Calculating the Financial Impact of DITA for Translation
by Amber Swope
Success in a global marketplace requires translating content into multiple languages. Moving to a topic-based XML architecture, such as the Darwin Information Typing Architecture (DITA), can help you control the translation process and save money.
The Ripple Effect: How Global Targets Inflate Content Volume
As companies expand into new markets that require deliverables to be translated into additional languages, the cost for creating, translating, and delivering the content increases dramatically. For example, when ABC electronics company decided to expand sales of their DVD players to five European countries, the cost to produce deliverables for the new countries required five times the amount of work in creating, translating, and delivering all the information necessary to support sales, marketing, and service activities. The following table illustrates how the number of deliverables quickly grows when you produce multiple versions of each deliverable and add additional translation requirements.
Content requirements per unit:
Content requirements per unit:
Owners Manual x 10 models
Owners Manual x 10 models x 5 languages
Service Manual x 10 models
Service Manual x 10 models x 5 languages
QuickStart Guide x 10 models
QuickStart Guide x 10 models x 5 languages
= 30 deliverables
= 180 deliverables
In addition to the increased translation work, you have increased the number of deliverables. This means you must also assess the implications, limitations, and suitability of all your processes in the following areas:
Authoring: Do you need to move to structured authoring? If you answer “no” to any of the following questions, then consider changing your content-authoring strategy.
- Does your information architecture scale for the additional work?
- Can you easily reuse content in multiple deliverables and support multi-channel publishing?
- Does your current authoring environment support additional character sets?
Storage: Do you need to move to a content-management system (CMS)? If you answer “no” to any of the following questions, then a content-management system may be a worthwhile investment.
- Can you easily find content to reuse it in multiple deliverables?
- Can you manage the content-development workflow for the increased number of deliverables?
- Can you easily identify the content that does or does not need to be translated?
Translation: Do you need a translation-management system? With the increased deliverable volume and additional language requirements, you need to be able to efficiently manage the translation process.
- Can you efficiently use your translation memory?
- Does your translation workflow support the additional work?
- Do your current vendors have the resources to do the work?
You can change content authoring to DITA without using a CMS; however, to realize true translation savings, you need a strategy for effectively managing the content creation, storage, and translation workflows.
Step 1: Evangelize DITA as the Information Model
DITA is an XML-based open standard that provides the technology foundation that controls and manages creation, translation, workflow, and publishing of content. As a topic-based architecture, it breaks down and structures information into reusable modules called topics, which you then assemble for deliverables using files called DITA maps.
Because its reuse capabilities eliminate redundant effort, it is a popular information model in today’s global, multi-channel environment.
For example, a manufacturing company who adopted DITA after a global expansion that required translation to 14 new languages was able to save more than $225,000 within the first three months of their pilot project on a single documentation set. This savings has enabled them to produce dozens of documentation sets in more than 20 languages.
Step 2: Show Savings for Translation Process
Traditionally, translation management (TM) tools send entire documents through a translation process, which includes previously translated content. Although the translation vendor does not charge you for translating the content again, they do charge you for the processing to determine which content requires translation.
In addition to the processing costs, there are costs incurred when you have to wait until the entire deliverable is complete before you can send it to translation. If you are required to simultaneously ship all the translated versions of the product, the company loses money for every week that the product shipment is delayed due to translation time.
Instead of sending deliverables to translation, with DITA, you send the XML source files. Because you can better manage the translation workflow, you can send content for translation as soon as it is complete, regardless of the deliverables in which it will appear, as well as send only those files that contain changed content. This flexibility allows you to better control costs and reduce release delays due to translation time.
For example, if you have a product guide that uses 200 topics (125 pre-existing topics and 75 new topics), if you are using DITA, you send the 75 new topics to translation instead of incurring processing costs on all 200 topics. In this case, the savings is over 60%.
Step 3: Calculate Measurable ROI
To calculate return on investment, you must first know how much it currently costs to produce and translate your deliverables. In addition to the actual translation costs, consider the following factors that impact content-development cost:
- Direct labor hours per page
- Cost per art file
- % of content rewritten before release of First Edition (revision / churn rate)
- % of new content in future revisions / variants (carry-forward reuse rate)
- Number of expected future product revisions / variants
- % chance to reuse content from this document across other products
- % chance to reuse content from this document across other documents for this product
- Current reuse rate achieved on iterations (before)
- Number of output formats
- Cost for file pre- and post-processing
Once you understand your current costs, you can calculate how much it would cost to translate your content if it was in DITA.
The following example shows how to calculate the cost for product manual that contains 100 topics of 200 words each.
First time translation
$.30 x 20,000
$.30 x 5,000
$.20 x 5,000
50% exact matches
$.10 x 10,000
$.30 x 5,000
50% exact matches
Because you only send new content or the DITA topics that have changed, you do not pay to process unchanged topics. In addition to this savings, because at least 30% of the new content will be phrases already stored in the translation memory, you save at minimum $450. This means that you pay $1,050 per language. If you release new versions of the product twice a year and translate into 10 languages, then your non-DITA cost to update is $70,000 compared to the DITA cost of $21,000, which results in $49,000 of savings.
The biggest opportunity for savings is to send the source rather than the deliverable to translation, especially if the deliverables contain duplicate or similar information.
Step 4: Select Technology Partners
To successfully create, translate, and deliver XML-sourced content, you must have a unified content-development environment (UCDE). This environment contains multiple applications that integrate to provide a seamless workflow.
The first application to consider is the authoring application. Because moving to a topic-based, typed architecture will be new to many authors, it is important that the authoring tool provides an intuitive GUI for authoring topics, and that it is easy to organize topics with DITA maps. Other key features include ease of controlling links with relationship tables, ability to customize the GUI, and support for integrating with other applications in the UCDE.
After you create content, you need to store it; therefore, a content-management system (CMS) is the ideal way to store AND manage your content. Typical CMS capabilities not only provide security, audit, and linking, they also add the power of workflow and metadata support, which aids in reuse.
To publish deliverables, you must transform topics and maps. Each deliverable type requires a different transform. Although the DITA Open Toolkit provides support for many output formats, you need to customize them to produce deliverables that meet your corporate standards.
Content translation management systems provide translation workflow support and translation memory. As illustrated by the above example, translation memory management is critical to localization savings.
To make the case for moving to DITA and XML for content authoring, you need to understand how translation affects content development and production. Although teams who do not have translation requirements will also realize cost savings through improved efficiency, content reuse, and content sharing, translation costs are where a topic-based architecture really makes the difference.
About the Author
Amber Swope is an experienced information architect with almost 20 years in the information development field. At IBM, she led the first HTML-to-DITA migration project for the Rational division and implemented DITA in a production environment. Amber is a member of the OASIS DITA Technical Committee and participates on the Learning and Training Specialization subcommittee. Amber has authored numerous papers and articles on information design, development, and architecture, and has presented at leading industry conferences.