In July 2005 SDL acquired Trados, becoming the largest supplier of software solutions intended to support a global content value chain. We recently spoke with Keith Mills, SDL's chief technologist, and Terry Lawlor, VP of marketing, to hear about SDL's plan for melding the software offerings of the two firms to support its vision of enabling global information management across an enterprise or supply chain. SDL aspires to offer language services plus a globalization technology backbone that extends from the desktop to the data center. When CEO Mark Lancaster announced the acquisition in June, he committed to a five-year grace period during which the company would support both SDL and Trados product lines as they worked toward integrating the two.

As a result of the acquisition, SDL has 50 offices around the globe. It claims more than 100,000 end-user licenses, most of them for translation memory, with Trados outpacing SDLX two-to-one. Installations of SDL’s three translation workflow systems – Trados GXT, Trados TeamWorks, and SDL Translation Management System (TMS) – account for more than 100 enterprise installations. Now that he has had a chance to roll up his sleeves and review what SDL bought, Mills told us that he is pleased with Trados’ features and architecture. We discussed each of the sectors in which SDL fields a product offering and the integration plans for each.

• Translation memory. By year’s end SDL will release interoperability versions of SDLX 2005 and Trados 7 to demonstrate integration progress. The still separate products will share Trados file filters, integrate Trados MultiTerm, add SDL’s Intermediate Translation Document (ITD) support to Trados 7, and improve TMX support between the two products. Besides sharing features, SDL will focus on improving round trips and the user interface. In late 2006 Mills plans a new desktop platform that he promises will be more than just taping the products together. While innovation tops his list, his main concern is that neither SDLX nor Trados customers lose any data in the transition.
  
  Our take: Preserving client investment will keep current customers happy, but too much caution could create a window for another software vendor to target SDL and Trados users with fear, uncertainty, doubt, and a strong TMX-based conversion story. This rival could go after the 90 percent of translators not using TM today – with no legacy shackles restraining its innovation.

• Terminology management. The November release of MultiTerm 7.1 will unify SDL’s TermBase and MultiTerm, with the Trados architecture winning out over the home team’s offering but with Mills saying that there is room for improvement in usability. Based on a relational database, the unified version will feature an improved online offering, deeper Unicode, better Word support, and searching of multiple termbases. To ease the transition for its own users, SDL may layer a TermBase interface over MultiTerm. Mills told us that he will swap out Trados’ search in favor of Phrase Finder – a component of SDL’s Knowledge-based Translation System (KbTS) – where possible.
  
  Our take: SDL took second in terminology management to Trados’ first place, both in market share and perceived feature-function. The greater power of the MultiTerm architecture combined with the KbTS Phrase Finder sounds like a good match. A TermBase skin on the MultiTerm engine plus a good migration story should keep SDL users happy.

• Translation memory server. One of Trados’ break-out-of-the-localization-box features was its server-based Language Server TM, a service that CMS vendors such as Documentum found unobtrusive compared to the widely overlapping functions of GXT. Later this year SDL will release a new version armed with a documented application programming interface (API) and standard internet connectivity. Mills also said that the company plans to introduce a software development kit (SDK) to complement the API, thus making it easier to integrate with database, content, and document management systems.
  
  Our take: Given the absence of an SDL TM server, no one lost any sleep over this decision. The addition of an API cum SDK means that third parties – LSPs, system integrators, software vendors, and customers themselves – could integrate Language Server into any application they could conceive.

• Translation workflow. GXT, Trados’ 1999-era legacy product from Uniscape, is entrenched at HP, Intel, and John Deere, but has not sold well due to heavy upfront development cost. Mills said that TeamWorks, a thick client/server product that Trados aimed at LSPs, is selling well as an upgrade to Language Server and desktop. Meanwhile, SDL has been pushing its own TMS into its language services accounts as a way of improving translation throughput.
  
  Our take: The product category formerly known as globalization management systems (GMS) poses the biggest problem for SDL, as the firm now owns three systems deemed strategic by their users. With each solution requiring much customization we believe that SDL will need to kill two of these products fairly quickly before they suck up too many development resources

• Machine or automated translation. The Enterprise Translation Server (ETS) has not been that visible in SDL’s marketing in the last year, although it is a key part of its Knowledge-based Translation System. Noting that he sees opportunity in improving SDL’s own translation memory via its MT engine, Mills said that the company does not compete actively with MT systems from IBM and Systran. SDL will focus on improving dictionaries, its rules base, and integration with the rest of its product line.
  
  Our take: With no MT in Trados’ product portfolio, which machine translation server to invest was another no-brainer decision for Mills. We think that computer-aided translation memory and automated machine translation will merge into a single offering in the coming few years, so Mills’ emphasis is well placed. However, competition from statistical MT solutions will marginalize SDL’s offering in the interim.

• Project management. SDL recently introduced Synergy, slimmed down workflow that it hopes to sell to smaller language service providers (LSP) for US$8-10,000, much less than it charges for GXT, TeamWorks, or TMS. Synergy is meant to help these shops track jobs, cost their efforts, and leverage translation memories. However, the initial release will not support Trados TMs.
  
  Our take: Many LSPs need solid, reproducible project management, but most tend to muddle through with manual processes or build their own solution based on Microsoft Access or Project. To win over this crowd, SDL will have to convince language service providers that it is on their side. It will also have to target a lower cost of total ownership by limiting the complexity of use with intelligent defaults, wizards, and straight-from-the-box integration with Office, CMS, translation memory tools including those from competitors, and financial packages such as QuickBooks or NetSuite.

• Content optimization. Earlier this year SDL rolled out AuthorAssistant, a tool for English-language authors that connects them to managed terminology bases and translation memory. Long term Mills says that SDL will add knowledge and rules about constructing sentences in other languages so that authors can “write for translation.”
  
  Our take: As we have noted over the last few years, text quantity and quality have spun out of control at most companies. Those who worry about the problem invest in optimizing the translation part of their global content life cycle, but they often ignore the opportunity to fundamentally improve the source. While AuthorAssistant tackles this problem at the front end of the content cycle, SDL will have an uphill battle selling this tool to content authors.

SDL comes to the table with a profitable services business and a big portfolio of technology. The company must move quickly to take advantage of its status as the only trans-Atlantic supplier of a full suite of language technology meant to support global information management. SDL must also lever its strong sales presence in North America and Europe against current rivals and upstarts (see Tables 1 and 2).

• Become the king of language middleware. When SDL rolled out its TMS 2005 offering earlier this year, only SDL-supplied client software could call its web services. Exposing SDL web services to any content authoring or translation tool could change the calculus of managing and transforming global content.

• Improve customer service. Prior to and following the acquisition, some SDL and Trados customers told us they feel the firms over-promised and under-delivered. Complaints include a lack of responsiveness, unchanged pricing, and apparent unwillingness to discuss issues now that SDL is dominant.

• Live up to its promises. Public statements from SDL emphasize the strong firewall between the company’s language services and tools divisions. It must work hard to convince doubting LSPs that SDL can be a friend, not only a rival.

• End the word-count wars. LSPs and buyers spend too many hours discussing word-counts and fuzzy matches. With its integration of the leading translation memories, SDL could set a word-count standard for all its tools.

• Integrate fast and furiously. SDL needs to avoid the integration chaos that often characterizes software buy-outs. It needs to be more like an Oracle or Symantec in rapidly assimilating what they acquire. This means making tough decisions like ending life support for GXT in favor of TMS product innovation.

• Change the business model. Most language workers don’t use linguistic technology, and those who do don’t use enough. By dropping the price of its desktop software SDL could empower more translators – and grow the market.

• Submerge the engine. SDL and other GMS suppliers have grandiose visions of globalizing all corporate content. A more realistic plan would be to create an embeddable TMS for use in CMS systems, thus masking its intricacies, simplifying integration and support, and establishing strong channel partners.

SDL’s urgent concern will be to convince existing users to give it a chance. But at the same time it must work hard to build a strategic role in a bigger market.

By Donald A. DePalma
October 2005

Copyright © 2005 by Common Sense Advisory, Inc. Chelmsford,  Massachusetts,  United States of America.

Published by:

Common Sense Advisory, Inc.
8 Minuteman Drive
Chelmsford, MA 01824-4646 USA
+1.866.510.6101 or +1.978.256.7621
info@commonsenseadvisory.com
www.commonsenseadvisory.com

No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning or otherwise, except as permitted under Section 107 or 108 of the 1976 United States Copyright Act, without the prior written permission of the Publisher.

Permission requests should be addressed to the Permissions Department, Common Sense Advisory, Inc., 8 Minuteman Drive, Chelmsford, MA 01824-4644, +1.978.256.7621, E-Mail: info@commonsenseadvisory.com. See www.commonsenseadvisory.com/others/citation_policy.php for usage guidelines.

Trademarks: Common Sense Advisory, Global DataSet, DataPoint, Globa Vista, Quick Take, and Technical Take are trademarks of Common Sense Advisory, Inc. All other trademarks are the property of their respective owners.

Information is based on the best available resources at the time of analysis. Opinions reflect the best judgment of Common Sense Advisory’s analysts at the time, and are subject to change.