Structure and Semantics: The Pillars of Intelligent Customer Experience
How will customer experiences become intelligent, flexible, and capable of personalization? Content.
Content drives all digital customer experience (CX). Whether clunky and single channel, or fluid and omnichannel — none of us touch or interact with any organization or software without content as a mediator of the experience.
And yet, most organizations are trying to change customer experience without looking after the energy that powers all of it: the content itself.
- recommending the ‘next best’ content to customers, Netflix-style
- proactive pushing useful content, instead of requiring the customer to dig many levels to find what they need, or never encountering it at all.
- meeting a customer’s stated or implicit intent with highly useful responses, search results, emailed resources, or other satisfying resolutions
All of it depends on content. But not just any content.
Intelligence starts in the content’s DNA
The very roots of content, the dimensions that give it life, live in the invisible parts. Peering into the invisible nature of content gives us abilities to work with that content in entirely new ways.
Structure and semantics are like the DNA of content. Invisible to the naked eye, but essential to the intelligence of life. Content lives its life with the capabilities of its DNA, like all living organisms.
Content can grow, change, relate itself, and recombine in accordance with the capabilities encoded into its DNA. When systems have no access to the content’s DNA, nor instructions to act upon, content just sits where it gets placed the first time it is placed there.
I like to think of content structure and semantics as existing already, and that as content strategists and engineers, we reveal and manifest parts of a reality already hidden in the content. Regardless of how poetic our interrogations of content, we certainly give content more power by enriching it with semantic associations and expressing its structure in machinable ways.
Discovering, harnessing, and bringing the inherent nature of content into representation within technical systems, gives content extraordinary capabilities. We open up a universe of personalized, relevant customer experiences when we attend to the content’s DNA.
When structure and semantics work together within content sets, suddenly we have a whole new world.
As publishers of intelligent customer experiences, teams need the ability to author and manage multiple variations of content components.
Content, in our new world, doesn’t sit in a chunk in a place. It lives in structures that enable pieces and parts to be dynamically assembled as relevant experiences, based on the context of the customer’s touchpoints.
Publishing teams, whether in marketing, documentation, or support, need the ability to create a content component once, then reuse and deploy it many times, in many ways. This requires structured content with an intelligence shaped and connected by semantics.
And it starts with a model. Structuring content within a well-defined content model makes content scalable, reusable, adaptable, and measurable. We cannot create real-time, personalized conversations until we have structures, or handles, by which to work with that content.
In this article, we explore the role of content structure and semantics, governed by a content model, to reduce friction for both the content’s authors and consumers. We will also examine key distinctions between structure and semantics.
Structure shapes content with an object-oriented approach:
- Content is organized as reusable content objects, rather than hard-coded, unstructured blobs.
- Objects have containers, which can be manipulated, transformed, annotated, reused, and managed from a central location; they can be pointed to from anywhere.
Content semantics is the contextualization of content structures:
- Content semantics define the entities, associations, and relationships for a given piece of content within the metadata.
- Semantics is how machines can understand content connections and relationships.
We will also explore content and semantic models:
- The Content Model is an orchestration layer for structural schemas.
- The Semantic Model is an orchestration layer for taxonomies and vocabularies.
Why Content Must Become Intelligent
Content is what we acquire, manage, and leverage in order to engage and inform people. In order to effectively engage consumers, our content must be capable of movement and flow. Intelligent content is content that has been structured and contextualized so that it can:
- Flow effectively and efficiently between people, machines, and applications.
- Deliver individually optimized customer experiences at scale.
- Maximize the return-on-investment made in its acquisition and management.
Content intelligence is a moving target, as the content itself constantly evolves, driven by consumer demands and needs. The channels for delivery are seemingly endless. We are also challenged with creating intelligent content that is ready to serve the demands of existing and future interactive channels.
Although most of us are focused on technical content creation, we must acknowledge the need to seamlessly integrate our content with other divisions of the enterprise. Working towards leaner, smarter processes that take advantage of automation and customer experience (CX) platforms helps eliminate costly barriers to customer engagement.
Pre-sales, marketing, and post-sales content must all work together to optimize customer experiences and ensure both customer retention and revenue growth. We accomplish this by structuring and contextualizing our content so that it can flow freely throughout the enterprise.
At the heart of this process of making content intelligent, the Master Content Model® (MCM) ties everything together: it defines the relationships between content types and elements across systems, variants, and channels. We explore models further on in this article.
Context-based Content Personalization Hinges on the Interplay of Structure and Semantics
In order to create intelligent content, we combine structured content with semantics. This powerful combination adds consistent, predictable order to content, which in turn enables machines to work with it. The addition of semantics also optimizes the human value of the content by adding relevant context.
Once the content has become structured and enriched with metadata, transformation, reuse, and adaption to many different presentations for customer interaction become achievable. This is the key to automated, real-time personalization.
The Master Content Model® (MCM) to the Rescue
The Master Content Model® is used to facilitate the complete, end-to-end content lifecycle and to enable optimization of content for each stage to plan, author, manage, and publish. It is a single integrated model across content sets and representations that defines how content assets will be structured throughout their lifecycle. The MCM focuses on content structure and it accommodates semantics in the form of contextualized metadata.
The many different applications of the MCM must address all of the ways that content will be displayed and used by customers during the content’s lifecycle. To maintain flexibility, the MCM must be software and vendor agnostic. In short, the MCM connects the many ways that content is created to the many ways that consumers use it.
A Foundation for Scalable Personalization
The application of a Master Content Model® significantly reduces friction typically experienced in content creation, management, and publishing. The MCM’s representation of publishing variations for structural elements makes dynamically connected and reusable content possible.
Content modeling injects static content with enough technical intelligence to make it both dynamic and interactive. The MCM simplifies the development of content tools, apps, and APIs. When paired with the context provided by a Customer Data Platform (CDP) or Digital Experience Platform (DXP) database, content can be largely automated and used to generate real-time personalization.
DITA as One Form of Content Structure
The most common solution for structuring technical documentation is the Darwin Information Typing Architecture (DITA), an open standard governed by OASIS. It was developed to solve problems of content reuse, interoperability, and omnichannel publishing.
DITA also helps provide the content’s baseline XML representation for the Master Content Model®; as such, it substantially reduces the number of transformations that are needed. Because DITA is component-based, it can be managed effectively with a Component Content Management System (CCMS). Unlike traditional Content Management Systems (CMS) or Enterprise Content Management (ECM) systems that primarily work at the document level, a CCMS can facilitate content reuse at the component level. Interestingly, components can exist inline, within a paragraph, which allows for remixing of content without requiring separate fields for every component. This significantly increases the content’s ability to flow across systems and solutions.
Enterprises, especially departments who create technical documentation, should care about DITA because it supports:
- Open standards
- Topic-based orientation
- Core structures
- Efficient reuse mechanisms
The basic DITA structures that are the most relevant to technical publications are topics and maps.
The illustration below shows content types and content elements typically found in structured technical documentation.
Schema.org Structure Wires Things Up
It’s important to note that DITA is just one structural standard, albeit a well-typed and useful one. Schema.org is another useful standard used by marketing organizations especially, to encode content with the Linked Data necessary to make it discoverable by Google and other robot consumers across an ever-more semantic web.
And there are many more broad and industry-specific standards around content structure. These structures can all be represented in a content model. The model becomes a schema by which content gets organized. Even if no one really agrees on content modeling approaches and terminology (it’s still a wild west), models matter at many levels.
[A] recommends enterprise publishers adopt a superset schema, a Master Content Model® which may include DITA and schema.org structures, but is not limited by them. Content is too bespoke across enterprise and department needs to live in one set of boxes someone else made up.
Content Structure, Semantics, and Metadata
Content structure: Structure is the organization of things. Content structure is the physical organization of content; it addresses the fundamental needs of publishing — presenting content to people.
Content semantics: Semantics is the study of meaning; meaning emerges from information within systems. Content semantics is the association of specific meanings with specific content structures. Meanings are established as a context within a semantic model.
Semantic models define and organize concepts within knowledge domains. At [A], we have found the following three types of models for semantics to be the most useful:
- Taxonomy: A classification scheme
- Thesaurus: A rubric defining and relating terms
- Ontology: A formal, machine-interpretable domain model
With the added intelligence we have described thus far, we can apply more rules and more properties to content objects to enable various types of useful machine interpretation.
There is an entire ecosystem in the content technology landscape around semantic services. You may wish to read, “Surveying the Semantic Services Vendor Landscape”.
A Master Semantic Model is a comprehensive model that defines concepts and their relationships. It also defines the terminology for identifying concepts across a domain and across systems; this informs the Master Metadata Model that is incorporated into the Master Content Model®. The relationship between our three main models is illustrated below.
At [A], we have found that Master Semantic Model development works best as a four-step process:
- Inventory: Collect any existing semantic models and supporting information.
- Analyze: Identify semantic requirements and evaluate existing models.
- Design: To define semantic models, as well as relationships and context.
- Demonstrate: A test demo or pilot to illustrate how the semantics apply to existing content assets.
Establishing and testing metadata is an iterative process. We eventually establish a Master Metadata Model (MMM) that helps define the structure and potential context for content, working within the Master Content Model®.
The Role of Metadata
Metadata is the critical intersection of structure and semantics. We use it to reflect the semantics we need in our structure so we can support both our content applications and our semantic applications.
Semantics are connected to content via metadata. The graphic below illustrates this connective role of metadata, which can produce different presentations or behavior for the same content asset.
At its most basic level, metadata tells us what is potentially going to happen to a thing or a piece of content. Meta means “after” in Greek. It indicates what can happen to content “after” an applied action or a certain usage.
Metadata resources are:
- Related to external systems which ultimately control the source data.
- Intended to be useful to downstream processes.
- Reusable across content sets, across systems, and across downstream processes.
Think of the Master Metadata Model (MMM) as the map of connectors between structure and semantics . Metadata may include system and integration data, content strategy and business value data, and content operations and workflow data. It’s a huge area that impacts content in all aspects of a lifecycle, and metadata needs discipline and planning to accrue throughout a content lifecycle as effortlessly as possible for busy authors and content admins.
Structure and semantics come together in various representations, including:
- Knowledge Graphs
Bringing together structure and semantics
Structured content, enhanced with semantics and metadata, empowers our content assets to decrease effective costs and deliver relevant content interactions to the right audience at the right time. The velocity of content throughput and distribution increases return-on-investment for pre-sales and/or post-sales content.
At [A], the orchestration of structure and semantics guides content intelligence initiatives within client organizations.
Realizing that content structure is the physical organization of content helps focus teams attention to managing that organization consciously. The form that content takes when represented in a base, like XML, or within other representations of a content object’s schema, directly impacts the ability of that content to transit systems and become useful to channels. When structure agrees between systems, or maps, cleanly from element-to-element, content can move between those systems.
Content semantics provide the essential context for content structures, making it possible to provide advanced content experiences, such as automated personalization.
Finally, metadata is the way we associate specific content structures with specific semantic contexts as defined within one or more semantic models. We make these associations by “annotating” content with metadata.
Content structured for motion creates more value than static content ‘at rest’. Content in a traditional, static forms (e.g. PDF or chunky unstructured HTML) still serves an important purpose, but it has limited findability, narrow utility, and it cannot dynamically serve context-driven interactions.
Engineering content for portability makes intelligent customer experiences possible. Over time, static content will get less and less relevant for customers in digital spaces because it will not be as available.
A Note About Books
Books and long-form content experiences will become even more treasured as assembled experiences start to become normal. The happy accidents of information discovery browsing a library’s shelves, or flipping through a book, or lost in deep reverie with a page-turner…these moments important to healthy humans cannot ever be replaced by automated digital interactions.
Moving Content Grows Value
Content value grows as customers engage with it, and as it drives awareness, interest, attention, and behavior. In other words, content in motion is content that matters.
To move and flow, we must imbue content with structure and enhance it with semantics. We must make the content’s DNA visible and accessible.
Empowering content takes work across strategy, engineering, and operations — and it pays off in the effortless, omnichannel experiences that drive the foundational value of customer-facing organizations.
Content semantics: The association of content structures with specific semantic contexts given one or more semantic models that define and organize the concepts within relevant knowledge domains. Content semantics are used to provide essential context for content structures, making it possible to provide advanced content experiences.
Content structure: The physical organization of content; it addresses aspects such as how content will be modularized, identified, addressed, and reused.
Master Content Model (MCM): A single, integrated model that defines how content will be structured throughout the complete content lifecycle.
Master Metadata Model (MMM): Helps define the structure and potential context for content, working with the Master Content Model.
Master Semantic Model (MSM): A comprehensive model that defines concepts and their relationships.
Metadata: On a general level, it is essentially a set of “statements”; each statement consists of a property or element and its associated value. Within the context of this article, metadata is the way we associate specific content structures with specific semantic contexts as defined within one or more semantic models.
Ontology: A formal description of knowledge as a set of concepts within a domain, and the relationships that are held between them. By having the essential relationships between concepts built into them, ontologies enable automated reasoning about data.
Semantic models: Define and organize concepts within knowledge domains. At [A], there are three types of models: taxonomy, thesaurus, and ontology.
Semantics: Represents the formalization, documentation, and publishing for use of the concepts, with their interrelationships, that are meaningful within the business of an organization.
Taxonomy: A classification scheme that helps you to organize your content and assets into hierarchical relationships. Content assets classified in a taxonomy are better optimized for internal and external searches.
[A] Editor’s Note: This has been modified from article is published by [A] on simplea.com, and was adapted from an original publication in the August 2019 issue of ISTC Communicator. Contributors to this article include Cruce Saunders, Joe Gollner, Maxwell Hoffmann, and others.
Visit the [A] Treasury for more assets about content intelligence.
Introducing The [A]cademy
The Cornerstones of Content Intelligence video course is available for license online through the [A]cademy — [A]’s new learning management system.
It features material on the [A] Content Intelligence Framework, the Content Services Organization, the Master Content Model®, architectural patterns, content supply chain dynamics, and new Content Operating Models.
Learners can watch the videos as their schedules allow, and use the accompanying downloadable materials for reference and continued learning.
For more information, visit simplea.com/academy
Join the Content Order
Be prepared for what lies ahead and receive the latest in intelligent content trends and best practices from [A], the content intelligence service. Join the Content Order to get access to exclusive, members-only content, peer roundtables, event invitations, and more.
Membership in the Content Order is complimentary and open to real participants invested in content. Stay up-to-date and apply to join.