Podcast Ep. 26 – Content modelling in the age of headless CMS and AI w/ Larry Swanson

In this episode, Noz Urbina sits down with content expert Larry Swanson. Together, they explore the intricate world of content modelling, the rise of headless CMS, and the importance of understanding user needs. Larry shares his journey from the early days of the internet to the current challenges of orchestrating decoupled, multi-channel content. In this episode, seasoned content strategists and newcomers alike will gain invaluable insights into the constantly evolving content industry.

What you will learn:

• The history and evolution of content modelling and its relevance today
• The challenges and benefits of adopting a headless CMS
• The importance of understanding user needs and shifting from a publishing mindset to a user-centric approach
• The role of metadata strategy in organizing and making content navigable.

Full session transcript

THIS IS AN AUTOMATED TRANSCRIPT

Noz Urbina 

Hello, everybody. Welcome to this episode of the omnichannel podcast by OmnichannelX and Urbina Consulting. My name is Noz Urbina. I’m here with Larry Swanson, who has an extensive background in all things content. And we have lots of lots of questions for him. We’re very excited to have him here. Larry, want to say hi to the group.

Larry Swanson  

Hey, welcome. Happy to be here. Thanks for the warm welcome and looking forward to talking.

Noz Urbina  

Awesome. Okay, so I want to I want to jump in we’re going to talk a Larry’s got a diverse, as I said, diverse set of arrows in his quiver. And as worked in many, many aspects of content. But we’re talking this year, a lot about content modeling. We’ve got the content modeling series going, and it’s it’s kind of an evergreen topic. I we’ve been I’ve been talking about content modeling my whole career. So that’s like the.com, boom, kind of 2023 years, we’ve been discussing content modeling, structuring content. Now we’re talking about all sorts of variations, like knowledge, graphs, and domain modeling and ontologies. And so, Larry knows a lot about all this. So that’s gonna be the kind of the main focus but we’re also gonna bring in a lot of headless CMS. So in the age of headless CMS, which is such a hot buzzword and very popular with IT departments and teams around the world, how to have all these things come together in an end Surprise rollout where you have all these things possibly mix in the mix in one environment. So Larry, let’s start off by talking. Can you tell us a little bit about your background, what kind of projects you’ve been on and what kind of roles you’ve had in those projects?

Larry Swanson  

Well, I’ll try to be brief. I’ve been doing this a long time, I go back to the I was browsing the internet before there was a web in the late 80s, early 90s, and all command line interfaces, and CompuServe, and all that old stuff. And then early on, I was working in book publishing at the time. So I’ve been a content person since the late 80s, early 90s. And even then I was exposed to like a headless CMS in the sense of like, there were these SGML based systems that handle the manuscripts, I was acquiring and put into production. So kind of been around this stuff for a long time. And even even during my publishing career, I was really interested in technology, I was always the guy teaching the computer class or showing how to use email or something like that. But as soon as the Internet came on, like the mid 90s, I left book publishing, did a couple of did about three internet startups kind of the mid to late 90s, early 2000s, then got into SEO consulting, which was largely about content and all this, all those startups I did, they weren’t the cool, like platforms, things that make you a millionaire, they were like crummy little publishing projects that, you know, make you poor your whole life, which is fine. I love content that much. And so anyhow, I got into a lot of SEO and sort of like internet marketing, kind of consultations with folks. And a lot of what I now realized, in retrospect, that was content strategy, consulting, even back then I just didn’t have the word to call it that. went from there, I did some of my own publications I launched as a entrepreneur launched a few things in the kind of mid to late 2002 1000s. Era. And then about eight or 10 years ago, I decided that I was kind of done being a publisher, I was much more identifying as a UX and human centered design practitioner. And so I sort of shifted my identity from publisher to kind of from a pushy publisher, like giving you stuff to like a curious designer going like, hey, what do you need? How can I help you, which is why the and so that’s in a combination of that as the last few years, I’ve really got into like service design, Omni channel, strategy, metadata, strategy, all the things that all the skills and tools that you need to take all this kind of crazy, new, decoupled, multi channel, multifaceted, multidevice content, and put it together in a way that helps those poor users.

Noz Urbina  

Excellent, thank you, I’m gonna jump in there and mention SGML cuz you dropped it dropped the four letter for letter one in there. So that was way that we were structuring content, your pre.com, boom, there’s many projects that lasted well into the 2000. And somethings, it’s rare to see in SGML implementation these days. But it’s very interesting, because a lot of there was a period where people were thinking, Oh, well, what, how could you possibly still be relevant? You know, if you were working on projects, way back in the yonder, younger days before we invented all the cool stuff. And then what’s interesting, what’s happening is we’re coming full circle, because in the early days of the Internet, we were we were doing everything we were going, there was no rules, there’s nothing, nothing was mainstream, yet there were no streams. And so we were doing all sorts of things. So structuring content and trying to do multi format and intelligent content, what we now call it semantic web, all that stuff goes back 30 years, even to the pre web days. What and what we’re seeing is, we went we know, we we ran with what was easy, you know, creating web pages, that was easy, everybody could do it, we just take pages and throw them up there. We put in like a word processor type interface, we could create pages, and now we’ve created this mess for ourselves, was too much spotty stuff, and nobody can find anything. I was actually thinking about this recently. I remember when the iPhone came out, it was we were also excited what there was a phrase, there’s an app for that.

Larry Swanson  

Remember that? Oh, I had forgotten that phrase. But you’re right,

Noz Urbina  

exactly. Because back will go oh, man, you know, there’s an app for that. We’re all excited to discover the new apps. And now we go crap is frickin app I have to download. So we we’ve gone from the excitement of Oh, wow, we can have web pages too. Oh, no, there’s too so many web pages. And so now we’re coming back to organizing all this stuff, structuring it making it filterable leverageable. So that we can get away from having to deal with pages and start to deal with knowledge and answers and you know, accessible value which you moves me along my journey. So for for me all all of us who you know, have go back 20 plus years in the web are bringing are bringing, you know, a very narrow was a very small community who have that kind of experience in how to manage something that isn’t web pages. So that’s I’m very excited about that. And you’ve teed me up nicely with your transition from push, you know, I’m going to publish which which what a lot of the thinking was doing I’m going to pop wish my pages I’m going to publish my knowledge. I’m going to publish my my messages or my, my materials, to how can I help you? You know, how can I understand you better rather than what I want to say? What do you want to hear? And what is going to be valuable so that we’re having a conversation about something that’s going to move this discussion along to somewhere that benefits both the person who has the content and the person who wants the content, which tees me up nicely to say that’s why Larry Swanson is now our latest member of the Urbina Consulting team.

Larry Swanson  

Yay, Larry. Thanks. Happy to be here.

Noz Urbina  

Yeah, so it’s, and I think that it has to do with that philosophical alignment. We’re really excited about this idea of building these relationships, through content that supports that dialogue. How do you rather than say, how do I, how do I get my stuff out there? How do I understand what the market needs and understand how to build content services, which are well designed for the requirements that the market has, it’s a different mentality. And I think the totally the future, which takes us to this idea of modelled content and headless CMS. So Larry, and I were about to kick off a big modeling project for a household name, which we can’t talk about yet. But you guaranteed all know them and grew up with them in your house, big publisher, and they’re in the classic situation, a lot of headless CMS projects these days, we’re going to we’re going to consolidate the zillion CMS is that we’re running around in the enterprise to one headless, uniform. Sorry, not uniform, the company one, one uniform, because there’s a company called uniform, what one consistent, headless CMS in the background, and we’re going to move everything in there. So I thought this would be a great opportunity to kick around the topic of modeling, but also how, what are the challenges and opportunities that are related to content modeling, in the headless project, because that’s something that we’re hearing so much about these days, we’re gonna move headless, we want to move headless, how do we go headless, and the two of us who have been doing decoupled and headless, you know, our whole careers, to talk about today with a current crop of headless systems, which are very popular. And there’s like a handful of names that are that are kind of becoming the new Darling children of the headless market. Everyone’s kind of talking about moving in this direction. But there’s a lot of issues there, and not a lot of experience in those issues. So I wanted to kind of start out what are the kind of the potential gotchas what people have to think about when they’re moving to headless?

Larry Swanson  

Yeah, I think the first one is, like, again, being old, I remember the the rise of desktop publishing, and everything. This is backup publishing metaphor that you could just publish whatever you wanted, you didn’t have to have an editor or publisher, all those things. I think it’s equally important now, just as like, we lost sight of the roles that were involved in creating a publication. There’s a lot of roles involved in orchestrating this new environment of like, really, truly decoupled content. It’s all about, you know, more back to the purpose of the content that, you know, like, typically, we’re gonna actually I want to do a quick aside, I just want to say real quick that you inspire the thought of me that like the tools that we need the foundational intellectual and standards based tools that can can guide a lot of this whether you actually adopt the particular standards, but like the DITA standard that arose like about, what 1520 years ago, you know, should have topically structured content. And then the semantic web standards, owl, and RDF. And all that is about W three C standards, that kind of gives you a way to think about how to structure stuff, and ascribe meaning to it, and put it back together in different ways. Again, whether you use it at a specific standards and technologies, I think the knowledge about how to do this as that comes out of our kinds of heads has been ensconced in places that we can draw on so so I like our odds, you know, as we tackle products and projects like this, we have some good good models and skills to draw on. But there’s still a lot of challenges there, you know, the, and just kind of from beginning to end, like you’re going to be authoring challenges. Authors are used to like WordPress and Squarespace and Wix and, and WYSIWYG tools to just that where your Microsoft Word or Google Docs. Exactly, yeah, were any of those where you what you see is what you get. And that’s going to be that’s not going to be true in these in headless decoupled architectures. So there’s that there’s going to be the management of the content itself. I think, in those systems, you’re going to be gone reading I think you could have I don’t know exactly, but it’s projects like this. You might take stuff that was previously done in social media channels and email and on the website and in the apps, and maybe not, it doesn’t all go in one place. But a lot of that stuff will be kind of reorganized by purpose and intent in a management system. So just the way you think about a content management system and how you as an administrator, that system work that’s going to change. And then the front end, I mean, we’ve been dealing with this for a few years now everything kind of started with the responsive web and, you know, just stylesheets that could help you adapt to different things. But now that’s just gone on steroids. With the need the need as a as the output of that content management system. Now, instead of just templatized, web pages, or docks or something, you’re going to be pushing out API’s that a web developer or an app developer, a kiosk person, you know, print publisher might be drawing on. So I think there’s challenges across the spectrum. But again, like I said, there’s standards and knowledge and lore in place, like our odds of being able to tackle this,

Noz Urbina  

you know, I think, I think so I think there’s a lot there. And I’m gonna, I’m gonna keep track of your of your acronym dropping, so we make sure I know, we have our audience out there. That so we’re gonna go back in order. So when we’re talking about understanding and having things to look at, you mentioned, DITA and RDF. These are two standards for how you can structure your content, they have different focuses. DITA is more about how you actually author it in like reusable modules, how you mark it up, etc. And RDF is more about how you make relationships between those modules. So how to use so if I’m making a product and a product has features, product has add ons on a product has related products, it might have parts and spares, or it might have you know, markets or particular applications, all these relationships between different ideas. That’s more on the on the RDF side. But I the reason i Be very careful, very careful about dropping these acronyms is because when you google this stuff, it’s incomprehensible to the to the average content professional, and definitely to the average team leader or middle or definitely senior executive. It’s just gobbledygook. So although there’s the knowledge out there, it’s very frustrating. Well, it’s good. If you’re a consultant like us, I’m kind of putting myself in our audience’s shoes and going, it’s very frustrating, because you it’s very hard to look this stuff up. It exists. But the resources are not designed for the people who have to get it done. They’re designed for the the technical people who will help them get it done. The same thing I I think applies for headless, if you’re looking for headless CMS is it’s a very developer oriented product set right now. So my big issue with this with the headless tools that I’m seeing, and they’re I think they’re pretty open about this, and they know that they’re kind of made for the IT departments right now is that they’re just not nice or clear or easy to adopt for the actual users. So the business end of the project. So when you’re going to one of these projects, you kind of have to if you’ve coming from any of the simpler CMS is that we’re you just create pages and you create pages and folders and those folders are the navigation on the website, then, or you know, your tag thing. And that goes to a certain spot on the website. If you’re from that world headless. Can. There’s a lot of freefall, you know, there’s a feeling of WoW, where’d all my stuff go? And how am I supposed to like, what am I looking at here? What I see in a headless CMS, I’ll give it a give a concrete example. So I’m working with modeling content. And we were doing this at the university. And we were teaching the kids how to model recipes, which is a classic, some people who are old content modeling people in the audience will be groaning. It’s a classic thing to model modeling, because you’ve got kind of the description part of the recipe, you know, a little bit of maybe some kind of written paragraph stuff, maybe could be quite long form. Like if people started waffling on about this was my grandmother’s recipe, and she gave it to me on our deathbed, like, and then you get into the actual ingredients part, which is very structured, because you have the each ingredient and it has its amount and it has its how do you separate like a handful of pinch of salt versus eight grams of pepper, etc. The all the different ways you might represent the ingredients, but then that’s a very structured reference table. And then you have the instructional part, like how do you write good instructions and make sure they’re clear and unambiguous? And how do you want to create add media or notes to each instruction? So it’s actually a very interesting example. And what we found was we’re trying to do this on a headless CMS, and there was not a very good vocabulary of how we could do content types. Everything in this CMS was a content type everything. And so we you kind of end up ended up authoring in a sort of a I’m snail shell, where you’re going, you know, I want to create a recipe, a recipe has an ingredient, an ingredient has its quantity, the ingredient itself like pepper, salt, onions, and then the, quote, quantity of what so if we have three, three, what, three kilos, three grams, three, pinches, etc. And what we found was that every, every single recipe, all of the ingredients in the recipe became their own objects, their own content types, not just pepper, but one pinch of pepper. And then if I have another recipe, I would have two pinches of pepper. And then if another recipe is half a pinch up, so I ended up with like, all these 1000s of objects that don’t really make any sense just kind of cluttering up my CMS. And it’s very weird, it’s a very weird way of offering. Because then when I say if I want to reuse something, I have this enormous list of every possible permutation of of how much pepper and it’s saying, Do you want to point to one of those, I’m like, I don’t know, if this list of 10,000 objects to is somebody else already said, one pinch of pepper. And so it’s, you know, the whole idea of these systems is reuse. But actually, some of these new systems which are being coming out here, yeah, they can do reuse. And, you know, we’re we are implementing them, and we do consult on them. But you have to kind of know how to design these systems in a way that’s going to be effective, so that the reuse actually happens that, for me, that’s one of our big challenges is we’re buying the idea of reuse and single source and omni channel. But making that happen in these systems is not trivial, you have to have a very well thought out model to do that, to make your system navigable and viewable.

Larry Swanson  

Your money, I’m still sorting out in my head, the relationship between content modeling and information architecture. But a lot of what you’re talking about is like a huge, obviously interplay and overlap between those. Because that notion of like, an in right in between the two of those that would put a metadata strategy that like, what you’re talking about is like, what do you have, you have these things are some of these things are written in terms of like, you can have a metadata strategy, both about how you identify that thing in your own skin taxonomical way of talking about the things you have, and then the the kind of property JSON metadata about like, Oh, and this is a thing for sweetness or savoriness, or, you know, ethnic recipes or various kinds or whatever. So, yeah, there’s, yeah, I like, again, I like rods, because the the I think metadata strategy is much more sophisticated and developed these days, I think largely, not entirely, but I think practices like knowledge, graph, ontologies, and things like that. help inform that. But we’ve all been doing, again, I think back to my information architecture practice, 20 years ago, you do the exact same thing, just modeling, you know, figure out what the entities are, what are their attributes? How do you represent them technically? And then build it? Let’s

Noz Urbina  

be I know, I’ve always I’ve, I talk a lot about models versus domain models versus ontology versus knowledge, graphs, I leave Information Architecture out most of the time, because it’s too it’s almost too confusing. And I, you know, we probably should talk about that, what is the difference between taxonomy, and especially navigation taxonomies, like, categorizing your stuff, for how you’re going to sort it? And how you’re going to display it so people can navigate it. Versus content modeling? Like, what is the difference between content modeling and information architecture? It’s one of the it’s one of the few kind of disambiguation, which I’ve kind of avoided, I avoid talking about not not intentionally. So this is interesting. So feel free to how do how do you differentiate them?

Larry Swanson  

Well, I make one, so I made an attempt a while back, because I identify, like, I really identify now as a service designer with an omnichannel focus, but But what I call myself among content, friends, is a content architect. And what I did is I mapped all these things we do as content people, to Jesse James Garrett, you know, that’ll map of the levels of UX, strategy, scope, structure, skeleton surface, and they I mapped, and this is kind of ham handed and sort of, like, you know, click Beatty, I guess, but like, but basically, I imagine, you know, content strategy to the strategy layer those line up perfectly. I’m that modeling to the scope layer, where you’re going to figure out what do you got, what do you you know, your domain modeling, content modeling all that stuff that happens there. I’m an information architect that structure level where you’re sorting out the Lego box and how you label it and in which colors go with which and all that all that structure. So I I put information architecture at the structure layer, I kind of it gets a little fuzzier at the top like the difference between content designing UX writing, I don’t know that the profession is has sorted that out. And the roles are often conflated. But those top two layers of like, I think the second layer is clearly about information design, as opposed to presentation level design. That’s sort of the distinction between the selection and the surface. And whereas this and so I kind of map content design to the skeleton, like the scaffolding that you hang your content on in the in the interface, and then I line up UX writing to that. So that’s that’s one. So

Noz Urbina  

But where’s the information architecture there? Because I, I use content modeling to refer to both to what you set for both for information architecture and content modeling, which is, what are the things that we’re going to be talking about? Which, you know, it starts with generally understanding the subject domain, you can start with a domain model. And going okay, what are we talking about? We’re talking about why wines have regions and grapes and particular brands and wineries. So those are the things we’re going to talk about, what are we gonna say about them? Well, wines have years and Ryan’s have particular brand names and wines have awards that have original recognition, certificates, and blah, blah, blah, blah, blah. So that detail of what are we going to say about this stuff? How do we want to say it, and also, when you put a year do you type, you know, you put in a number. And if you put in a number in words, like if you write in 1995, it’s good, the system is going to reject, it can only it will only take a four, four digit year, that specification of how we want to receive the content. I refer to all of that as content modeling, information architecture, in the parlance that I’m used to using it has to do a lot with navigation architecture, how am I going to categorize this? How are like, when we come to whatever it is the site, the watch the chat bot, what is the architecture of how you can move and maybe chatbots are the same Yeah, chat, conversation design, the organization of these things that we’re writing about, because that could be very different, depending on what touch point we’re talking about. So in a headless system, we may present and organize these things under a very different navigation and categorization system for one market or one application or one audience, but at the same information, or some slice of the same information. But then another touch point might have a completely different navigation architecture and categorization system. So I refer to those as the information architecture. Content modeling, for me is the back end stuff that says this is how we do all of our content is for example, in a headless CMS. And then on the different heads, you might have different information architectures. And this is this is the interesting thing like we work together. And when you get into the nitty gritty, we don’t even know no one has alignment on all these definitions. I’ve never met two people who are like, Yeah, okay, we see all the things the same way.

Larry Swanson  

This is also where we have job security because nobody agrees on this terminology, and half of the job of the consultant is getting people to agree on terminology. So we got to do that. Got to eat our own dog food, as they say. But I think but I think we’re aligned on like, you know, the general idea of how we think about it, just we just have to tidy up how we label things. You know, one thing I’ve seen a number of folks have attempted to simplify that that five layer model and I’ve seen other like enterprise architecture models that people have tried to tidy things up. And a couple of them lead to my death done did it for inter books or what wasn’t best done, Beth done wrote a book called cultivating content design. I can’t remember exactly how she sorted out but it basically it was three layers of like strategy, structure and and surface something I think that was how she simplified it. But more more more germane maybe to our work is the the guys you know, factor firm, Bram and Gary Carl and Graham muslin. Gary Carlsen, in Seattle, they have this model that they show for their like an information architecture consultancy, but I can’t remember exactly how they identify it. But they have this three level enterprise architecture model. It’s like systems, information, application stuff, you know, like kind of bottom to top back end, the front end. And I think that aligns really well with like my service design mentality about front of stage backstage services underlying it. It also kind of gives you buckets where we could I think we could put everything we’re talking about information architecture and content modeling in that middle layer with some like the navigation stuff, migrating into that top layer of the application level. And but also a lot of the Taxonomy and Ontology stuff going down to the systems layer. But it’s that I think that’s the that’s the point of their model is that like, that’s the thing ties it all together is that information layer in the middle. And that’s where all of this stuff rests, I think that we’re talking about

Noz Urbina  

yeah, I’ve got a quote one of our one of our clients at IKEA, so there’s, and she was also on our podcast, herself recently, which was Timmy A trainee from IKEA. What’s her? She’s

Larry Swanson  

a caller. What’s your lesson? I forget it’s a hydrogen. Yeah, there, she’s

Noz Urbina  

gonna kill me. Yeah. Can you still balca Alcala, so Timmy came out with a I think it’s her term, but she’s one of her first who shared it with me which I love, which is the content stack. In the idea that, you know, your whole content strategy comes in these layers of how you organize it structure it, you know, have you have the actual layers of standards, I used to talk about, like the standards pyramid, where we had like enterprise wide standards. But there’s certain conventions and rules that go across the whole enterprise of how you do content, not just how you present the logo or what colors, but how you like, for example, certain language standards, you have certain tone of voice standards, which might be company wide, you might have certain terms you use and don’t use a company wide, others will be functional. So you know, you have a layer so that you have a kind of departmental layer saying okay, well, within support, we can do, we can say this, but we wouldn’t say it that way in marketing. And then and so on, and so on until you get down to stuff, what you define for an individual project, which is, you know, I’m going to create new content for this purpose, which no one has ever talked about before in the company. And so you it’s ultra unique, but it’s all based on this whole pyramid of standards. And so we’re talking about the content models, and taxonomies, and all that stuff, living somewhere in that pyramid. So coming back to recommendations, like if you’re going to a headless CMS, you need to be thinking about this pyramid of however you label the layers, start to think in this layer separation of concerns that we have to be implementing. Whenever we’re thinking about how we’re going to tag things, what is the relationship between those things that we’re tagging? So I gave the example of wines, you know, a great but a year and a vintage? Like, where does the price live? Maybe the price lives in a whole different system than the content. So how are we going to manage those relationships between the price data and the actual stuff we write about the products. And this is true for all industries, like it’s, it’s this separation of concerns, they may they may not be in the same system. And I think headless CMS is doing a start up. I’ll call it a mistake. Headless CMS is kind of making the same mistake as AI companies are making that companies have been making my whole career, which is that shows up goes I am the new single source of truth. I am the way I am the one path. Now you can now you can put everything in me. And all your problems will go away. We’ve heard story every decade, you know, since I got into this game, and and surely since you did, too. You’re going to have your headless CMS, but you’re still going to have other content sources. There’s other things in your strategy, like like like, like pricing data, like support knowledge bases, like whatever, etc. Unstructured stuff like, Oh, God, save you, if you’re using SharePoint, there’s gonna be libraries and stuff all over the place. And you have to have this layer cake of standard saying, Okay, we’re going to, we’re going to have a categorization system, which we’re going to agree across these departments. And occasionally, maybe you might get to a point where you can say, okay, enterprise wide, then content models, how can how, how realistic is it for me to do an enterprise content model? For example, Bruce, Chris Saunders, who is part of our panel of experts that we’re doing as part of the same series on content modeling, talks about introducing the the master content model. So if you look for cruce, see Are you see E. Saunders introducing the master content model, then he’s talking about this idea of we have a master content model, but then we have more specialized derivations of that for different purposes in the business. So this kind of layer cake thinking, I think, is the fundamental takeaway from this, this bit of the discussion that you have to you have to not swallow the pill of oh, I’m gonna have a headless CMS. Now I have one CMS for all my things and everything else can just, I’m done. And then we just raised the for very important that this is coming up again, and again, the design of the information within a head is a separate concern, but related to the design of the information in the head less back end. Like if I’m going to display it in these ways on this touchpoint that information has to be available in the back end. So although we are separating concerns, and we are going to this kind of neutral omni channel resource, you have to know what you’re trying to do on with the experience you’re trying to deliver and the delivery layer, so and you have to so if you’re not in control of all that within your company, you have to be having an active dialogue. Yeah, the next the next big one I want to come up with the next big challenge which I have an issue with, which is Is this situation where the IT department or some enterprise central function? Is that? Okay, we’re all standardizing on X headless CMS. What do you do if you already have a really good system in place, you have structure, you’ve done a lot of work. It’s very well built the argument like what do you do if you want to maintain or your your you own the pricing system, for example, you own the PIM system, you have a structured system. And now you have this headless CMS movement in the organization. Now navigating that discussion,

Larry Swanson  

yeah, that’s a that’s an interesting one that I, I kind of dealt with that a little bit, I was on one of my last things. One of my gigs last year was Expedia. And one of the things that there was a worked on enterprise wide CMS planning group, which was they’ve got some content there, let me tell you, like, literally 10s of petabytes of data about every airplane seat in the world. But, but I think that kind of gets back to the the foundational meaning of that separation of concerns, like there, there was, there was it was early days, when I still early days, when I looked at I have no idea where that project went. But but one of the things we talked about was one CMS, you know, which might make sense. Maybe not, though, you know, that like one of the things that the separation of concerns may be that like, this kind of CMS is really good for this family of stuff. And we don’t want to just do our Taxonomy and Ontology stuff in the CMS, we want to sort that out is like something we do integrate a pool party or something like that

Noz Urbina  

external taxonomy dedicated system that is a service to the other systems.

Larry Swanson  

Exactly. And I think in an omni channel environment, I bet my prediction is we’re going to see a lot more of that. Because all that stuff, you were talking about the very start the need to align different siloed organizations, you know, entities in the organization, on terminology. Boy could like doing that, anyway, but you’ll at least have a chance, you know, if you’re doing it if you have like, a person who’s doing their best to align people across the Oregon on stuff, and it can just be as simple as saying, like, Hey, I heard these people in the port talk in this way about this thing and or in sales or talking about this way. And in customer support they’re talking about it helped me reconcile this. And then and again, back, I won’t, I won’t repeat the acronyms. But there are technologies and standards that help you organize and, and align similar terms that mean the same thing. Kind of back to that, that Google notion of things not strange, you’re talking about the same thing. But you just have given a different labels in different contexts. It’s kind of like what anthropologists or sociologists call a boundary object, you have a rock sitting in the middle of a table, you know, in, in for the warrior culture, that’s like, yeah, that’s how you kill people. And for the farming feature, like no, that’s how you plow the ground, you know, so that’s so and so they call it a killing machine over there in a hole over there. But it’s still the same thing. And so I think we have an enterprise content management that you’re going to have things kind of hovering in the organization that people call different things. And there are both procedural like, you can get people in a workshop and help them, you know, agree on terminology. And if you can’t win, there may be good reasons to call things different things in different contexts, then you have some kind of technical mechanism to align, say, like, oh, when they say this, it’s the same thing as this. They’re saying over here? Yeah,

Noz Urbina  

I agree completely. I think the taxonomy being a separate concern, and even a separate system is really, really, really important and powerful. So we’ll come back to the AI thing. If your AI is being trained on your content within your enterprise, if you have a universal taxonomy, that at least it can get some tagging and get some understanding of relationships, that is concrete. So it’s not trying to just interpret this by reading your your words, but you say no, this is a this is a, this is a product, it’s a it’s a drug, it’s for this market, it’s for you know, it’s not for pregnant women, like if you have certain clear labels on your stuff, which is absolutely reliable. And that is present across multiple silos connecting multiple silos than an AI or a human being, he’s gonna have a much better chance of understanding that and connecting the breadcrumbs. I think that that’s another kind of Getu that we need to tell our our listeners about taxonomy. You know, it’s not all of these headless CMS is do taxonomy at all. Some of them have plugins that do it, some of them don’t. So how are you going to make, you know, how are you going to categorize things? How are you going to filter things? I see. We’re gonna talk a little bit, I think for the last little bit, because we’re covered probably coming out on time soon, I see that we’re gonna see a future of tools there that are layers on top of your headless CMS. What I’m seeing right now with these headless CMS is you can go in, you can do your content model, you can author your content, but that’s not really what they’re good at. And I see a future where we have heads within the business that are for internal use that make the headless CMS usable, and that might be very good. different depending on what your needs are with that with that headless CMS, so you can have an authoring layer. And we’ve actually done this in some projects already, where you have like content curate, you have content creation, where you do go into headless CMS, and you create your objects and you relate them etc. And then we have kind of a lower level of content creation, where we’ve created a website where you can come in and put together Lego pieces and customize them and you know, rename the groups and reorganize them. But you never have to log into the headless CMS, you are creating objects in the CMS. So you’re creating new collections, you’re creating new, new organizational structures, but you’re not creating, you don’t create the raw objects, even though you can put on like notes on things. But your job is not to be a content creator. So you you but you are, as far as you’re concerned, you are creating content because you’re taking stuff for a particular purpose. Putting together the Lego bits that you want, giving it new names, giving it a new presentation, navigation, and then delivering it out to the market. So that but on steroids, so that you can have completely different authoring tools, ways of authoring and in different interfaces for different parts of the business. But all facing back to this common back end, I really see that as a as a potential thing.

Larry Swanson  

Yeah, I was I was talking to a guy a couple of weeks ago, CMS vendor here in the Netherlands, about consumptive AI is generative AI, and why ideas like consumptive AI. And I don’t know, I don’t know if that’s a real term. But that’s what I was calling it. That notion of curation, like using AI and machine learning technologies to help with curation and reusing existence that totally is tied into this whole structured content, heedlessly metadata, strategy, stuff, you know, being able to do more with it. And that’s it, for lack of a better word, I came up with a consumptive AI that, like you’re just consuming stuff you’ve already created. I think all the fuss now is about creating content with AI. But I, I think there may be, as you just said, I think there may be as many opportunities to reuse old stuff.

Noz Urbina  

Yeah, no, I think that’s one of the more mature applications of AI is to categorize, read, summarize, auto tag, you know, to to read everything and say, Where are you talking about? What, what are the potential terms that you should be having in your, in your dictionary in your taxonomy of terms? And then, you know, I think summarization is one of the most interesting things when you think about in your content model. So we have, I haven’t seen this yet, I’ve told this for a few to a few vendors, and I should probably get paid for it, because they were very excited about the idea. But when you combine generative AI and consumptive AI, it will and in terms of a content model, so I have got a content model that says I’ve got products, and then I’ve got podcasts, so and then a podcast or a product might they both have social posts about them. And they might have, they might have email newsletter entries about them, they might have whatever other things in my content model that are about the about products and podcasts. So can I specify that if I write the long description of a product, or the long description of a podcast, I want you to generate me the one line summary that’s supposed to appear on the widget in my on my website, like on the on the landing page, where I’ve got all the episodes of my podcast, or the catalog page that has all my products, there’s gonna be like a one sentence description, or two sentence description for like a LinkedIn post or like a one paragraph description for a LinkedIn post. So you, can you I think there’s a huge opportunity, headless in AI to say, over here, I’ve got one incarnation of the description. And elsewhere in my content model, it may be in this content type, and maybe in other content types, I have other incarnations. So a way to specify to the headless CMS Look, here’s the long one, give me the short one. Or if I write a short one, you know, exploded out by consuming the other things I have in my business and writing me a long one. I think that’s a huge opportunity, which we’re not seeing in any implemented animus systems yet. We’re bringing together

Larry Swanson  

that’s brilliant. And I think that my very first publishing lesson back in the Flatiron Building in New York City many years ago, my boss wrote this really long memo to me She used that old now everybody’s heard this. I apologize for writing such a long letter. I didn’t have time to write a short one

Noz Urbina  

quote from

Larry Swanson  

exactly I think, I think it’s anyhow I forget the exact source but basically that idea that like somebody likes it Yeah, but you’ve done the hard work of getting everything there. And these tools like AI are so good at summarizing it like, what would take you three hours they can do in 30 seconds. You’re like, that’s a win. I’ll take it. So yeah, it’s concept and workflow. Holy cow. Yeah.

Noz Urbina  

30 seconds plus two minutes or review and wordsmith it. No, I

Larry Swanson  

think that’s an important point that what they call in some junctures the human in the loop, you know, to have like human oversight, especially like, well, I know you’re doing stuff in pharma these days. And, boy, you’re not going to be shipping like raw AI generated content in that world, I’m sure. So but in whatever it is, but you can also picture like who you were talking before about, like having your own enterprise corpus of your own stuff, that learning loop I get really faster and tighter when you’re working with your own content. And the the learning can be ensconced in a systems that much quicker, I think,

Noz Urbina  

yeah, well, that’s what our big enterprise clients, they’re looking at training their own large language models, they’re not going to just use chat, GBT, like we do, they’re going to train their own internal model on their own content. And so the more you can give it structure as a as a reliable back end, and you can point you know, different things. And say, specify reliable structures, the more effective those internal AIS will become, AI is going to become this to to quote, well, I don’t remember who it is. But someone said that AI will be like spellcheck. People talk about you know, how am I going to use AI to do my x ai will be integrated into all of our tools. So that when whenever we hit File, New, or we or new X, there’s going to be an AI contributor, I just saw that in one of those types of CMS is, when you’re doing your content model, you can say, AI generated content type. And you can say, I’m creating a podcast overview object, and it will come up when it came up, or what to come up with, I came up with like, episode number guests headed titled short description, you know, the basics, it’s there for you, you have to do things like say, No, I want my guests to be a pointer to my person object. So that the guest is actually a human being and that that human being then has a first name, last name, headshot, etc, you can do two AI models, and then just explain to the AI look, this one should actually this guests thing for a podcast should actually point to this person object over here, you have to do that manually. But, you know, every day, we’re seeing a new point in the workflow where AI is being embedded. So rather than thinking of AI being some this thing that I use, and then I go back to my regular tools, it’s moving so fast that all of these tools are going to be embedding these things. And then can you in an enterprise context point, these different tools to your one unified like body of knowledge? Or, again, we’re not going to have that either, because we will never be one unified anything in a large enterprise. Can you federate it so that at least there’s like, a, a butler AI? Who knows how to talk to all the others, and bring you back the right answer for what you’re doing?

Larry Swanson  

Yeah, as you talk about that, I’m thinking there’s many practitioners who are listening, there’s a lot in the new kinds of professional development, as we go forward to understand these technologies that will use them, like I was, I had coffee with a friend in Amsterdam last week. And he’s a, he was a poet in college. And then he became a Content Designer, he just learned Python, so he could do AI better. So that’s like, maybe the extreme, you know, from poetry to python programming in a couple of years. That might be extreme. But I think the point there that like, you know, being conversant with all these technologies, and and just so you can identify those opportunities, especially if you’re doing any kind of Enterprise UX, or enterprise content stuff to like, all the stuff you just said is like gold, and I think will save millions of dollars. I don’t know. It’s not hard to picture good benefits.

Noz Urbina  

Yeah, that’s what that’s what I’m hoping for. All right. So I want to wrap up with a cute couple. You know, Larry, you and me, we’re available on LinkedIn, of course. And this will be going up on the omnichannel. X website, as well as one of our podcast episodes, do check it out if you’re listening to this, the rest of the content modeling session series. So we have kind of these these overarching, let’s say virtual events that collect podcasts and webinars, we have the content modeling experts panel, which is part of the same group, where we have crusaders who have already mentioned myself, Megan Casey, Carrie hain, Jeff Eaton. So they’re kind of a bit of a who’s who and the content modeling world. I wanted to recommend a couple of articles if you’re looking to learn more about content modeling, that you know, this is this is all stuff written prior to the headless movement that are term headless, is only about five years old. These are articles that go back, you know, double that and, and some of our work goes back before we were even calling it content modeling, we were kind of conflating it with content strategy. So when I was doing this with an Rocklea, and you know, just after the.com, boom 2000 to 2003, we didn’t use the term content modeling, we just said it was part of your content strategy was getting down into all these details. So the terminology is evolving, but these articles you can check out that will prepare your for your headless migration. So from Jeff Eaton, you can Google up battle for the Body field. That’s for me, that was a seminal article, that was a kind of a one of those kind of penny drop moments. In terms of bringing this discussion to the the world of the page oriented world, the people, which is not where I come from, I come from the structure oriented world. Kerry’s got a content modeling guide. So if you look for Carrie Hain content modeling guide, you’ll find her guide. Megan, Casey’s got a workshop up on our website on the omnichannel. X website, the which is part of the same series called the content modeling the translation layer between teams that’s free. If you register to watch that you can sit for a whole workshop with Meghan, we’ve all already mentioned cruces article introducing the master content model. And, Larry, you mentioned when we were talking earlier up Marshall Marshall, Marcelo Lewin.

Larry Swanson  

Yeah, he runs it. Marcelo runs a site called headless creator. And he does just this whole series of really good content on content modeling. It’s more like more of the technical implementation level, like actually building the schemas for a CMS. But it’s, but there’s much more than that in there as well. But it’s just I think it’s just headless creator.com. And Marcel is a really gifted interviewer, good guy, and just tons of good content on that website as well.

Noz Urbina  

And you were just on this podcast. Yeah. Last Friday. Yeah, we

Larry Swanson  

could go. I was, we had a chat about authoring experiences in the a couple of architectures, which was

Noz Urbina  

exactly, exactly that’s, I’m super passionate about that. And then what else the other one I want to drop is, of course, Rachel Levenger. I think she’s she’s credited with the first kind of article that went viral on this topic. In 2012. She wrote content modeling a master skill on A List Apart. And that was a good article, I use that as a, I use one of her screen tour diagrams as a teaching aid. And I was always reading her because she, she wrote in one of her content types was was, I think, a banned page or something like that. And I’ve always always circled page, and I’m not talking about pages. But it’s an excellent article. It’s a really, really great. Her along with Carrie and Jeff, they’re really great at kind of making these these things accessible for everybody and understandable. So we’ll put all those links in the session notes. Otherwise, Larry, thank you. Welcome. Thank you for joining us on the podcast. Thank you for joining us at Urbina Consulting. And thank you for our discussion today.

Larry Swanson  

My pleasure now is always fun to talk to you. Awesome.

Noz Urbina  

Thanks, everybody. Talk to you soon.