their peers are elsewhere. Well, if everybody is in different locations and you need to put together a class, you either have 10 people travel plus the instructor or you just do an online thing of some sort. And so that really has pushed us into online.
Online, maybe asynchronous, which is to say e-learning or instructor-led, but that digital approach has been relatively late because it’s really hard. It is really, really, really difficult to deliver a great class online and to make it stick. So there was a relatively late adoption of this move to digital. Partly it was forced by COVID, and now it’s sort of picking up speed, because, ultimately, it is a lot cheaper to deliver e-learning or online instructor-led because you cut out the travel. Okay. So these are the sort of issues that have pushed back on digital for the learning experience. And this relatively later adoption means that now we’re in a relatively early stage in terms of doing digital delivery.
So complex deliverables. When you think about training materials, here are some of the things that I think about, like instructor materials versus student materials, test answer key versus test. PDFs that need to be delivered for compliance purposes, they’re not the live training materials; they’re what you have to ship to the regulator to get it approved. There are course variants, and those can be pretty sophisticated. And then we get into questions like adaptive learning, of going a little more in depth into a topic because this particular class is not good at that or extra interested, and so we just sort of veer off into that tangent.
We’ve got SCORM and other kinds of deliverables. So SCORM is a standard-ish that allows you to package up an e-learning class and deliver it to different learning management systems. But actually, delivering SCORM that works is hard because it turns out that every learning management system wants it set up a little bit differently. So you end up targeting different systems, which is a big old pain. And our particular client also had SCORM for licensees. So this community college model I was talking about, they needed to package up their content and deliver it downstream to their customer, which was not necessarily the end user, but rather the educational institution that’s sitting in the middle of that.
API connectors. If I want to control the intellectual property, then maybe what I want to do is put all the learning content inside some sort of a bucket and then have an API that connects to that. And if you’re not authenticated, you can’t have my content. So lots of stuff going on in there.
So when we break this all down and we think about what makes up learning content, it looks something like this: You have lessons and you have assessments, questions, test questions, that kind of thing. You have scenarios, you have learning objectives, you have glossaries, terms and definitions. You’ve got how do I do this thing? That’s basically a task, which we’re very familiar with from techcomm land. You have what is this thing, which is a concept. And then you have simulations and animations. So I’m going to run this thing in some sort of a simulated … my software, in a simulated environment where I can’t break things.
The most famous example of this probably is flight simulators, but there’s a lot of simulation software for you can’t just go around trying out network security settings, “Oh, what happens if I push this button?” No, no, no, no, no, no. That’s how we get like you come in on a Monday morning and nothing’s working because the cloud services are down. So simulation, love it.
Animation, video. So these are all things that go into learning objects. And then when you think about these objects, they get packaged up. So lessons consist of all those other things. And then you have all these different kinds of delivery types, instructors, student guides, PDF, this and that, and you have variance. And with variance, I mean, I’ve put in a couple, but these are very, very high-level.
But in particular, the audience. If I have database administrators taking a class on my particular database, they don’t need database concepts. They can just skip right past that. They just need, “Here’s how my thing is different.” So we start thinking about how do I store these objects? And then you get into the acronyms. We have LMSs, learning management systems. That’s where you store and/or deliver the courses. You have your learner records, your grade books, your quiz results and potentially you’re learning paths, like, “Oh, you already took this other course, so now we’re going to pass you through this one.” Or, “You passed, but it was like 67% out of a minimum 66. And so we’re going to give you some more information in the next course because that didn’t look great.” So that’s the LMS. The LCMS, the learning content management system, was where you create and manage your objects. This is a back-end system for authors. And I think it’s really important to understand the difference between these two because some systems will do both, and then we get ourselves in trouble. Some systems only do the back-end stuff. And when we talk about techcomm stuff, then almost always we’re talking about back-end systems. All right. So you have-
CC: Sarah, sorry to interrupt. This seems like a good time to jump in and talk about the poll results for our second question because we asked about our audience’s learning content stack. And the vast majority is general office tools. They’re using things like PowerPoint, Word. About 18% are using some learning development tools in addition to that, articulate, captivate, similar. We do have a couple of people in a learning content management system. No one on the call is in an LMS, at least that said so in the poll. And then just a handful of people in a CCMS as well. So yeah, just wanted to throw that out there. So it looks like these are good concepts to talk about today.
SO: Yeah. And I’m surprised not to see more learning development tools. I am completely unsurprised by the rest of it. But yeah, I mean, PowerPoint rules this world still, and we’ll talk about that in a bit.
CC: We do have a question, and someone’s asking if an LCMS is the same as a CCMS. I think you’re going to talk a little bit more about this, too.
SO: So it could be, maybe, but broadly, it’s more a question of audience focus. And by audience focus, I mean the people making the LCMS or the people making the CCMS. The LCMS people are targeting learning content and the CCMS people are targeting techcomm content. And I think you’ll see where I’m headed with this, is that our client decided to use a CCMS as an LCMS because they couldn’t make the other tools work for them. The learning optimized tools is what they ran into. They just could not get it to work at scale for their requirement. So on the back-end, create and manage the learning objects, create and publish courses, track images, do all the things. And then you sort of have this front-end, deliver the courses, keep track of the learner records. Now, learner records, front-end or back-end, it’s not authoring. Let me put it to you that way. The learning paths, arguably, are authoring because somebody has to think about the learning path and put it together.
So here we are, can we just use the LMS? Well, you can, but the problem you run into is that while many of the LMSs, so now we’re talking about something like Moodle or Canvas, many of them do have the ability to create and publish courses, but they don’t have the ability to create and manage learning objects. So you have a course and you can duplicate it and you can edit it, but there’s no real management of those objects down in the weeds. It’ll do all the other things.
On the LCMS side, and this was specific to our customer, they looked at this. They looked at using a dedicated learning content management system for their authoring. And despite the fact that they only have learning content and they don’t really have techcomm clients to consider, what they ran into was that the systems that were available to them didn’t meet the extensibility and scalability requirements that they had. So they tried this, or I mean, they didn’t build it, but they looked at it and they said, given the output pipelines in particular, the publishing requirements that they had, the LCMS didn’t have enough flexibility and enough options downstream to do what they needed to do.
And so stepping back from this for a second, there’s this gap, and this is what the person asking is calling out. When you think about learning content and you think about these different systems, the CCMS is not optimized for authoring learning content. Some of you will argue it’s not optimized for authoring any kind of content and it’s sort of awful, but not optimized for learning content authors at a minimum. The LMS doesn’t really do content management, not at the level that we need it or that our client needed it. The LCMS has a tendency to be sort of locked in. And if it solves the problems that you’re trying to solve, it could be a really good option for you, but if it doesn’t, extending it can get very problematic. And then automated publishing, which we’ve been talking about, again, on the techcomm side for a very, very long time, tends to compromise the learning experience if you’re not careful.
So those are the gaps that you see when you look down this tech stack and at all these different pieces and parts that you’re dealing with, and the question then becomes, where do we compromise? Do we compromise author experience or extensibility, or what do we do? And so that was kind of where our customer was when they reached out to us.
Now, by the time they got to us, they had pretty much gotten this far, they’d already done all of this. They looked at the LMS, the learning management delivery systems, and said, “No, this isn’t going to work because the authoring is not efficient enough for our scalability bandwidth requirements.”
I should also tell you at this point that our client had and has instructional content creators that are extremely technical, because they’re building technical training. That is an important piece of information, because they were pretty happy with, “Ohm yeah, it’s kind of weird and nerdy, but it’s great. We can do it.” That’s not the case for everybody. The LCMS, which I noticed I’ve spelled incorrectly, so that’s fun, had issues with the sort of level of flexibility that they needed and extensibility. They had some, admittedly, unique publishing requirements, and so they had some issues there. They looked at CCMS options after ruling out all the LCMSs and finally decided this is what we’re going to have to do and chose DITA.
Now, the very, very first conversation we had, they said, “We think this is the way to go, but we’re worried about the authoring experience.” I mean, I remember that very clearly because what we told them was, “Well, let’s get into this. Let’s do a proof of concept, let’s see how it goes, but let’s acknowledge that this is going to be a big change and a big lift, so we need to be careful.” So they made this decision before we ever got involved. This was not a matter of me talking them into it, although I’m perfectly happy to try, but in this case, by the time they got to us, they had done the hardest work and what was left was we’ve made this decision, now let’s figure out if we can implement it successfully. So it’s been actually really fun.
This was a different client, but this is sort of the universal attitude that people come in with: “We are doing too much work. We are running too hard, and we are not making any progress. It’s just there’s too much busy work.”
So with that in mind, let’s talk a little bit about the challenges that we ran into in building this. These were the day-one problems: multiple fragmented delivery channels and formats, fragmented workflows, a stunning number of content silos across the organization, just content parked in different places that we could not get to. Some big issues with localization, especially for Japanese, which is and was a major customer. So the Japanese version of this content, which needed to be at a very high standard, was a major revenue source and is. And there was a requirement to continue production during the transformation process, which is … I don’t know if I even need to mention this. This is just a thing that everybody does. Nobody ever says, “Oh, yeah, we can stop for six months. It’ll be fine.” So these were the big-picture things, and this is a big and challenging list of stuff.
One thing you don’t see here is the number one challenge that we run into in all of these projects ever, which is change management. We had almost no change management issues because everybody was on board. They had already made the decision internally. We were not bringing that decision in. And again, it was a very technical team that thought this was going to be fun, and they’ve had a great time building out stuff and working through it and doing all the things they wanted to do. So team-wise, this is what the vendor and client stack looked like. On the inside with the client, we had our learning experts and our domain experts in the sense of people that are experts on this kind of content and how it needs to be built and delivered and formatted and all the rest of it. Data Conversion Laboratory, DCL, came in to do the migration from the legacy formats into DITA. The CCMS, the component content management system, that they eventually chose was Heretto. And then my team did the DITA architecture work, information architecture, configuration and publishing pipelines.
Our client also had a lot of technical expertise and has become quite adept at those last four things. So over time, we’ve done some knowledge transfer, ironically. There’s nothing actually more stressful than giving training to trainers. That is a terrifying thing to do, but a lot of that, that bottom thing, we did a lot of the initial work, but it has now moved up into the client’s domain.
The migration, I believe that’s done, I think. So now it’s basically the client, Heretto, occasionally we get brought in to help out, but they are self-sustaining.
So we did some content ops work, content audit, what do you have? What’s the inventory of what exists? What are we going to carry forward? We did some content modeling, what does it look like to capture all these learning objects in DITA in this case? Was the system architecture look like? How are we going to take the CCMS, which kind of sits at the core, connect it to all the other things? We ended up with a centralized repository with some modular stuff and a workflow that supports reuse and localization largely in an automated fashion as appropriate.
They built on DITA. They used the learning and training specialization, which is another DITA layer, basically, that allows for quiz questions, assessment, learning objectives, all those learning-specific things. And then we discovered that we needed to do some additional extension or specialization for assessments, for questions. It turns out that what’s in DITA out of the box didn’t quite go far enough for the types of questions and assessments that we had in our requirements.
In terms of Heretto and system architecture, so there’s the Heretto component content management system, there’s a digital asset management system for images and videos, that kind of thing, there’s a translation management system for localization. We have publishing pipelines. We have multiple learning management systems that are consuming this information downstream. And authoring was done, it is done in the Heretto web editor. And then for the power users, they’re using oXygen. There were day-one questions like, “Oh, well, if I just download the files and write myself a little script, I can just vroom and run through all the files and make changes,” which, again, is great if you can do it and not a normal thing that we hear from instructional designers. So I really want to stress how not just engaged, but how technical these instructional designers are. So that was a big part, I think, of the success.
All right, so how do we build this thing? The CCMS, the component content management system, in this case, Heretto, is where the files are stored, the text files, and where the course authoring is done. You then have a digital asset management system, which is where the assets live. When we say assets, we mean images, video, that kind of thing. So non-text assets. All right. From the DAM, it goes into the CCMS. The TMS, the translation management system, is how you manage the localization workflow. So I’m working on English, it goes out for translation, it comes back as DITA files in Japanese or French, Italian, German, Spanish, and gets stored in the CCMS in that language.
There is and was a separate layer for labs, simulations and activities, that was not integrated into this process. So they are sitting outside of that. There are links, but the links go over to the LMS. So from the CCMS, we now have all of our assets, text and linked images in the CCMS. You publish from there, it goes out to the LMS. So this is SCORM, potentially, sometimes PDF for course delivery, and then the learner records are managed in the LMS.
In addition to the LMS push, and it’s actually multiple LMSs, there were also … it’s mobile-friendly, we have a PowerPoint output, which I’ll talk about in a minute, we have HTML, and we have PDF, and all of those are necessary for other deliverable requirements aside from e-learning.
So then there were some other things. xAPI is in here as a path to deliver things, and also LTI. So both of those were built into this process. And then, and this is maybe unique, they have downstream customer learning management systems, including Canvas, Moodle, Blackboard, and others. So we had to pay attention to the question of, will our content go there? Will it work? Will this particular packaging of SCORM work downstream? All right. So PowerPoint.
CC: I enjoy that slide.
SO: PowerPoint is a problem. And the reason that PowerPoint is a problem is because structured content has to follow a template and be predictable. And I don’t know about you, but PowerPoint is the exact opposite of that. I mean, I’m running a set of slides here that would not accommodate being in structured authoring. And so if you can get yourself to a point where your PowerPoint is pretty well-structured, it has headings and it has bullets and maybe an image occasionally, that type of thing, you can get there, and you can get from the sort of CCMS back-end to your output, but you can’t do arbitrary PowerPoint. You just cannot.
And so this is always an obstacle, because the question is, well, how do I deliver … if all the training is PowerPoint-driven, how do I do that? Now, I would argue that if we’re worried about learner experience, we should not be delivering slideware, and we should minimize the use of slides. And if it’s a few that are kind of predictable and we can kind of work with that, then great. But this is always an issue, because if your primary deliverable is highly customized PowerPoint that is visually all over the place, then we’re going to have big problems with structured content.
Now, you can separate out the PowerPoint and just keep doing that. You can source it from the structured content. There’s a bunch of things you can do, but just keep doing what you’re doing in PowerPoint, and still gain structured authoring benefits is not one of those options. So that’s the issue here.
Realistically, you can do something like this. You can have an image and you can have bullets and you can have headings, and you can go a little beyond this and have a couple of different kinds of formats and tables and various other things, but you’re just not going to get that wide-open formatting that you’re accustomed to, even if you’re starting from a template in PowerPoint. We built out some simulations animations, or I shouldn’t say we built them out, those exist. They are not stored in the structured content world, but rather inserted like an image, and you can do some interactivity. For our client broadly, these were generally done separately in their lab systems. So when you publish out to the LMS, the LMS stores the learner records and the learning paths, and then you have the requirement to deliver. This was tricky. Delivering to all these different things and making sure they all work is tricky, but what you want to do is really separate out the learner records and the learner behavior kinds of things from the core content. That was really the key.
I want to talk briefly about migration and how this worked. So the starting point was pretty messy, as it so often is. And we had a lot of different formats and a lot of content duplication. Again, DCL was the one that did the work. One of the things they did in collaboration with our customer was they brought in a tool called Harmonizer, which will look at the source content, the legacy content, and identify duplicates or near-duplicates. And what that helps you do is reduce the amount of content you’re actually converting in. You can then build out, or, well, they can then build out an automated migration system, which they did. And we worked with them to make sure that we got that exactly to where we needed it to be so they would migrate the content out of all these different formats into DITA, and then the DITA content got brought into Heretto. There was ongoing production during migration, so the instructional designers, content producers, and all the rest of it were still producing in the legacy systems before we kind of went live with the new stuff. I think if you ask them, they’d tell you, it was maybe not as painful as they expected, but it was a lot.
So down the road now, we have centralized content in Heretto, we have reuse instead of copy-paste workflows. There were a ton of copy-paste workflows. “Oh, well, this content needs to be in A, B, and C locations, and we’re copying and pasting.” We have consistent global delivery, localization is faster and arguably better, higher in quality, and the content creators now are able to shift their focus to quality, focusing on the content quality instead of the formatting quality. So instead of formatting problems, they’re focused on content quality, which is great.
The lessons we learned, I would say that, again, the level of interest and commitment that we had from the team was a big deal. So that was one. We did get this to work. We had structured content. We got the scalability and reuse that they needed and could not get to any other way. We aligned their content ops with the business goals, which, again, scalability, reuse, time to market, the usual.
There was really great cross-team collaboration between us and our client, but also all the vendors. Everybody was pitching in, finding issues, making it work. And so it was a really great team, and they were really fun to work with. And I think everybody is happy with the outcome in terms of how this went and where we landed. So it was a great experience.
And with that, I think I’m going to wrap it up, see if there are any questions. And I’m going to put that resources slide up again. But Christine, anyone yelling at me yet?
CC: Not yet. So if you do want to ask Sarah a question, now is a great time to get that in before the end of the show. One thing I do want to point out is that for our last poll question, we asked how many people on the call were considering DITA for their learning content. About 40% said yes, they’re actively considering it, and about 25% said maybe. So total, about 65%-ish of the people on the call are thinking about it.
Sarah, my question for you is, what would you recommend for people in that boat? Maybe they’re just starting the process thinking about it or maybe they’re a little further down the line, what should they be looking at, especially keeping in mind that they might not have the same in-house technical kind of experience that you were mentioning this client had that was fairly unique, it sounds like? Where would you recommend people get started?
SO: That’s a great question. And I guess, I did expect a relatively high number of people on this call to be at least interested because if you’re a hard no, you’re not here.
CC: It’s fair.
SO: I mean, I guess you could be hate-watching, and that would be okay. We don’t judge. We all need our entertainment sometimes. If this is something you’re looking at, first of all, probably you have a tech-docs organization, probably. And you may want to see if there’s some opportunity for a drag-along. Are they already instructed content? Do they already have some of these tools? Is there potential that you could piggyback in there and implement some of the reuse that we were talking about that you could be able to get out of the techcomm, again, if it exists, which was not the case for this particular organization? So that’s one.
Two is, I think the most critical thing is to do the work internally to talk to your people and understand what are the pain points in terms of content development right now and is there anything that you can offer that will be better if you go forward? Because the reality of structured authoring, in general, is that there’s a lot of … it’s not easy. There’s a lot of challenges in terms of learning the tools. And so if I just give you those tools and say, “Do this because it makes the content better for the organization,” but it doesn’t make my life better as an … or sorry, your life better as an author, because in this scenario, I’m the awful executive, well, where’s the upside for you, the author?
So I think really think about what would it look like to bring this into your organization? What would the change management look like? Is there interest? Are people really annoyed with all of their copy and pasting that they have to do? That’s usually the entry point. We can get rid of the copy and paste, but it’s going to come at a cost. We’re going to have to be much more mature about how we create content. That’s the trade-off. If people are ready to do that or at least ready to think about it, you can probably get there. So that’s one. Look at your authoring teams and who they are and what their appetite is for something like this. Really, this is a 360-degree problem. So talk to your authors, talk to your upper management. Is this something you can sell to upper management? What are they thinking in terms of systems and tools? And do they know that, assuming you’re facing similar kinds of problems, do they know about that? So rather than, “Hey, we need some software,” it’s, “Hey, we have this problem in our content ops, in our learning content, and the reason everything is six months late is this.” Start that conversation and see what kind of appetite you get for those kinds of things. Talk to your peers.
So if you’re learning content, talk to the techcomm people, talk to the support people, all the other customer-facing enabling-content people and kind of see if there’s something you can do informally to get things started. And finally, I would take a very hard look at what does your AI-enablement strategy look like? I got through 58 minutes before I said AI.
CC: I was going to say.
SO: What does your AI-enablement strategy look like and what is it going to mean to do some of this work or to have AI support some of the things that you’re trying to do? Because right now, if you can show that this will help with productivity gains, then a good AI strategy will probably help you get the funding. We have a ton of resources, and I am always happy to talk to you, just big picture, about where something like this might be going, but I would start with the question of where are you internally?
CC: Yeah, that’s great. Thank you, Sarah. Couple questions coming in. We have one that says one argument against DITA in e-learning is that it offers so little interactivity. Was that an issue for this customer?
SO: So what they did was they built interactive stuff, but they offloaded it into their simulation animation tools, and then you can either potentially embed the animation and get interactivity that way, or in their specific case, you send people off to … They don’t make flight software, but you send people off to the flight simulator where they do all the simulation and interactive stuff, and then you’ve also separately got the content, because ultimately, it’s not how do we build simulations in DITA, but rather how do we do the text content development more efficiently?
CC: Okay, yeah. We have another question here that says can DITA create editable PDFs forms that can be filled out?
SO: Interesting. So in order for a PDF to be editable, and now we’re at the edge of what I know about editable PDFs, but I think it’s all sort of underlying JavaScript. And so in theory, yes. In practice, I’d ask a lot of questions about why, but I think in practice, yes, you could pass the JavaScript through to the PDF rendering engine and modify the processing so that that JavaScript goes into the appropriate place to then enable the fillability and the interactivity. So in principle, yes. I don’t know that I’ve seen anybody do it that way. Usually, if it’s an editable form, you’re not doing them at scale, usually, and so those get hand-built, usually.
CC: Gotcha. Awesome. Yeah, well, thank you so much, Sarah, and thank you to everyone on the webinar today. Thanks for being here.
SO: Yeah, thank you. Thanks, everybody.
Subscribe to our newsletter to get monthly insights from Scriptorium
About the author
