2023 content trends: technology (headless CMS, composable, DXC, and AI)

Note: I’ve interviewed a number of folks in the industry, so citations and quotations are from interviews I’ve conducted, unless otherwise cited.

In the first of two articles, we talked about how the content profession is changing, and now we need to take a look at how content technology is changing.

You may not geek out on this like I do, so I’ll do my best to not go too deep. Even if you don’t consider yourself a technical person, you’ll want to understand where the technology is going. It can help you prepare for the future, and it might even be a way to grow your career.

While I’m going to talk about tools, I’m attempting to do so in a setting of how these tools and technologies solve real-world problems. It’s not about the “shiny new thing.”

Headless CMS

In a traditional content management system (CMS), content is tightly coupled to how it’s presented on a webpage. That means the content isn’t very portable or future-proof. Changes are expensive.

The answer to this has been what is called a “headless CMS,” which separates the storage of content from where it’s presented.?

A true headless CMS uses something called application programming interfaces (APIs) to create and publish content. APIs are how computer programs and services connect to each other, similar to the pipes that connect your home to the water system’s pipes.?

Headless CMSes like Contentful , Kontent.ai , Sanity , Prismic , and Contentstack have been around for several years, but there are more and more headless CMS products hitting the market (and legacy CMSes scrambling to overhaul to headless).?

There’s also a growing buzz about “headless” in the industry.?

It’s probably safe to say that if you’re not using a headless CMS yet, you will be.

“As more organizations move to a headless CMS, everyone has to move with it, whether they are ready or not,” said Carrie Hane , principal evangelist at Sanity.io and co-author of Designing Connected Content.

With more and more companies trying to publish personalized content to different user experiences and various media, a headless CMS enables these new capabilities, as well as operational efficiencies.

“Customers now expect content on a variety of platforms, in many different formats, and most importantly, they expect personalized content that is relevant to them,” said Yanko Ivanov , enterprise solution architect at Enterprise Knowledge, LLC . “As such, now it is more important than ever to be able to create the content once and then reuse it, tailor it, and deliver it in any format or channel that can reach a customer. A headless CMS is a key component that empowers this ability.”

Aaron Bradley , structured content architect at TELUS, echoed a similar sentiment about what a headless CMS enables.

“The rapid ascendency of headless is directly related to the things it facilitates: automation, personalized content generation, a flexible and scalable architecture and, especially, the ability to publish the same content to different endpoints,” Bradley said.?

In fact, the headless CMS is but one piece of an emerging view of content tooling that speaks to information technology (IT) concerns about how the tools they manage work together to serve internal stakeholders and their requirements.

The days of monolithic programs that do all the things are increasingly behind us, and they are being replaced by an approach that many are calling “composable.”

Composable, DXC, and the MACH Alliance

In modern software programs, code is often broken up into discrete pieces of functionality, instead of being one big mass of intermingled code. This approach called microservices architecture makes updating, testing, and replacing parts of the program cheaper, easier, and faster.

IT departments are seeing a similar approach play out when it comes to content tooling.

In the past, these departments may have used a single program—an experience manager—or Digital Experience Platform (DXP) to attempt to do all things content. However, they found that DXPs don’t meet all needs, they are expensive and slow to customize, and they are painful to migrate off of.

Now, IT departments are starting to use best-in-class vendors who each provide pieces of the content tooling ecosystem. In this approach, organizations choose the best CMS to suit their needs, as well as what they need from the best search provider, best taxonomy manager, best A/B testing provider, best digital asset manager, best product information management system, best front-end framework, and so on.

These vendors’ offerings connect to each other via APIs. In a real sense, it’s a microservices approach applied more at a macro level.

This approach is frequently referred to as “composable,” as organizations compose the tooling that meets their needs. Often framed as a marketing technology or customer experience concern, composable is something that ties directly into enterprise content strategy and content engineering.

“I’m excited about composable because it’s a far more flexible, extensible, and scalable approach to content engineering than what we’ve seen in the past,” Bradley said.

Composable is taking root in the industry. You’ll find a whole host of vendors trying to capitalize on the buzzword. But more than that, there’s a new not-for-profit called MACH Alliance that advocates for solutions that are “Microservices-based, API-first, Cloud-native SaaS and Headless.”

Additionally, in 2022, Gartner recognized a new market segment based on composable.

“With its rising importance, Gartner has recently coined the name … Digital Experience Composition,” explained Sana Remekie , CEO of Conscia and a MACH Alliance ambassador.

While it’s easy to focus on the technology and specific tools, the composable Digital Experience Composition (DXC) approach has come about because it’s solving problems for businesses that a headless CMS alone won’t solve.

“The appetite to turn to best-of-breed solutions has risen with the need to build both unique and consistent experiences that are not just limited to the web, but span the entire customer journey,” Remekie said.

“Insofar as it imagines a content ecosystem as a number of API-connected parts it’s a natural extension of the headless approach,” Bradley said.

“True composable is the promise of connecting the different technologies in the stack without needing custom glue code,” said Lars Birkholm Petersen , co-founder of Uniform . “Thus the more composable your stack is, the easier it is to add new technology, orchestrate data between technologies, or replace one technology with another.”

The composable approach addresses another issue that can arise from stretching the headless CMS’s capabilities and creating complex implementations.

Preston So is senior director of product strategy at Oracle. In a session at OmnichannelX in June 2022, So talked about the low-code movement and questioned why headless CMS are still high-code.

“Headless CMSes remain for developers and no one else,” So said, while talking about how it’s hard to optimize content tooling for both developers and content creators.

There’s a sense that “composable” brings tooling that fills in some gaps that CMSes don’t address.

“I think the space of Digital Experience Composition (DXC) is really interesting for content practitioners because it finally gives them back the level of control and flexibility that they need to be productive and reach their business goals,” Richard Bausek , director of product management at Uniform.

Petersen voiced a similar thought while pointing out that the coordination of these decoupled composable services is an area ripe for growth.

“With everything being decoupled, the need to orchestrate and create compositions from a mix of different systems is where most innovation is happening,” Petersen said. “All the innovation is happening to make the experience equally good for developers and business users and make the digital experiences fast and accessible to visitors from all over the world.”

Orchestration

CMS vendor Contentful talks about its “composable content platform” and another CMS, Sanity.io, talks about “the composable content cloud.”

In this emerging space, terminology is a bit fuzzy.?

Some folks may say that “composable” and “orchestration” are both talking about the same thing. They tend to be used interchangeably.?

However, I tend to think of “composable” being about the technology or tooling and “orchestration” as being about how content from a headless CMS is assembled.

It’s entirely possible to build content models in a CMS that allow for content orchestration, but it’s really easy to revert to coupling content with presentation logic in this approach, which adapts an existing tool for something outside its purpose.?

The content world needs something purpose-built for orchestrating structured content that can be used across channels in different experiences.

It’s been really exciting in the last couple of years to see a new crop of solutions—Uniform.dev, Conscia.ai, and Stackbit (acquired by Netlify) —taking on that orchestration gap. Each is taking a slightly different approach to the orchestration problem space.

“DXC solves the hard challenges around how content is put together by connecting content to design components and providing the contextual preview and rendering capabilities that business users need,” said Bausek.

“The tools outside of the CMS aim to unify the content silos as well as centralize the orchestration of experiences to create a cohesive and continuous user experience,” Remekie said.

That resonates with expert practitioners who see the need for orchestration.

“I am excited to see vendors starting to focus on the aspect of content delivery that allows marketing teams, product teams, etc, to take advantage of the existing content elements and piece them together in a way that serves their specific purpose,” said Ivanov.

With orchestration tools connecting content silos, organizations might even see new capabilities, like the ability to swap out an old CMS with a new CMS progressively. If the website’s user experience is driven by a composable DXP that orchestrates content from multiple CMSes, you can potentially iterate your way through the CMS migration instead of having a single high-stakes “big bang” migration.

Knowledge graphs

While a headless CMS can be extremely powerful on its own, and even more powerful when coupled with pieces of a composable DXP, it still can’t deliver the kinds of automation needed for large-scale enterprise content or even smaller-scale cutting-edge content-driven experiences.

“A headless CMS excels at structuring content, but less so at describing how different types of content are related to one another and to the real-world entities to which that content refers. So you know your content’s shape, but not its meaning,” Bradley explained. “Knowledge graphs bridge this gap because they provide a means of precisely describing how different elements of a content ecosystem are related to one another: content in your CMS, customer data, product information, analytics.”

In a CMS, content is essentially stored in a traditional relational database. Pieces of content can be linked together—usually manually—to create unspecified relationships. A knowledge graph specifies the relationships and sets the stage for rich machine learning.

Let me lean into my grammar background with a metaphor.?

In a CMS, content entries are objects, also known as nouns. A person must look at all the nouns in the system and figure out how one noun relates to another. Imagine an object that is person and another that is vehicle. If the CMS links them, we can tell they are related, but how?

With a knowledge graph, content entries remain objects known as nouns. However, the knowledge graph includes verbs that can connect one noun to another into a simple sentence with a subject, verb, and direct object. In the knowledge graph, we would know that person drives vehicle.?

That’s a different meaning than if the verb were something different, like fixes or sells. The knowledge graph lets us define whatever relationship our person/vehicle content needs to communicate. This pairing of a subject, verb, and object is called a “triple” in the graph world.

In the CMS when relationships are nothing more than links, the relationships are often too rigid.

“Information wants to be connected to other information, regardless of any hierarchy, which is so constraining. Connections are multifaceted, and that’s what knowledge graphs help us see,” Hane said.

In fact, connections are fundamental to the idea of a knowledge graph.

Many people decry the “silos” that exist in organizations. Indeed, silos can be problematic as content duplication can proliferate, creating inefficiencies and inaccurate content. Knowledge graphs can help connect silos.

“Knowledge graphs are fascinating in the way they allow us to connect things together regardless where these things reside in an organization's ecosystem,” Ivanov said.

But as Bradley alluded to earlier, it’s more than connecting content from different silos. Knowledge graphs connect content in the CMS with data in other business systems.

“They allow us to seamlessly integrate structured data and unstructured content to produce powerful applications like recommendations, chatbots, question answering engines, and so on,” Ivanov said.

“Knowledge graphs are an important tool for organizing, connecting, and understanding vast amounts of information,” said Cruce Saunders , founder and principal at Simple [A] . “They can help content people better understand the relationships between different pieces of content, as well as the context in which the content exists.”

So what should content professionals be doing about knowledge graphs in 2023?

“At this point content people just need to know what knowledge graphs are, and how they are different from relational databases,” Saunders said.

Indeed, knowledge graphs are an emerging technology. Forward-thinking content professionals should be starting to learn about knowledge graphs (and by extension knowledge management). Beware, though, that a lot of knowledge graph information, including The Knowledge Graph Conference , is rather academic—mathematical, even—at this point.?

There’s not a lot of content-centric knowledge graph resources out there, but Bradley’s Graph Lounge website is one place to start, along with The Knowledge Graph Cookbook (free download).

Tooling wise, many CMS providers support GraphQL, a query language that provides some graph capabilities for non-graph content. Hygraph leans into Graph QL, and seems to be doing some interesting things.

Additionally, many of the “composable” tools can be combined to yield graph-like capabilities. Conscia.ai offers a Digital Experience Graph offering that is centered around the knowledge graph idea.

“Even if the tech stack employed in any given composable exercise isn’t strictly speaking that of knowledge graphs (though it may be), composable is an emerging approach to getting to graph-like capabilities,” Bradley said.

Generative artificial intelligence (AI)

If you’re like me, your newsfeed at the end of 2022 and into January 2023 has been full of chatter about ChatGPT, Midjourney, Stable Diffusion, and DALL-E.?

ChatGPT and DALL-E are probably two of the most well-known players in the generative AI field. Generative AI is artificial intelligence geared toward generating new content (including text, images, or other media).

This has many content folks wondering what these technologies may mean to their jobs.

I’ve purposely not engaged with these technologies myself, but I think that content professionals should be aware of them and thinking deeply about the implications of them.?

I’ve chosen not to engage with them myself for a few reasons:

• I have limited bandwidth. I don’t want to add another thing to focus on.
• I’m not usually an early adopter.
• I’m leery of free services—especially ones that want my input to help train their technology.
• I’m not sure we as a society have worked out the ethics and implications of these technologies. This is not to malign anyone involved in working on the technologies, but more along the lines of “Did we realize the impact that social media might have on our society? Have we learned any lessons from that?”

However, having said that, I expect generative AI to become more mainstream, more commercialized. So, with the obvious caveat that I’m not an expert in generative AI, here’s my thoughts:

• These technologies were developed by humans. Humans have a lot of conscious and unconscious biases. Those biases will surface in—and possibly—be magnified by generative AI.
• Generative AI could be extremely useful and valuable.
• Generative AI is not going away.
• Content professionals (including artists and musicians) should expect to learn how to harness and use these capabilities instead of living in fear that the technologies will replace them.
• Content professionals could be vital to helping think through ethics and implications, and maybe setting up standards and best practices along with the technology creators.
• There’s a world of opportunity for content-minded systems thinkers and service designers to imagine how we put generative AI to use.

Applications of generative AI that excite me:

• Could the image generators be trained to create localized screenshots for technical documentation, with realistic-but-fake user data displayed in the interface?
• Could generative AI be used as an accommodation for someone with learning differences (either in school or in a profession)? I have a hunch it could be useful for people on the spectrum, people with ADHD, people with dysgraphia, and so on.
• Might a tool like ChatGPT make a good “collaborator” for remote workers or small businesses? I’ve heard first-person anecdotes from folks who used ChatGPT to refine their thinking, identify gaps in their logic, and find creative metaphors.

Conclusion

The “secret sauce” that underpins all the capabilities that emerging content technology promises is content that is intentionally structured and semantically rich.

As more structure is added, content becomes more flexible, artificial intelligence and machine learning become easier, and the ability to scale content increases.

Companies will need professionals who know how to structure content according to its own meaning in ways that can support user and business goals at scale.

This will require content engineers (see 2023 content trends: roles and specializations), and a modernized approach to content.

Beyond content engineering, I think there will be a new job emerging that will help show organizations the art of the possible in stitching together DXC solutions to create powerful integrated omnichannel experiences.

It will be a position for a tech-savvy, content-minded person who can use service design principles and connect the dots for engineering teams, content teams, and UX teams.

Others see new roles coming too.

“Another role that will evolve is that of an ‘omnichannel experience architect/designer’ who would be responsible for using a tool like the DXC to orchestrate experiences on multiple channels,” Remekie said.

Cover image by storyset on Freepik