Godfrey Baker
Group Director, Engineering
Organic
gbaker@organic.com
|
Search Our Catalog of Articles
Five Best Practices for Content Management Application Development
The hottest area in web development today is arguably content management application (CMA) development. For any large enterprise, one of the primary purposes of any web initiative, whether it is internal (intranet), or external (internet) development, is to display, in an organized fashion, large collections of company material. This material may be press releases, forms, human resources information, job postings or any other intellectual property, but it all must be located, entered, approved and published. This is the role of the CMA.
Implementing a CMA is a cross-enterprise undertaking. Almost all members of the organization are impacted, either through a higher demand for original content, distributed editing and publishing activities, or through more pervasive utilization of information. The Content Management Application is also an integral part of the organization's overall web strategy, since the content entered into the CMA is the fuel that powers the various
web initiatives.
Organic has designed and implemented dozens of CMAs over the last few years for a variety of retail, financial service, telecom and entertainment companies. Starting before off-the-shelf applications from companies like Broadvision, Vignette and Interwoven were available, Organic worked to create and apply best practices on taxonomy creation, meta-data tagging, and dynamic content retrieval to custom coded content management applications. As fully developed CMAs like Vignette's Story Server and Interwoven's Teamsite became viable systems capable of supporting rich and complex content sets, Organic continued to apply these best practices to the new pre-packaged solutions. Today, these well-developed technical strategies and information architectures guide the latest enterprise level CMA implementations and act as a foundation for compelling, maintainable, and effective web initiatives.
1. Don't skimp on the content inventory assessment
The
first key to a successful implementation is to ensure a complete understanding of the existing content on all web properties to be serviced by the CMA. This should be a detailed spreadsheet listing filename, owner, file type (word document, PDF, graphic), template type, and categorization. The content inventory needs to catalog all HTML and document content on all targeted sites. Correctly inventorying the content is a significant task and a large system can have tens of thousands of content items, but the team must not fall into the trap of generalizing content areas - every content item needs to be analyzed and categorized. A complete content inventory achieves the following:
- Establishes a content baseline
- Forms the basis for the content taxonomy
- Informs content conversion
- Establishes content mapping requirements
2. Determine content conversion needs early
The content inventory will inform the implementation team about content
that requires a repurposing effort before being deployed in the next CMA release. Content conversion may include applying new meta-data to content, migrating stored database content between data models, and "scrubbing" or removing embedded HTML and other legacy symbolics from the content. Additionally, during a content conversion effort, Word documents, Acrobat (PDF) files, Powerpoint files, etc. should be examined for proper branding, current logos, and approved formats.
Portions of content conversion can be automated, however, the programming effort for automated conversion must be evaluated against a "brute force" method of re-keying old content. While it is tempting to assume that all content can be programmatically parsed and migrated to the new CMA, structural differences can rapidly surface requiring special processing and complicated logic. Typically, a fast, automated program to migrate some of the content followed by a comprehensive effort to manually move
the rest is the most economically viable technique. In any case, the content inventory must be analyzed for the level of effort required for the migration and resources planned accordingly. Content conversion is critical for the project's success and should never be an afterthought left for the end of the project.
3. Develop a scaleable taxonomy that will support future needs
The taxonomy structure must be able to support content growth and content acquisition for the projected CMA lifetime. Since the taxonomy author cannot completely anticipate what new content will surface in the future, the structure must be flexible enough to support new content types. For instance, when classifying forms, the taxonomy should not contain an element such as 'HR Hiring Forms". This is non-scalable because we have many kinds of forms for many departments, and we would have to add every permutation to get complete coverage. Whenever support for a new form or department is
required, many new elements would need to be added to the taxonomy. Additionally, if we suddenly had to add, say, "reference guides" for each department, we would have to add a whole new series of taxonomy elements.
The scaleable approach is to have one category for "Department Topic" another for "Department", and finally one for "Information Type". Values are then associated with these categories i.e. "Human Resources" for Department, "Recruitment" for Department Topic, and "Form" for information type. To handle payroll forms as well as hiring forms we only have to add a new element, "Payroll", for the Department Topic category. Projecting forward, we see that if we suddenly find a "Web Site Reference Guide" we could fit it as Department - "Information Technology", Department Topic - "Web Site", and Information Type - "Reference Guide". Database designers will recognize this exercise as data normalization. A normalized taxonomy will scale to support multiple repositories
of content.
4. Keep meta-tags separate from information architecture (site navigation)
Applying the taxonomy is known as meta-tagging the content. The meta-tags categorize each content item and organize the content repository. Ultimately when the content is presented on the site, it is the meta-tags (or tags) that drive the extraction and placement of content items as specified in the information architecture.
When applying the tags to the content, it is important to tag independently of the information architecture of the site. It is tempting to consider the tagging structure as means for "placing" content on a site, i.e. by tagging an item as a form it will render in the forms section of the site. This tactic will, however, sharply reduce the power of the taxonomy and limit the usefulness of the meta-tags. Instead, each piece of content should be considered as an independent entity and tagged as such. A form should be tagged as a form because
it is a form; where it gets rendered, how it gets rendered, or even if it gets rendered is of little consequence to the meta-tagging structure. It is the information architecture that specifies the location and presentation style of the content items; the tagging structure simply identifies to the system the organizational attributes of each content item.
Strong partitioning of meta-tags and information architecture will allow:
- Workflow separation between content producers and information architects
- Logical use of content items in multiple areas on the same site
- Logical use of content items across multiple sites in the enterprise
- Specialized sub-sites delivering specific content types
- Use of content types in meaningful ways unforeseen by the original information architecture
5. Educate the stakeholders
The Return on Investment from a content management system deployment lies in the distributed nature of the
application. Content management systems allow business units to handle their own content creation and content publishing without the assistance or gate-keeping of an IT or E-business group. The unfettered business units can then focus on creating content for the enterprise while the E-business groups concentrate on building and maintaining the actual systems. The resulting win-win enables fresh, relevant content to be efficiently produced and deployed on stable and highly optimized systems.
However, distributing a web publishing system to unprepared business units will create unfavorable results. Web publishing has unique requirements that might be unfamiliar even to skilled corporate communication groups. Units not experienced in internal communication such as Legal, Compliance, and Facilities will have an even steeper learning curve. Creative aspects such as voice, tone and color have specialized requirements for the Web, and the depth and breadth of content required
to make a compelling and engaging user experience might surprise the authors and publishers. In some cases, the fundamentals of intranet or extranet management may be completely foreign to the stakeholders or effected employees requiring basic educational groundwork.
Stakeholder education is key to a successful content management roll-out. A CMA with embedded workflow represents a business re-engineering activity and needs to be handled with formalized change management tools. Senior stakeholders must be identified and paired with delegates. Content publishers and content creators must be solicited and sold on the benefits of the system. Training materials must be created and presented methodically to the organization. While ultimately a CMA is a self-serve application, its distributed nature and digital character require a thoughtful and thorough educational effort in parallel with the technical rollout.
Conclusion
Content Management Applications
are one of the highest ROI implementations available to the E-business community today. Off the shelf packages take the technical pain out of the systems development and can significantly enhance the project timeline, but development hurdles remain nonetheless.
The team should start by executing a thorough content assessment to baseline the current content inventory. This will lead to a solid content conversion strategy and will help inform the taxonomy. By extending and normalizing the taxonomy, the team will create an organizational framework that will support content additions and changes throughout the program lifecycle. As the project moves into development, this taxonomy can be applied through a meta-tagging effort capable of supporting web initiatives that extend beyond any one project, creating, in effect, an enterprise-wide content repository. During project rollout, by educating the stakeholders and effected employees, the true power and effectiveness of
distributed systems, intranet and extranets, and content management applications can be communicated. The end result - distributed teams of efficient and satisfied producers creating meaningful and compelling content capable of powering enterprise-wide web systems.
|