The Global Content Lifecycle: How integration generates results

Buyers have become conditioned when evaluating Language Service Providers (LSP) to quickly assess how the vendor will apply Translation Memory (TM) and/or create other efficiencies by utilizing some sort of technology. An example would be a shared combination of technology between the LSP and the client. The overall goal is mutual: to realize a greater return for their translation investment. So, too, are the benefits, which are most often measured monetarily or strategically

The technologies most frequently used for language translation can be categorized as “data management” or “workflow/business process automation”. A need that transcends both of these technologies is quality control. Regardless of what quality standard you subscribe to, it is likely that in some manner you collect and evaluate the quality of your translated content and process data.

Another rapidly growing need is for Content Management System (CMS) integration. This requirement is born from the goal of streamlining the hand-off process from source content management to content translation--two historically disconnected processes. This integration should not be a novel concept, really, to develop a single process workflow for both mono and multilingual content.

DATA MANAGEMENT: The default technology when one thinks of language translation data management is Translation Memory (TM). Regardless of your preference in vendors, most TM products operate in a very similar manner. Source text is stored in a local relational database with a direct association to a target translation. The TM technology was originally created and primarily used by translators for reuse of existing translations, improving consistency, and reducing costs.

WORKFLOW: Language translation workflow technologies vary significantly from vendor to vendor. Many LSP’s use home-grown workflow or collaboration tools to facilitate the global content workflow. Email is still a popular method of moving project files from one point to another. More recently, website project submission portals (e.g., FTP) are growing in popularity. This may be a portal to submit project files and project-related data that is then processed in a traditional manner. Or a web portal may enable the project submission and also serve to communicate project status to users.

TRANSLATION QUALITY: The assessment of language translation quality most often occurs at or near the close of the project by the LSP reviewer or client reviewer. This is an important step that not only helps improve the quality of the file(s) being translated, but can also help educate the translator on stylistic preferences for future projects. If the LSP is managing the quality review process and the client reviewer is performing the assessment of the quality, then the LSP will collect the data and pass it along to the translator, so the TM files with revised content can be repurposed on future projects.

INTEGRATION: A variety of technologies exist that enable data or application integration into other business systems. Some corporate clients may use existing technologies to facilitate workflow and rely on their LSP for data management using TM tools. Usually, the data itself is rarely used to serve other business functions outside of the tactical language translation process. Examples of how this can be done range from supporting many content types, serving the monolingual source data to authors in the content creation process, and streamlining the process by directly integrating with a CMS application. These examples represent strategies implemented by advanced language translation users. For the most part, users are unaware of the potential returns and value they can realize by integrating more broadly or approaching the entire translation process differently.

Each category listed above and the individual technologies that enable them can independently yield positive value. The challenge is that these technologies are most often incongruent. If Language translation buyers could weave together all technology components seamlessly and transparently, that technology can generate exponential returns.

The opportunities to improve both the data management and workflow process are becoming clearer as those in the language translation industry learn from lessons derived by the content management evolution.

The manner in which traditional TM files are managed is one of those opportunities. Let’s assume an average LSP serving a high-volume client can manage 100 TM files, organized by subject matter. It is highly probable that the same source segment exists in at least a few of those TM files. It is also probable that different translations exist for that same source segment. This is a significant problem and is the exact opposite of what the language translation buyer wants to see. While a return may be found when searching the TM files, the integrity of the quality and consistency is highly compromised.

This method of organizing the translation data does not offer an environment for purifying the data or implementing tight quality controls. The deficiency in this model is only further supported by the recent emergence of federated search tools--tools designed to search many TM files quickly in hopes of finding a result for the translator or LSP. While this technology offers the ability to search many TM files quickly, it offers no more benefits than using Google Desktop Search software. It is the equivalent of comparing unstructured data to structured data. Is your goal to find a return or to find an accurate return? Your requirement will drive your decision.

When workflow and data management are not integrated, you lose the ability to properly manage content versions, perform real-time quality control, and maximize the speed in which content moves through the global content lifecycle. If a language translation buyer has the ability to detect content inconsistencies immediately within the workflow, it will result in a more streamlined, valuable solution.

Workflow should span both internal and external users, and should extend to all participants, integrating milestone data. Workflow should not, however, serve only as a glorified replacement for email communication. In addition, workflow should update and advance systematically and automatically. To be accessible to all global content users, workflow should ideally be web-based, requiring no installed software. This could be offered by your LSP or obtained internally, if application development is within your core competency.

Performing language translation quality assessment can be dramatically improved if the content is reviewed, evaluated, and reported at the immediate time of disposition. If a translation segment is evaluated and altered by the approving party, and the update to the database occurs immediately, this action will avoid latency in what historically is the TM updating process.

Further, this method offers an unbiased report on quality. Capturing quality data directly from the reviewer without LSP input or control yields a true assessment of language translation quality.

The measure of an LSP should not be if they get it perfect the first time; it should be on the methods they use to assess and control for continual improvement and reuse. It is not a secondary step requiring additional time. Additionally, when translated content is altered within the workflow, version control and the ability to tightly manage content control is achieved. When language translation data is properly managed within the workflow process and stored in a single, well-structured database, you reduce the time associated in content update and increase content quality and contextual accuracy.

When considering the opportunities for improvement, a user must consider integration into CMS or other business systems. Content Management Systems are deployed to manage content, offering an improvement over what was a decentralized or unstructured content environment. The same principles apply to how we should view multilingual content.

The traditional CMS application does not typically offer the features to deliver rich value in the localization process. Therefore a company using CMS should consider direct integration of their existing CMS into some type of advanced Translation Management System (TMS) in order to improve their localization process.

The global content lifecycle of tomorrow is available today. Language translation data is directly integrated into an adaptable and dynamic workflow technology that offers tight and strict control over content, yet offers flexibility to adjust to ever changing business needs. Language translation data is no longer stored in a ‘decentralized’ manner—many TM files stored in one location—but instead is ‘centralized’ in a structured manner to yield accurate returns.

This constitutes a different kind of repository that extends well beyond the original requirements of a translator needing a TM database. The needs defined by a translator so many years ago are different than those of a corporate client with multiple product lines, business units, and/or company divisions.

Begin with the heart of what you need—a language translation database. This database should not be flat in design, but should be centralized into one physical database (not just one location with multiple file), and logically partitioned (structured according to context.) This will deliver maximum, and most accurate, reuse. It is proven to generate greater returns over traditional TM. If you, as a Language Translation buyer do not have the means internally, request it from your LSP.

Next, seek a web-based workflow solution. It should directly update or write to a centralized database. User access is managed through roles, to support the variety of users involved in this process. As each one of the complex steps of language translation takes place, your global content is being written to this centralized database (not to a myriad of TM files) and you are able to manage version control and purify your content.

The workflow solution should also offer integrated feedback. As a translation segment is reviewed and assessed, any changes are categorized and comments directly input into the system. This integrated feedback most importantly accelerates the process, but also eliminates what could be a biased interpretation by an LSP. A quality translation is not done by default, but is the result of a well-structured process with tight controls.

Now that your data is well-structured, organized and tightly controlled, it makes sense to integrate it with other business systems to maximize the value of this strategic asset. Seek from your vendor an open integration component. This enables interactive access to and from your data in a meaningful way to benefit your enterprise. If you presently own a CMS application, incorporate language translation into it. This is actually relatively easy to do and streamlines the global content process significantly.

If you’ve implemented all these steps, you have one more untapped frontier to conquer--source content consistency. Since we set out to cover the global content lifecycle, it would not be complete without addressing this key function.

Your repository of well-structured source and target content is not only for translation reuse, but also can deliver significant value if used during the content creation process. Applications exist that allow you to interact with your centralized database so that you can make the subtle adjustments to a source sentence that now makes it a 100% match with a previously translated sentence. This offers exponential return. For every source sentence that you make consistent, it can save you 15-25% in new or fuzzy translation costs. This is not meant to institute a strict controlled authoring application, it is meant to offer an easily accessible tool which drives business value.

While the technologies described exist, they are not common with most LSP’s. This is not a fault of the LSP, but rather a reflection of how they have evolved and where their strengths and core competencies reside. It is also a reflection of a dynamically changing language translation landscape where technology is now delivering value that traditional LSP’s cannot extend.

Technologies of 20 years ago can be dressed up and re-branded, but when all is said and done, they are still technologies from 20 years ago. Expect more and draw from what has been learned in the content management environment: moving from paper to electronic versions represented an improvement. Advancing from disparate electronic file storage to centralized file storage showed greater promise. Quickly, indexing and advanced search methods offered more meaningful methods of managing and using that data. Technologies evolve in countless other environments; language translation should be no different.

ClientSide News Magazine - www.clientsidenews.com

Submit your article!