E.g., 11/14/2019
E.g., 11/14/2019

Localization World Madison Review

By: Shelly Priebe (McElroy Translation)


24 November 2008

Lively discussion revolved around the place of machine translation and collaborative work environments in workflow, new concepts and possibilities of shared TMs, development of open source workflow solutions, and creative new approaches to pricing that would allow us to educate clients by providing true transparency into our cost structures. It may also appear as you read this summary that it is necessary to attend this conference just to keep refreshed with industry acronyms!

GALA Speed Networking

In addition to its usual pre-conference business meeting GALA hosted a speed networking session. Although I approached the session with some hesitancy about the value of the format, I walked away a fan. My presumption that years of GALA meeting attendance had resulted in “knowing” most other members proved false on multiple levels. The speed networking format resulted in introductions to 100 percent of attendees, and I soon realized that my usual “colleague comfort zone” was much less than 100 percent. Furthermore, with only a few minutes to “meet” my colleagues and with the mutual focus on how our relationship might be beneficial to each of us, we were all “on task” to maximize the value and substance of the communication. I missed “speed dating” by a couple of decades but now I understand the phenomenon. I concede that a networking style that originally struck me as contrived can be a highly efficient mechanism. And did I mention fun?

Keynote

The keynote by Jeff Howe set the tone for the conference. Jeff Howe also gave the keynote last year. A writer for Wired, Jeff coined the term “crowdsourcing” a few years ago after beginning research on the phenomena of the crowd as a vehicle for fringe musicians to leverage the frenzy of their fan base through virtual presence. When he came across some early commercial applications outside the music industry he suspected that he had stumbled onto something big. His keynote last year was fascinating but seemed only connected to the translation industry by the faintest of dotted lines. Howe’s book “Crowdsourcing: Why the Power of the Crowd is Driving the Future of Business” was released in August, and it is the culmination of much more research. This year his presentation applied more directly to our industry. And this year audience members provided specific examples of the concept in action. In one case an entire website for an AIDS foundation was localized at large by translators who donated time through an open platform. Lionbridge is focusing their technology development on providing a thin layer of technology enabling translators to work collaboratively and in real time together on projects. 

Howe referred to “technology natives” who have grown up interacting with the virtual community in ways that corporate America has only begun to understand. But we respect the power of what we see—and what we don’t. Much as a geyser is a wondrous site it is only the physical manifestation above ground of much more that percolates beneath the earth’s surface. The subculture that brings crowdsourcing into the corporate boardroom is similarly invisible and powerful. The pivotal point to success, advocates Jeff, is to, “Ask not what your community can do for you, but what you can do for your community.”

Managed Communities

Applying the community model to our industry does not necessarily mean getting work done ad hoc and for free. It can mean saying yes to clients who want their vendors to provide the managed service of hosting communities to optimize productivity. This may entail a thin platform using technology to unite community or it may be more evolved with administration of the platform within the client’s technical environment (using Clay Tablet Technologies as an integrator, for example) plus project management. 

In the session “Community Localization: Practical Solution or Pipedream?” there was group consensus that the top vendor management tool currently in use is Excel, followed by Outlook. But that is changing. ProZ manages 1,000 translator requests a day. Lionbridge has created a brokerage site where translators and clients can meet online. Bob Donaldson, Vice President of Technology and Strategy at McElroy, was spot on when he joined our company in January 2007 and identified vendor management technology as the missing technology link in our industry. Now McElroy has much to “show and tell” when it showcases the McElroy Translator HUB at another conference this November. I am genuinely excited about ideas I took away from the conference on how our technology can be adapted to enable McElroy to increase its pro bono workload. 

As more folks look at managed communities as a valid option, questions to be asked are:

  • How do you rate professionals? (Editing and project management will be working together to define what this looks like in our system)
  • How do we manage quality?
  • How are work files organized?
  • How are rates and payments managed?
  • What new kinds of relationships with clients, SLVs and communities are possible?

This was an excellent discussion with a group of forward thinkers. 

Machine Translation

I also attended two sessions on machine translation. One was entirely focused on Machine Translation Post Editing (MTPE). Presented by two agencies that both do $30M+ in sales, the grand conclusion was that translators are not naturally proficient at MTPE and productivity rates vary based on individual proficiency and target quality as directed by the client. Not earth shattering. The result of editing of Human Translation (HT) is generally the discovery of typos or omissions while the result of MTPE is more often the discovery of untranslated or mistranslated key words. It helps to train the client to develop source material that:

  • has only one topic per sentence,
  • uses nouns instead of pronouns,
  • uses short sentences of less than 20 words,
  • does not omit articles.

Eventually I did glean a relevant data point from this session. Target productivity for MTPE is 5,000 to 7,000 words per day. One of the speakers compared the time that it would take to conduct MTPE to the time that it would take to edit 80 to 90 percent fuzzy match TM. She presented extensive data, but in the end we learned that her test group was based on the translation of one document by eight translators. This seemed like a rather insufficient pool of data. 

The next session entitled “Machine Translation – Is it Now?” was much more direct and substantive. One of the founders of Moravia, who was on the panel, spoke about how introducing MTPE to some clients prompts the language service provider to ask itself some important questions. 

  • Where is there money to be made?
  • How will we fairly compensate our people?
  • Which is the most promising technology?

The Moravia panelist was convinced that the Statistical Machine Translation model, such as that used by Microsoft and Language Weaver, is unquestionably where commercial success will lie. This provided an interesting dynamic as another panel member represented rules based machine translation (RBTM) vendor Systran. 

The Localization Value Change – Past, Present and Future

Paula Shannon of Lionbridge skillfully moderated a session of panelists from Welocalize, Jonckers and SDL. She used Charles Dickens “A Christmas Carol” as a literary reference to guide the panel’s analysis through past/present/future with attention to business models, pricing margins, and overhead/infrastructure that typified each era. 

It occurred to me that with McElroy’s later arrival to the localization table we typified until recently the “past” description. This 90s timeframe was described by panelists as an era when just getting to market was a success story for clients and internal workflow and delivery felt a bit like the “Wild West.” Funds released fairly easily and competitive straight translation margins were padded with charges for downstream services like DTP, testing, publication and PM. The “Wild West” of sourcing, production, and delivery sounded close to home. 

But as it turns out the panel agreed that the present is considerably less exuberant and more sobering. Competitors report profits being wiped out by the growing overhead of increased headcount and worldwide sites. Sales expense of more complex sales and the opening of worldwide offices, coupled with participation in corporate audits and the complexities of global finance make doing business now unwieldy and unprofitable. Open standards and education of clients have resulted in client ownership of downstream functions that used to line LSP pockets. In the end the industry has moved from translation of “skinny words” to providing more add on services and now is back at only being able to charge for “skinny words.”  Language service providers (LSPs) helped clients innovate but to what end?

The value proposition LSPs can provide is in the process and the solution, not simply the product. That is the potential of the future. The trend toward Business Process Outsourcing (BPO) is supported by the general increase in services outsourcing and the potential for new types of outsourcing that exists as technology supports BPO. The key to success is open ended partnership with clients with business innovation as the focus. 

In 2006 McElroy Translation Co. engaged futurist Derek Woodgate of “The Futures Group.”  This was a turning point for me as a leader as I transcended from operational coping to engage in strategic “future proofing.”  The action items from our work with the Futures Group gave McElroy a jumpstart on developing a collaborative work environment, integrating machine translation in workflow, and identifying ways to add value for clients via consulting on the best combinations technology, processes and humans. I’ll be the first to admit that just a few years ago much of what we discussed with The Futures Group seemed more like futuristic psycho-babble than business strategy. But while attending the Localization World sessions, I reviewed the outcome of that exercise—just two years later—with a new lens. If you are not innovating you will suffer, because someone else will innovate in your absence. I borrow from Arturo Quintero of Moravia who hails this industry era as a time of “disruptive innovation.”  And from that standpoint, Localization World was relevant and affirming. 

Shelly Orr Priebe is the President of McElroy Translation Company, based in Austin, Texas.