Search Intranets
Current Issue
January/February 2012
Editorial
Columns
Features
News & Tools
Read_Me_File

Services
About Intranets
Subscribe to
Intranets
Past Issues
Sample Issue (PDF)
Content Choreography: Content Management for Personalization
- November/December 2011 Issue Posted Nov 1, 2011 Print Version  
Page 1

Content providers face a host of challenges-new information sources, new channels, new formats, a new social networking culture, new competitors, and above all, higher consumer expectations. As the information environment grows more complex, users paradoxically expect the information experience to be both richer and easier.

There are more moving parts to manage, more applications, and more complex models of information components. The solution to information and application proliferation is not, as some vendors recommend, yet another application. It is the choreography of content management components to create a harmonious whole.

While content choreography may appear to be something content management professionals have been doing all along, the concept conveys a new approach. Traditional content management processes are often siloed and monolithic-single-dimension taxonomies, disconnected processes, content assembled based on simple scenarios. Content choreography brings together the best of several disciplines in order to selectively present information to users as they need it.

The concept emerged when a client was having problems with a complex website. As we articulated our approach, we kept hearing that there were groups responsible for each component. User experience? Check. Search site? Check. Personalization? Got it! SEO? Yup. Content creation, content modeling, taxonomy, metadata, IA, tagging, user testing, socialization, and training? Check, check, check, etc.

But there were still problems pulling all of the elements together. What was missing? It appeared that each element was considered to be in isolation. But the whole is greater than the sum of the parts. The system is more complex than its components. It turned out that this client faced two major obstacles: Communication and collaboration were not taking place among the groups, and user scenarios were not accounting for all the variation in how components needed to interoperate to meet business objectives for personalization.

Just as choreography leverages technical knowledge and expertise but combines elements into something with an intangible character, content choreography contains technical elements assembled with judgment and creativity. There are many variables that can be adjusted to optimize the outcome for a unique set of requirements. Achieving an elegant user experience requires paying attention to each of the elements and considering how these elements interact.

Successful content choreography begins with an overarching strategy that unifies each of the components. The strategy itself requires consideration of multiple dimensions: business, editorial, technical, deployment, management, and solution architecture.

Content choreography requires a platform designed to support dynamic content delivery. Diagram 1 shows a representative platform architecture. This architecture brings together many typical components. Most importantly, from a content choreography perspective, we need to think of the "platform" as involving more than the software applications. It's the platform applications blended together with critical processes and core design assumptions. Some of these include site search facets, the ingestion and indexing process, taxonomy, and governance.

There are quite a few components that must be configured together; this may seem overwhelming to an organization with diversified ownership of the various components. However, once we set out to achieve specific goals, certain relationships between elements come into focus. Diagram 2 highlights elements that are key to addressing SEO across a large repository of content components. We can now see that there are new relationships between the taxonomy, the thesaurus, translation management, campaigns, and page-naming conventions that were not apparent before.

If we take another objective, say, improving the reusability of content components resulting from content authoring and tagging processes, a new set of components require attention and re-integration. Terms still come from the metadata schema and taxonomy, but they are now considered in the context of content authoring. Governance is a bigger consideration since there are many processes that need to operate on content creation, curation, and tagging standards. There are issues around change control that are now apparent in this process context.

Personalizing Content

One of the most important aspects of correctly choreographing dynamic content management systems is an understanding of user context and associating that context with appropriate content. Defining context takes persona analysis to a new level.

Beyond identifying what a user wants to do, we need to define what content will provide the most value. There are a number of approaches. One approach is to identify content related to a process or a specific step in
a process. For example, for a salesperson engaged in developing a proposal, related content might consist of competitive analysis, market studies, or case examples. When a salesperson searches for related proposals, we can anticipate the value of this related information. This context of informational needs is derived by understanding the users' line of thinking when they are executing their tasks: What they need, what they look for, and what supporting information will assist them in the process.

A second approach is to infer needs from how users reached a particular page: what terms they searched on, where they came from, and what they were doing prior to arrival, what pages they viewed.

A third approach is to make inferences based on what we know about the users. This knowledge may come from asking for preferences such as topics of interest, authors they read, their stated demographics (age, income, education, etc). We can also make inferences based on knowledge of their geographic locations (either at a region or hyperlocal level, for example, shopping in a particular area) or where they are going or coming from (which is a little more difficult to determine unless permission is provided to track detailed location).

Both passive and active personalization requires that the designer of the system (typically an information architect) develop a list of characteristics that are matched up with preferences. For example, if I am
an attorney (and this can be determined by matching login credentials to a profile), then present information from the legal databases; if I'm an engineer, draw from the engineering databases. Passive person-
alization can be as simple as presenting corporate news to people who work in headquarters and branch news to people who work in the branch.

It is a good practice to provide users with opportunities to broaden their selections-therefore, it is important to make the personalization choices visible to users. I may be an attorney doing research for an engineering contract, in which case I want to view engineering content.

Content choreography is the integration, assembly, and management of various elements of a dynamic content system. These elements need to be considered in relation to one another and not simply as work streams or independent elements. Creating an elegant, satisfying user experience that presents information in the context of the users' tasks will reduce the perception of information overload and lead to higher conversions, more successful content reuse, and greater return on investment in content tools and technologies.  

Print Version  
Page 1