Chapter 2. Intelligent content

One of the challenges facing anyone considering a content strategy, whether on the scale of a single web offering or a global enterprise, is sustainability. It’s only with intelligent content that it becomes possible to talk about a sustainable enterprise content strategy. Automation can be used to minimize the time, effort, and money needed to apply a good content strategy. However, automation doesn’t just happen; content must be consciously designed to support it. An intelligent, unified content strategy establishes a coherent plan under which content will be designed, developed, and deployed to achieve maximum benefit to the customer and the organization while minimizing the cost to the organization.

What is intelligent content?

Historically, content has been managed as documents. Metadata is applied at the documents level to facilitate document search and retrieval for both the customers and the content creators. Unfortunately, metadata applied to a complete document can only adequately describe the content at a very superficial level; it can’t identify the many types of content within the document. The searcher must still examine the complete document and extract the desired information.

If we make the content intelligent by tagging and structuring it, designing and preparing it for discovery and reuse, we can be freed from managing it within the “black box” of a complete document. We can move forward to actually managing the content itself once we take the step of making it intelligent.

Intelligent content is content that is structurally rich and semantically categorized, and is therefore automatically discoverable, reusable, reconfigurable, and adaptable.

Too much of today’s content is stuck in formats that don’t allow you to easily publish to various channels. For example, content destined for a print magazine can’t easily be displayed on the Web with the same look and feel unless you store it as a PDF. And you can’t easily take interactive content destined for the Web and publish it as an attractive, print-based magazine. The problem is that most content is locked within that formatting and changing it takes a lot of additional work and expense. You can’t hope to be responsive to mobile let alone the latest flavor of device to hit the market when your content is wrapped in the straitjacket of format-specific information.

Instead of thinking about how we visually design the content, we need to start thinking about what content is required, by whom, when, in what circumstance, and in conjunction with other content or interactivity. To do this the content has to be structurally rich and semantically enabled.

And it’s not just about format. If we’re to truly make our content accessible to customers, it has to be discoverable. When you have unstructured, untagged, unintelligent content, the information you or your customers are looking for is likely to be hard to find. You have to rely on brute-force search methods to find information. Intelligent content allows you to take advantage of the information contained within the content to make the content more discoverable.

Understanding intelligent content

Let’s take a look at each of the pieces of the definition for intelligent content to understand it better.

Structurally rich

To make our content intelligent so that the system can automatically process it, we need to add structure. Structure is the hierarchical order in which content occurs in an information product. An information product can be a web page, a book, an eBook, a brochure, a training course, and so on. Information products have recognizable structures that are repeated each time the information product is created. Information products consist of components (topics) that also have structure within them.

Structurally rich content is easier for organizations to manage across different products, channels, and departments and it’s easier for authors to write.

Structure is everywhere in content. The more consistent and detailed the structure, the easier it is for customers to read and use the content, and the easier it is for authors to write it.

Structure also makes it possible to manipulate the content. For example, we can automatically determine how to publish content to multiple channels (print, Web, mobile) by mapping the structure of the content to a particular style in the output. Or we can filter out some content (for example, tables may not work as well in the mobile environment, so we design a different method of displaying the information that doesn’t rely on tables). We can perform searches or narrow our search to the particular type of information we’re interested in (for example, all occurrences of a word in the context of a specific element such as a positioning statement).

Structure also frees authors to think about the content itself, rather than the way it should be organized and written, because that’s already been laid out, for example, through templates and guidelines.

Understanding structure

To understand structure, let’s look at the structure of three different recipes:

• Celery salad

• Chocolate-dipped strawberries

• Beer can chicken

In these examples, the semantic structure of the recipes is illustrated.

What does “semantic” mean? It’s a word you hear a lot these days, often without much explanation.

The Oxford English Dictionary defines the word semantic as:

1. the study of the meanings of words and phrases

2. the meaning of words, phrases, or systems

We’ve italicized the key element in both of these definitions—the word meaning(s).

For example, in a Microsoft Word file we have style tags (headers, bullets, and so on), but they’re all about what the content looks like. Headers are larger than the body copy, bullets are indented, numbered lists have numbers in front of them, and most of the other text is tagged as Normal. You can see what the content looks like, but it tells you nothing about the content.

The styles in Microsoft Word imply a hierarchy through the use of font size and indentation, but there is no real hierarchy. For example, a paragraph follows a section title. You can delete the section title and the paragraph remains and Word displays no error messages. In a structured document, the paragraph would have been inserted in the section as a structure that is “under” (subordinate to) the title. If you attempted to delete the section, the system would warn you that you are also going to delete the following dependent elements: section, title, and paragraph.

Compare that with semantically structured content. The text that may be tagged as Normal (in an unstructured Word file) could be tagged as Ingredient in the structured content. This tagging immediately informs the author what information must be placed in that location, and allows us (in the future) to search those places and know that the information we find there consists of ingredients, not just random text. It can also enable us to display the ingredients differently depending on the device. Similarly, in Microsoft Word, a numbered list implies a series of steps, but Word doesn’t “enforce” this structure. In a structured document, a Step must contain an instruction (chop carrots) or action. So as soon as we say something is a Step, the author knows that an instruction or action is required—just as when we say Ingredient, the author knows exactly what to write.

Recipe 1

See Figure 2.1 for the celery salad recipe and semantic structure.

Figure 2.1. Celery salad recipe and associated semantic structure.

Recipe 2

The semantic structure of the chocolate-dipped strawberries recipe is pretty similar to the first one, with one addition. Notice the extra information after the set of instructions. This is a suggestion to the cook, so we’ve added a semantic structure called Suggestion. Notice as well that the step numbers in this recipe have a period after the number, and they don’t in the celery salad recipe. However, the addition of the period is format (how it looks), not structure. Format is handled by the stylesheet, so we can ignore this difference. Stylesheets map your structure to a defined layout, for example, all numbered lists are automatically numbered and have a period following the step number.

The combined semantic structure for both Recipe 1 and Recipe 2 is shown in Figure 2.2.

Figure 2.2. Chocolate-dipped strawberries recipe and associated semantic structure.

Recipe 3

Look at Recipe 3 in Figure 2.3. There’s actually a recipe within the recipe: the beer can chicken recipe contains the rub recipe. So if you create a structured component for recipes, you could then include a recipe within a recipe or use a number of recipes to create a larger recipe.

In addition, both recipes contain a Description at the beginning.

We know, many of you are probably saying your content isn’t anything like a recipe. We use recipes to communicate the concepts. Refer to Chapter 12, “Content modeling: Adaptive content design” for a model for a value proposition and Chapter 16, “It’s all about the content” for a model for product descriptions.

Figure 2.3. Beer can chicken, Italian rub, and associated semantic structure.

The importance of structuring content

By creating and using well-structured content, you create more opportunities for reuse across information products, product families, audiences, and channels. In a structured-authoring environment, where authors follow the same rules or guidelines for each element of content, the potential for reuse is greatly enhanced.

When content isn’t structured, many problems arise. Not only is unstructured content difficult for customers to follow, it’s also difficult for authors to create. And it’s almost impossible to automate your delivery processes without structure. Content in one format can’t be automatically converted to another.

To quote Anne Mulcahy, former CEO of Xerox Corporation, on this subject:

Unstructured content is stupid and old-fashioned. It’s costly, complex, and does not generate a competitive advantage.

Benefits of structured content

There are a number of benefits associated with structured content:

• Reduced costs: Structured content is less costly to create, manage, and deliver. Authors spend less time creating content, and reviewers spend less time reviewing content. Costs of publishing can be virtually eliminated and the costs of adapting your content for multiple devices can be significantly reduced.

• Speed: It’s faster to create content when there’s a pattern to follow. It takes a lot of the guesswork out of trying to determine if “Every recipe has an introductory paragraph” is a structural rule. A structure guides the author in creating the appropriate content.

• Consistency: When we read or use content, we get used to seeing the same types of information in the same place. When things change (when the formatting or structure differs), and there’s no obvious reason for the change, our comprehension slows down. Unexpected or unexplainable change reduces the usability of the information. Developing comprehensive, effective structures can eliminate the inconsistencies that drive customers mad.

• Reuse: The creation of structured content ensures that reusable components are truly reusable, that their reuse is transparent, and that all content appears unified, whether it’s reused or not.

• Predictability: Predictability drives consistency. When information is presented and structured consistently, customers get used to the patterns and structures they see. They can find the information faster and understand it more easily because it’s predictable. Predictability is also very important for automating publication. It’s easier to create stylesheets and automated processing instructions for controlled structures than for ad hoc structures.

Semantically categorized

When we talk about semantically categorized content, we’re talking about content that’s been identified as “meaning something” and is related to other, similar content. We do this by applying metadata to the content, or tagging the content.

So you might tag your content as being related to a particular industry, for example, industry=medical or industry=pharmaceutical. You might identify it as being written for a particular audience: audience=physician, audience=pharmacist, or audience=patient. Or you might tag it with information defining its subject area: subject=diabetes or subject=hypoglycemia.

Using those metadata, you could later find the pieces of content you need to automatically build customized information sets for the industry, audience, and subject.

Without metadata, it’s very difficult to automatically, let alone manually, find the content we need.

Easily discoverable

Every piece of content, including text, video, and audio, can be described and therefore understood if it’s tagged with metadata. Metadata makes it possible to discover content.

We use metadata tagging to drive the search engines and we use the intelligence that we’ve built into the content to allow us to sort through the myriad pieces of information to discover exactly the content we need.

Efficiently reusable

Content reuse is the practice of using existing components of content to develop new materials. Reusable content reduces the time required to create, manage, and publish content, and it also significantly reduces translation costs. We can create modular structured content that can be either easily retrieved for manual reuse or automatically retrieved for automated reuse.

Text-based materials are the easiest to reuse. It’s easier to reuse graphics, charts, and media in their entirety than it is to use portions of them, but it is possible to create reusable media.

Most organizations already reuse content by copying and pasting it wherever they need it. This works well until the content has to be updated. It’s usually very time-consuming to find and change all those places where the content has been used. Not only does this waste time, but you run the real danger of missing some instances of content, which can result in inconsistencies and inaccuracies. In a recent Substantive Audit (refer to Part 3, “Performing a substantive audit: Determining business requirements”) our client indicated that in a recent update they had to manually change content in 48 places! Imagine the consequences of missing one.

Why reuse content?

Reusing content can dramatically improve the way content is created in an organization. Improvements include reduced time and costs for development, review, and maintenance, reduced costs of translation, and increased consistency and quality.

Reduced development, review, and maintenance

Based on our experience, most organizations have a minimum of 25 percent reuse, but it could be as high as 80 percent, depending on how the content is reused. These numbers apply whether you translate content or not.

Development costs are reduced because the amount of content an author has to create is reduced. Authors don’t have to research and write it again; they simply reuse it.

In addition to taking less time to create the content, less time is required to review the content. When approved content is reused, it’s not necessary to review it again, reviewers need only ensure that the reused content fits or makes sense in the current context. This frees them up to do their “real jobs.”

When content is reused, it can be updated automatically everywhere that particular content appears. And if you want to update only certain content but not other content, a smart content management system (CMS) makes it possible to selectively update content (refer to Chapter 20, “The role of content management”).

Translation

You can significantly reduce the cost of translation through reuse. Some areas of savings include:

• The cost of translation is reduced by the percentage of reuse (typically a minimum of 25 percent).

• The cost of reviewing the translated content is also reduced by the percentage of reuse.

• Desktop publishing/post-translation formatting is typically reduced by 30 to 50 percent.

• If four or more languages are translated, typically all costs can be recouped in less than 18 months (including the cost of purchasing a CMS).

Translation memory systems (TMSs) use pattern matching to match content that’s already been translated so the content doesn’t have to be translated again. Every time content is sent for translation, that content is run through the translation memory tool to identify content strings (text) that have already been translated, and as a result, the existing translation is reused. However, each time someone creates and recreates a piece of content, the greater the number of variations that are introduced and the less likely it is that the TMS will find a match. Even if the change is as small as a space or a comma, the TMS will mark it as different. When you ensure that content is reused, translation costs are reduced even further.

Translations vendors charge you a cost to match content, even if the content matches exactly. The cost of matching identical content is a lot less than you pay for actual translation, but it’s still a cost. With a good CMS and modular content you send only components that need to be translated, reducing your costs even further.

Translated content can also be rapidly reconfigured and brand new information products can be delivered from existing information products, without ever having to send that content to translation and pay additional costs.

The less easily measured benefits of consistent structure, consistent terminology, and standardized writing guidelines also help to reduce the cost of translation.

When content is formatted with format tags, translation costs you even more. For example, if you have content in HTML for the Web and the exact same content tagged with format for print, the TMS sees the content as different because the formatting is different. In addition, content often has to be reformatted for publication. If you have structured content, particularly if it’s in XML, it’s easy to automatically reformat content, regardless of language. Translated content is automatically formatted with stylesheets, so only minimal rework is required.

Increased consistency and quality

When there’s no reuse, the chances of inconsistencies in content increase, either because the content’s been rewritten by many people, or because it’s been copied and pasted and some of the occurrences of the content haven’t been updated properly. Often, when content is copied and pasted, the versions of the content begin to diverge over time.

When we examine samples of materials, we find examples of content that’s similar but not exactly the same. On average, we find five to six variations of content. The worst we’ve seen is 56! Yet the majority of the time when we and our client really look at the information, we realize that the content could be identical or there could be a limited number of variations.

When content is written once and reused many times, the content is consistent wherever it’s used. This consistency leads to higher-quality content.

Reconfigurable

Reusable content is modular content. In today’s rapidly changing world, products and customer requirements are constantly changing. Modular, reusable content makes it easy for organizations to rapidly reconfigure their content to meet changing needs. You can easily change the order of modules, include new modules, exclude existing modules, and use modules to build entirely new information products to meet new needs.

For example, a company sold training courses that addressed the issue of diversity in the workplace. They’d been very successful sending their trainers around the world to provide classroom training and had recently developed both virtual training and eLearning. However, more and more customers wanted to license the courses for their own use. Some wanted to have the courses customized for them so all they had to do was deliver them; others wanted access to the source materials so they could customize the courses themselves.

No two companies are alike. Customizations may be as simple as adding a logo, but most companies wanted a learning module or two from one course, a different exercise in a particular module, different terminology, and even localized content. The company ventured into customization but then pulled back. It was too much work and they didn’t have the resources to make all the requested changes and still keep their regular content fresh and up to date. They turned to intelligent content, modularizing and structuring their content. Once the content was intelligent, they could respond to a request for customization in a matter of weeks rather than months.

Completely adaptable

Structured content is content in which the look and feel of the content (format) is not embedded in the content. That makes it very powerful. When we know the structure of the content, we can output that content to multiple channels, adapting it to best meet the needs of the channel, or we can automatically mix and match content to provide what customers want when they want it and the way they want it. We can even transform (reconfigure) content from one structure to another, but only if we know what the structure is in the first place.

We frequently create our content for a particular need or audience, but content can be adapted (used in a different way) to meet a new need often without our knowledge. Think of mashups. We don’t know how our content is being aggregated, but we know that it can be because we’ve structured and tagged it intelligently.

We can also automatically adapt our content to the device it’s delivered to, not just visually, but structurally. Using structure we can identify what content should be displayed when or how. For example, content that is presenting linearly on a web page or in print could be presented as a multitabbed representation of content on a smartphone.

Intelligent content and content strategy

So what does intelligent content have to do with a unified content strategy? Everything! With the speed of change occurring in all industries and with the rate at which new devices for content consumption and interaction are proliferating, organizations must seriously consider how they’re going to make their content accessible for their customers. If you go the route of “handcrafting” your content for one channel only, you risk providing only a partial content strategy for your customers. Customers who can’t get what they want when they want it and in the form they want it in rapidly go elsewhere despite the quality of your products and services. Just look at the dramatic change in the publishing industry to see how a business can be imperiled by having content locked into old formats and technologies. An intelligent, unified content strategy is adaptable to the known and unknown changes you’ll face today and in the future.

Summary

Intelligent content is content that is structurally rich and semantically categorized, and is therefore automatically discoverable, reusable, reconfigurable, and adaptable.

• Structurally rich: To make our content intelligent so that the system can automatically process it, we need to add structure. In addition, the more consistent the structure, the easier it is for customers to read and use the content, and the easier it is for authors to write it.

• Semantically categorized: The word semantic means “meaning.” Semantically categorized content is content that has been tagged with metadata to identify the kind of content within it.

• Automatically discoverable: If the content has semantic tags and is structurally rich, it’s a whole lot easier for customers and systems to find exactly what they’re looking for. The addition of semantic tagging makes it possible to zero in on the required content.

• Reusable: Reusable content, content which is created once and used many times, reduces the time required to create, manage, and publish it. And it reduces translation costs. We can create modular structured and semantically rich content that can either be easily retrieved for manual reuse or automatically retrieved for automated reuse.

• Reconfigurable: Reusable content is modular content. Modular reusable content makes it easy for organizations to rapidly reconfigure their content to meet changing needs. You can easily change the order of modules, include new modules, exclude existing modules, and use modules to build entirely new information products to meet new needs.

• Adaptable: When we know the structure and semantics of the content, we can output that content to multiple channels, adapting it to best meet the needs of the channel, or we can automatically mix and match content to provide what the customer wants when they want it and the way they want it.