Most software problems have nothing to do with technology or standards, but they have everything to do with people. I learned that from Bill Gates in 1989. When I interviewed him for the German edition of Forbes magazine, he told me the following. When programmers say that something cannot be done, in 80% of the cases, it means that they do not like your idea. Fifteen percent of the time it means that they don’t know how to do it, and in only 5 percent of the cases does it mean that something really cannot be done.

I was moving from journalism to multimedia publishing back then and supervised producers and programmers. Ten years later, I moved onto localization and found that Gates’ statement also applied to the language industry. If translation professionals claim that a particular tool doesn't work, they either don't like your (typically management's) idea of using it in the first place, or they do not know how to use it properly. Granted, some software really stinks, but only in a very few instances are translation tools really useless.

User Acceptance Is the Number One Challenge

In every translation technology integration project I have seen, whether in the public or private sector, the number one issue is always user acceptance. People struggle with a new work environment, new ways of getting things done, different means of sharing data, additional workload and uncommon ways of communicating. Every function is included to some degree: Project Managers, Translators and Terminologists – just as much as the very same IT people who are helping to integrate the new technology.

To illustrate how difficult it can be to change user habits and attitudes, I often compare a Localization Project Manager to a parent. If you, as a parent, have a babysitter whom you like, who is always available, affordable, and whom your children like, you will not share that phone number – not even with your best friend. Yet, Global Translation Management Systems (GMS) ask the impossible from Localization Project Managers when they do just this – by requiring that they share access to their best Translators. Once they find a Translator who is good, reliable, available and affordable, the best Project Managers will not easily share that name – not even with their boss. In addition, these same managers really believe that they, not the organization, ‘own’ access to a Translator. After all, specific Translators have helped them to become and remain successful. Making those resources available to others can jeopardize individual success. So, why share? Who cares about all the bells and whistles of a translator database?

Technical Problem or Adaptive Change

But instead of working on users, software companies and buyers work on technical changes to the product, often introducing even more complex solutions that lead to even less acceptance. However, successful implementations of translation solutions and standards require a sensible approach to adaptive change.

Ronald Heifetz, co-founder of the Center for Public Leadership at Harvard University’s Kennedy School of Leadership, explains: “The difference between an adaptive problem and a technical one is key. There are problems that are just technical. I'm delighted when a car mechanic fixes my car, an orthopedic surgeon gives me back a healed bone, or an internist gives me penicillin and cures my pneumonia. That's a key question: is this a problem that an expert can fix, or is this a problem that is going to require people to change their values, their behavior, or their attitudes? For this problem to be solved, are people going to need to learn new ways of doing?” (For further reference I highly recommend Heifetz’s book, Leadership Without Easy Answers).

More TM users – But Less Experienced

To help produce the desired results, technology requires changes in the way people work together, communicate, process and hand off projects. For example, the introduction of translation memory (TM) requires translation professionals to share and manage a knowledge base with other colleagues and freelancers. It asks Linguists to leave a comfortable editing environment. It means that Localization Project Managers must adopt different steps to qualify projects and new ways to handle files. It translates into additional support and system management from IT personnel. Just to name a few.

LISA’s recent survey report on Translation Memory (TM) hints at the difficulties that users may experience when integrating this technology. While the report indicates a strong growth in TM usage, many of the companies surveyed are using TM for their first projects, and only a third use TM for all content. A mere 6% of respondents use TM for tasks other than translation, such as analysis for quotations, verification of client proprietary needs, or consistency checks. In fact, the percentage of users performing other tasks has declined since 2002. Interest in open standards, such as TMX, has declined by 50%, despite the many advantages they bring for sharing knowledge. These data tell me that while more people are using TM, less are using it to its full potential.

Where Implementations Go Wrong

I have observed five main areas where translation technology implementations often go wrong.

Stakeholders

The World Bank's Business Practices Survey (a joint effort with LISA) reveals that Translators have a vote when buying a translation memory tool only 30% of the time. This is a recipe for failure – users need to be included. I say it over and over again: the best tool is the one most accepted by users, i.e., Linguists, Translators and Terminologists. For example, it does not matter which translation memory product produces the highest number of matches, but rather which one will be used by more people and for the longest period of time.

Workflows

Every introduction of technology requires a review of current processes and project communication. Whatever flaws appear in an existing localization workflow, technology (and especially GMS) will amplify them. Changes will need to be made in project administration from qualification to archiving. File handling and data management must be handled sensibly. Of particular importance are new roles and responsibilities; for example, who will approve the translated segments to go into the memory, and what will the process be? How will projects be analyzed and what data will need to be shared, and how?

Leadership

In times of distress, people look for leaders to help them out. If nobody is responsible for the implementation, users will find their own ways to work around the technology. If you are serious about the return on investment in translation technology, you need an Implementation Manager or a Computer-Assisted Translation (CAT) Manager who takes on this role. I know what some of you are now thinking: “Another position to be added to the headcount.” But this is one well worth it. When I became Managing Director for Central and Eastern Europe for Berlitz in 1997, that was one of the first decisions I made. At the time, even the appointed CAT Manager wasn’t sure if responsibilities would justify a full-time position. In the end, she was so instrumental in bringing production margins up that we turned a profit within only four months.

Training

It is rarely enough to provide only a general introduction to a new tool. Eighty percent of learning occurs within the workplace. Proactive guidance and immediate user support are key to reinforcing new processes, techniques, roles and responsibilities. This can be done by the CAT Manager, but depending on the size of your organization, the help of the IT group and the support of power users may also be necessary.

Performance Standards

I believe that the vast majority of employees want to do a great job. But change asks them to undergo a sustained period of stress in which it is unclear if they will be able to perform. The introduction of translation tools requires professionals to acquire new competencies. Old skills that once were desired are often no longer needed. The question of job security becomes an issue immediately. Managers can help by defining the organization’s new definitions of good performance and success. For example, is it really important how many words an in-house Translator can translate per day on average? Or, is it more relevant to track the number of words that same Translator can manage using technology integrated with a pool of freelancers?

I believe we are at a crucial point, where technology developments have achieved their peak, but have left the users behind. The question is not, what additional features are required, but rather, how can we facilitate adaptive change, so that translation professionals maximize the use of technology we have available. I am not saying that we do not need better technology and standards. We simply also need to adopt a better and more enlightened approach to their implementation, along with the inclusion of users throughout the process.

Andrew Draheim oversees the implementation of the World Bank’s translation framework. Prior to his current assignment, he managed the organization’s translation and interpretation unit. Before joining the World Bank, Draheim worked as Berlitz’s Managing Director for Central and Eastern Europe. The views expressed in this article are his own and do not in any way represent the view of his employer. Draheim can be reached at adraheim@worldbank.org.