Skip to content

Robust Metadata: The Secret Sauce of Relevance


People feel overwhelmed by the volume of content available. They want only content that is relevant to their needs and interests. But where does relevance come from? The basis of relevance is not a mystery, but it is driven by something that can seem mysterious: metadata.

Metadata’s overlooked role in communication

Metadata helps to make content relevant to your audiences. So it is important to understand how metadata is relevant to you as a content publisher. Metadata acts like background information: the kind of information that’s useful to know, but you don’t want to think about much. I call metadata a secret sauce because when it works right, it’s invisible: You don’t notice the metadata; you just notice that things work right.

That’s what we aim for in our content.

We interact with metadata every day: It organizes appointments in our calendars, helps us sort through our documents, and lets us find things on the web. It’s a constant companion, but because it acts in the background, we may feel we have only a vague or incomplete idea of what it is.

I like a definition of metadata provided by Mikael Nilsson who said, “Metadata is descriptive data about identifiable things.” That thing may be your content, but it may also be your customer, or the device your customer is using to view your content.

This is an important point: Metadata is about more than just your content.

Metadata is a language, much like the words in this article or the formulas in a spreadsheet. As a language, metadata enables conversation. In the case of metadata, the conversation is between items of content. Good conversationalists use language skillfully to convey what they’re interested in. Metadata needs to be understood and to be convincing.

If you’re a publisher, you are already using metadata. But you may not be using it to its full advantage.

Indicating how something fits in

When we want to indicate how something is relevant, our first inclination is to identify what the thing is and where it fits within a structure. But that might not be enough.

Suppose a new hire or a prospective client asks you to explain the organization where you work. You might trot out your firm’s organization chart showing the structure of your firm. You would point to various boxes: Here is Bob, he’s VP for Sales, and Mary is VP for Customer Service. The structure presents lots of boxes that show how people are formally related to each other. There’s a hierarchy: The box at the top is more important than the boxes at the bottom.

But if you wanted to get a more complete picture of what’s happening at the firm, you might ask: Who are the key experts on a topic? Who has worked with whom in the past? Who is the real rainmaker? Answers to these questions establish the context of the organization and help clarify what an organization is about.

We hear a lot these days about semantic structure. People involved with search engine optimization often talk about structured data and semantic search. Content strategists emphasize the importance of semantically described, structurally defined content to provide the right information to people.

When people talk about making content semantic, they normally only mean describing what something is, not what it is about more generally.

A good example is the markup that websites use to improve search ranking. This markup helps Google and other search engines identify what the content is about. But Google isn’t giving you any choices about how you show your content, such as whether to show some content to certain audiences and other content to other audiences. markup is very good at telling what something is, but it is not very helpful telling your computer systems how to use the information in different situations.

To make the best use of content, we need to know the context in which an item of content belongs. Metadata has two roles: to indicate what something is and, more broadly, to indicate what it is about and where it belongs.

People often compare content to food, which some people regard as the most interesting content of all. So let’s consider how your content is like a cooking ingredient. When my wife and I visit our traditional neighborhood food market in Rome, there are numerous stands selling a dizzying assortment of produce, meats, and dairy products that come from local Italian farms. We see Mediterranean fish that we wouldn’t see in the United States. We can ask what the fish is, but sometimes we are no wiser after hearing the answer. Here’s what we need to know: What does it taste like? How do we prepare it? What goes well with it? Knowing whether to prepare the fish with lemon or with tomato sauce suggests appropriate accompanying side dishes and wines.

I check out the pasta stand. There are many pastas, all the same color and identical in chemical composition, but of various shapes and sizes. There are dozens of names for pasta types. The important thing to know is why they are shaped the way they are. I’ve learned that each shape is created to accommodate a certain kind of sauce: chunky, smooth, or gooey; oil-based, vegetable-based, dairy-based, or broth-based. Knowing what type of sauce is meant for a type of pasta suggests the right accompanying dishes and wines. Stated more broadly, it’s about knowing the intended context for an ingredient.

What’s the tie-in with metadata? It’s not enough to say what something is, whether that something is a piece of pasta or a piece of content. People need to understand how that thing interacts with other things (with pasta sauces, say, or with other pieces of content).

Here’s another analogy. A lot of metadata is like the boring guy at a party who only talks about himself. Metadata needs to describe not just isolated pieces of content (as a boring guy would do) but also the wider environment in which the content is situated (the rest of the party).

Instead of being self-centered, our metadata needs to apply the maxim of the late American children’s TV host, Mr. Rogers: Let others know they are special.

In short, metadata needs to describe more than the pieces of content themselves; it needs to also describe how those pieces of content relate to other pieces of content – or to other things.

Attributes: Defining the context for the content

Our customers want to feel special, and our content needs to make them feel that way. To do that, our content needs to adapt to various situations. When metadata provides contextual information, it can enable the content to adapt to the precise needs of our customers.

We provide contextual relevance through metadata attributes. Attribute metadata extends the descriptive data about your content to give it more precision. The attributes are additional comments about the topics you have tagged; they enable you to indicate how a topic is to appear in a given context.

Most metadata is about semantics, saying what something is. Semantics are like our names: My name is Michael. Attributes are like our body language. They signal interest and express intentions. They establish the relationship.

Semantics define the topic being discussed. Attributes define the context. Robust metadata includes both.

To provide content that adapts to the changing needs of audiences, we need metadata to specify attributes that indicate the preferred context and what that content is about in general.

In what ways might content need to adapt? Brands face a daunting range of situations where customers might want their content. Not all your customers are the same. They may comprise different segments. They live in different places. They use different devices. They have different priorities at different times of the year. All these variations can affect how they need to receive content.

That amount of variation may seem overwhelming, but robust metadata can help you manage it.

Taming variations through metadata

Here’s how to tame all those variations through metadata.

  1. Start with a solid base of content.
  2. Prioritize the situations that have the most significance to your customers and the most impact on your business goals.
  3. Use metadata attributes to describe the variations you want to make available.

Metadata attributes: An example

Metadata attributes can benefit your business in numerous ways. A simplified example gives you a taste of how they work.

Let’s suppose your business is a chain of hospitals. You produce a series of advice articles about wellness to promote your wellness programs. One of the articles is about how to meditate to reduce stress. Once you develop a default version of this article – your solid base of content – you can modify it to address the most important variations, possibly the following:

  • The device the article will be viewed on
  • The time the article will be delivered
  • The location the article will be delivered to
  • The audience that will be reading the article

Let’s say you’ve identified several opportunities to reach high-value customers on the basis of their situational needs. You want to prioritize content variations that address those situations.

So for your article on the topic of how to reduce stress by meditating, you might create a plan that looks like this:

  • Default version – a three-paragraph article, with the third paragraph inviting people to attend your wellness-program orientation
  • Variation 1 (audience = working mothers) – additional opening paragraph addressing the needs of working mothers
  • Variation 2 (audience = retirees) – additional opening paragraph addressing the needs of retirees
  • Variation 3 (device = smartphone) – additional link to the hospital’s smartphone app
  • Variation 4 (location = airport) – additional opening paragraph on meditating in airports and information on the meditation room at that airport
  • Variation 5 (event = holidays) – additional opening paragraph on stress during the holidays and information on special holiday public programs

Airport business woman on smart phone at gate waiting in termina

When robust metadata is incorporated into a well-designed system of intelligent content, the content can adapt to various circumstances of use, such as the user, the device, the time of day, and the location. For example, a traveller in an airport who is reading an advice article on meditation might find additional location-specific information about that airport’s meditation room.

The variations in parentheses are the attributes represented in the metadata. People viewing each variant are unaware that they are seeing a special version; they simply focus on content that’s relevant to their interests. In the background, the metadata enables the variations. And the hospital can use the metadata to track how each variation performs compared to the default version. Was it a good choice to target working mothers instead of working parents in general or instead of also including a working father’s variation?

Perhaps the hospital finds that meditation, unlike other wellness topics such as controlling blood pressure, does not appeal to retirees. With standard web analytics, the hospital content team might see that few retirees accessed content about meditation, but they wouldn’t know whether low access was due to a general lack of interest in the topic or to a failure of the default content to appeal to the segment. By producing a content variation targeted at retirees, customizing the content for the target segment, the hosptial could more accurately measure the level of interest in the topic.

Attribute metadata helps teams to focus on what is special about the needs of a high-priority segment, and what needs to be special about the content. It enables content variations that go beyond what’s possible using semantics and structure alone. It may enable content to adapt to the person using it, the time and place of use, the device being used, and other things that vary with circumstances.

Robust metadata and adaptive content

Adaptive content requires that content be described in a manner that enables it to interact with IT systems to deliver contextually appropriate content. For example, many brands are investing in customer experience management technologies, location-based marketing solutions, and other tools that promise to deliver more precise outcomes.

To realize the promise of these tools, the content must adapt to the needs of customers rather than simply being pushed at them uniformly. Robust metadata is the critical bridge between content variations and marketing technologies that can enable the content to respond to changing needs of customers according to their context.


Some people regard metadata as a technical detail of limited significance. Since they don’t see it, they don’t understand it, and they assume that it has little value.

But metadata can be a strategic asset, much like your audience-facing content. Robust metadata allows your content to present itself more versatilely: It can help you gain attention in a noisy world.

Unfortunately, not all technology vendors support robust metadata that describes both semantics and attribute variations. They assume that primitive metadata is fine; they may instead offer proprietary features that deliver only limited content customization. But it’s important not to be stuck with yesterday’s thinking. Adaptive content delivery requires the flexibility that attribute metadata offers. Without attribute metadata, you can’t deliver truly adaptive content.

Content professionals, whether involved in editorial planning, content promotion and marketing, or IT development and deployment, all need to become familiar not just with the concept of semantic metadata but also with attribute metadata. Organizations must ask their vendors to show them how they can deliver and track usage of content variations according to audience segment, device, location, or other dimensions. Brainstorm some scenarios of variations you might want to try, and learn whether your existing capabilities can offer such variations.

Has your team used attribute metadata to create content variations? If so, tell us about it in a comment below.

Learn more

For more on the role of metadata in content variations, check out these articles:

Thriving in a World of Change by Karen McGrane

Adaptive Content for a Future-Proofed World by Lacey Kruger

Adaptive Headlines: The Right Genre for the Right Context by Michael Andrews

Who Wants What Your CM Puts Out? by Bob Boiko

Want to learn more about intelligent content? Sign up for the Intelligent Content weekly email newsletter. Not only will you receive our new posts, but you’ll also get an exclusive article from Robert Rose each week.

Cover image by Joseph Kalinowski/Content Marketing Institute