Wednesday, July 28, 2010

The Coming Disintermediation and Disruption in the Translation Industry

I spent a few days in a very hot and dry Las Vegas last week at the IMTT Vendor Management Seminar, which I would wholeheartedly recommend to anybody who wants to understand how to better manage the relationship with translators to mutual benefit and also get a sense for changing production models in the industry. IMTT does a better job than anybody else I know of including translator perspectives into broader LSP and enterprise level localization discussions. They do this in a smaller and more engaged and interactive format than larger events.
There were many great presentations on the nuts and bolts of vendor management, but two high level industry presentations stood out for me. One by Arturo Quintero who shared many stories of how Moravia's supply chain works and evolved over time (" Supply Chain to Partner Management: Moravia’s Strategy"), often by learning from their customers and also provided much insight into how large enterprise buyers think about translation process and efficiency. The twitter stream with #imtt_vms from @ebodeaux @ebodeux does a great job of summarizing Arturo’s (and several other) presentations. Secondly, a relatively informal discussion session between Renato Beninatto and Bob Donaldson on “Disintermediation: How Power is Going Back to the Translators”. I review these two presentations together with my own thoughts on this subject in this entry.

So what is “Disintermediation”? According to Wikipedia:
In economics, disintermediation is the removal of intermediaries in a supply chain: "cutting out the middleman". Instead of going through traditional distribution channels, which had some type of intermediate (such as a distributor, wholesaler, broker, or agent), companies may now deal with every customer directly, for example via the Internet. One important factor is a drop in the cost of servicing customers directly.
The supply chain in the professional translation world looks something like this: Global Enterprise content creator > Internal localization department (Buyer) > Multilanguage Vendor (MLV) > Single Language Vendor (SLV) > Freelance Translator. Given the huge increases in global-customer-relevant content and the growing collaboration infrastructure many are rethinking the traditional supply chain. This approach made sense sense in the Web 1.0 and L10N 1.0 world but it is increasingly being challenged.
At the VMS Renato (RB) stated that he saw the possibility of SLVs being squeezed as MLVs go directly to translators, however Bob (BD) felt it was more likely that the MLVs would be marginalized as Buyers start going directly to SLVs. There was an interesting discussion on where the most valuable project management was being done with BD saying it was at the SLV and how the value of “project management” in localization was constantly falling as more efficient process and collaboration technology infrastructure reduced the need for PMs. There also seemed to be a sense that that the role of the Vendor Manager will increase in value relative to Project Management and that more VMs would be required and fewer PMs. They both commented on how the technology had reduced the need for “in-country” translators as it was now possible to easily find the best people in a decentralized model independent of physical location. However they both pointed out that collaboration models will still continue to require quality management and intelligent work assignment. RB pointed out that companies like SDL have very focused performance measurement systems at an individual level but BD countered that often what is measurable is not necessarily most valuable, so it is important to be aware of the bias that a ranking system can create. RB pointed that while there is change, multiple models of production will co-exist for a while.  Their discussion raises some interesting questions for us all:
  • Where is the greatest project management value added in the supply chain?
  • Does collaboration and process automation infrastructure reduce or even eliminate the need for project management? 
  • What are critical vendor management skills needed to build a high value supply chain?
  • What is the right criteria to use to measure vendor, employee quality and contribution?
  • Will domain expertise and competence with technology and collaboration platform become a differentiating factor for LSPs in future?
One point that RB made was that real innovation usually comes from outside the industry e.g. Microsoft snatched the PC market away from IBM (yes, asleep again), and Google did the same to Microsoft with the Search and Web Portal markets. Innovation can also be quite disruptive, often causing changes in the market eco-system and leadership. Late Update: Here is some of that conversation:

An example that I used some time ago that I think is quite applicable to translation, is the world that was in existence when word-processing technology first emerged. In the early 1980’s IBM sold lots of IBM Selectric typewriters and business documents were produced by pools of typists (freelancers?) who got content handed to them in voice or handwritten form by administrators (PMs?). This was a business that IBM dominated with 75%+ share. A company called Wang introduced a new and better way to do this with CRTs and "memory driven" typewriters. They and other similar innovators very quickly grabbed the market away from IBM and this market eventually evolved to PC based word-processors. Later we saw Wang miss the innovation opportunity as the technology evolved to PCs and become irrelevant. IBM was irrelevant in all these new market developments because IMO they never really focused on their customers real needs, or their thinking was confined by the existing revenue base. While typists did lose their jobs, the new technology created better jobs (new eco-systems) for the most skilled typists who became office managers /administrators, marketing coordinators, executive administrators and even DTP experts. I think MT and collaboration technology are similar drivers of structural change in professional translation today. The global enterprise is facing a data deluge and wants partners that will help them cope to make this rapidly multilingual. TEP production processes have a place and do work for static content, but are not useful for the dynamic and fast flowing content streams that global enterprises MUST deal with today.  Continuously flowing content streams demand translation production lines.

So how disruptive could disintermediation really get?
I paint some possibilities (complete speculation and nothing more) below:
  • Enterprise Content Creators could go directly to translators (both professional and amateur) via highly efficient collaboration and translation production infrastructure (which includes data driven MT technology) with a small role for internal localization teams to manage quality. Translation becomes as much a utility as word processing and Powerpoint production is today.Standards Rule!
  • New translation utilities emerge with highly automated translation and collaboration technology platforms (continuously learning MT, next generation TM, MT, huge portfolio of linguistic assets and super efficient editing tools, simple and powerful collaboration) and the ability to match buyers and sellers of translation services (freelancers and amateur SME who do post-editing) through efficient and effective collaboration platforms.
  • Enterprise Localization teams go directly to multiple SLVs through standardized and open collaboration infrastructure.
  • Enterprise Localization teams go directly to freelancers with some MLV quality management through standardized collaboration infrastructure (standardized infrastructure by definition excludes SDL).
  • Enterprise Buyers decide that vertical stacks like SDL/LIOX are the best solution and direct all translation activity through them.
  • Nothing changes because everything is great and it is not possible to improve the existing supply chain. We see the same people at all the same shows and we are all happy.
A tip of the hat to Ultan for reminding me again of the need for an alternate view on the building forces of disruption. Take a look, it is worth reading his blog and watching the video. Ethan Zuckerman wrote what I think is the seminal essay on the future of translation and “peer production” models. Yeeyan is another example of disruption: 150,000 volunteers who translate tens of millions of words on a regular and timely production schedule because they view what they do as high social value. The supply chain here is Content provider > Passionate organizer > 150,000 volunteer translators > Content consumer. 

So we see that change is already here, and more is coming, and it is wise to keep your eyes open and stay alert. Prepare, adapt, learn if possible, since as George Santayana said

“Those who do not learn from history are doomed to repeat it”

and I bet that at least a few of today’s market leaders will move into irrelevance or obscurity as other change laggards have before them. But yet, these are exciting times and evolution is a universal imperative.

Friday, July 16, 2010

Is SDL Getting Serious About MT? Does It Matter?

When I turned on my PC yesterday morning,  I was immediately greeted by a message joking, (I think), that I was fortunate not be working for SDL, "I bet you are desolate this morning - you could be working for SDL!" since I used to work for Language Weaver some time ago. :) In fact I was the VP Sales & Marketing for almost 3 years.

Before I knew it, Twitter and the blogosphere were abuzz with the “news” that SDL had acquired Language Weaver. The most interesting blog comments came from the politically correct
Common Sense Advisory, the skeptical Renato Beninatto and the surprised and slightly far out ruminations of Dave Grunwald. Of course, I could not leave out the succinct and ever insightful Ultan aka @localization who said: “Out of touch L10n news #1: Still digesting the industry impact of SDL acquiring Amptran in 1995, and now this LanguageBeaver thing.” (He of course offers this branding advice at no cost to SDL.)

There were also
comments from Forrester which state how important MT has become, but completely miss the ambivalence and uncertainty that many in the professional translation industry feel about any SDL owned technology and the SDL partner (sic) model in general. They seem to miss the fact  that it does not generally make sense to buy strategic and core production technology from a competitor.

Some of my favorite comments from these blogs (emphasis mine) starting with CSA:

On the reasons Tapling gave for selling to SDL: “a “sincere” offer to the market, focused on giving customers a real choice of a single vendor (doesn’t choice mean at least two options? and this sincerity is somewhat insincere ;-)) with both technology and service competencies;”  and
“Tapling said that Language Weaver… didn’t have to be acquired by SDL.” (Does that mean he was going to do a lot better on his own but decided just to be nice and hand over the reins? Also, there really is not much of a story to tell VCs is there? Hmmm maybe he did have to be acquired.) 
”we figure that the current venture capitalists, …. were like many investors whose patience beyond a five-year investment horizon was getting thin.;” (And most of them did make a respectable if not spectacular return on their investment.)
Renato had some zingers in his blog:
“I have been thanking SDL for the great job that they are doing at alienating their technology customers by providing sub-par customer service and support.” (Yes a highly developed “customer hostile” strategy that is a model for others who hate their customers.)
Language Weaver's main client is the US government and the main language pair is Arabic-English. In fact, the announcement points out that in 2009, the company had a loss of $1 million for revenues of $12.2 million.” ( Yes Mark, what's "good enough for government work" is often crap to the global enterprise. I wonder how many commercial market customers are using the Arabic system. Pray do tell.)
Dave asked some great questions in his blog:
”My first thought was “what, only 42 million? Why so little? Shouldn’t they be worth more? What does this mean for the MT industry as a whole?” (Yes unfortunately not a great valuation in the world at large, but pretty good for MT, especially MT that is not particularly competitive quality wise.Tapling should be commended for getting it that high. Guess who was the better negotiator? Tapling and a few others will walk away with a significant bundle of cash.)
Why is the CIA bailing on Language Weaver? (Was it ever their responsibility to keep them afloat? Now that they have Moses and access to the best Arabic TM around, do they really need LW? Moses is hard to use but they could find people that do it for them. They are probably painfully aware of the Google vs. LW quality differences and if they could get private access to Google MT quality guess what they would choose? The Feds like any rational users will choose quality if it is available though they do tend to like to buy American.)
Firstly, is this really a surprise? The writing was on the wall when they announced the “partnership”, and this was always one of the possibilities of the relationship, especially given LWs limited business success outside of the US government. They have never done well in markets where translation quality really mattered, hence their quiet exit from the localization and professional translation world a few years ago.

So does selling the company now make sense? I think so, from what I can see, this is probably a peak valuation event for them. As the MT market heats up, customers are doing much more comparison shopping and I am sure that many are realizing that there are several solutions in the market that produce better MT quality than Language Weaver. I have always believed that quality is the real driver of expansion of the overall MT market and that this happens most often with cooperative human linguist involvement.
(I felt strongly enough about this that I joined Asia Online, who also believed this). So, SDL does make sense for LW.  A company (LW) that decides to avoid regular interaction with professional translation companies is not likely to advance very far in terms of quality. As painful as it is, MT vendors need to talk to professional translators to understand how to make the technology better and more useful. This constant and ongoing feedback is essential to continuing improvements in MT quality. This is something that Language Weaver has consciously avoided for the last few years. Also, if after 9  years in business you get to $12M revenue (with a $1M loss) and are basically an Arabic to English translation solution for US government and G7 countries, maybe you should consider options other than doing more of the same. (I also know firsthand that the revenue expectations in 2008 when I was last there were SIGNIFICANTLY HIGHER than $12 million.) So we have a company that has been in a virtual revenue growth standstill for 3 years. In spite of doubling manpower and sales focused resources. Not to mention that it is has become increasingly clear that Google has better baseline systems,  Try it and see. Clearly 85% of the investors thought something along these lines.

Is this good for customers? In a word, NO. Remember that SDL is a company with three TMS products where data cannot be interchanged without some loss. Remember that Idiom used be the “Switzerland of TMS” which meant they were neutral and open and relatively transparent. It made their customers feel safe. Data could be moved around as needed. That day has come and gone and
it is now difficult for products not owned by SDL to get data in and out of any SDL software infrastructure. Does anybody believe this is going to get any easier with this new arrangement? The Idiom user experience is probably a good model for what might happen. It is not all bad, but not good either.

While customers who do want an all-SDL solution may not care, most of us live in world where non-SDL components exist and are necessary. Best-of-breed is much more acceptable nowadays, than one-stop-shopping. The demands of business translation in the future will require always-on, rapid response to streams of content that come from many different content creating products and Web 2.0 environments that are not owned by SDL. Openness, data interchange and transparency have
become increasingly important but SDL cannot even provide standard data interchange between their own products. So quite possibly we are heading into a bigger software roach motel where you check in but you never check out. At least not with your legs intact. Standards and easy data interchange are going to become critical and I think you should always be thinking how safe and portable your data is when you choose a vendor.

So is this all bad news? Not really, I think there is some possibility that they raise the profile of MT in general now that they have just spent $42M on it. This is about twice the amount they spent on Idiom. This does suggest that they think this is 2X as important (or that they seriously overpaid, which I think is the case here).  Possibly they could even get a couple of key content production components really tightly integrated, so the market begins to understand the value of integration. It is even remotely possible that SDL opens up and provides clear APIs to enable data to easily enter and exit SDL software environments for the benefit of their customers. (Umm probably not). MT technology can be a powerful tool in the hands of motivated and skillful users and perhaps SDL will learn to use this technology to maximum benefit. If they do, they could develop a long-term and enduring competitive cost advantage with their technology, especially in dynamic data environments. However, MT is complex and requires new kinds of thinking and many would say that SDL has not had a good track record with complex products or with out-of-the-box thinking.

Across, Globalsight, Andra and other TMS systems have benefited from the Idiom acquisition and it is quite likely that all the MT competition, especially Asia Online will benefit from this acquisition, as they are committed to a professional translation partnership model. The larger MLVs should be concerned about this latest SDL move, since they will face serious price pressure if SDL eventually figures out how to build functioning translation production lines and raise internal productivity on a regular and consistent basis. It would be prudent to explore your MT options now so that you are not marginalized or left out of many of the “new” kinds of translation work that will be coming in the near future.

I am also skeptical that this provides any leverage for SDL in the US government market. The spend on MT is miniscule compared to what the Feds spend on human translation and SDL has little or no share in the huge HT spend. So how does LW change this? Government procurement is a very specialized world that will require new partnerships, and we already know SDL does not do this well even if FOCI is not an issue. Also the government has already had unsatisfactory SDL experiences with Trados and Idiom and are unlikely to want to expand on this experience. Playing well others is an important skill here and the SDL lads just don't do that.

Some humorous asides:  The folks at the Examiner needs to re-examine themselves as they actually state that SDL/LW will be a threat to Google Translate.(Really !?! SNL Style)

Mark Lancaster maybe should have done his homework better and get his facts 
straight. He actually states that Franz Och was a founder of LW in this interview. Wrong. He was not, even though he had a short stint there and chose to go to Google instead of staying at LW for a tiny share of equity. (And Mark, just because somebody says I am best-of-breed it does not mean it's true. You do have to check this. They call it Due Diligence and they sometimes teach it at CEO school)

Here is a video discussion on the issue as well:

Also, I was asked by SDL PR representatives if eMpTy Pages
(apparently an influential industry blog) wanted to interview Mark & Mark. :-)  So what will happen next? Who knows really, it is too early to tell. I think SDL will find out that what I say above (especially about the fundamental quality issues) is true over the next few years. Regardless, this is one more sign that MT is strategic and it is worth your attention to figure out what your strategy is. Somebody is going to get MT right one of these days.

Wednesday, July 14, 2010

What is Holding the Wider Adoption of Machine Translation (MT) back?

I have been seriously distracted the last few weeks by the World Cup which is really the only sporting event I really ever connect to anymore. This year I watched more games (in HDTV too) than I have since my childhood and I am glad that the clearly better team won the final. My opinion on the real winners: Germany, Uruguay, Spain and Ghana, the biggest and possibly worst loser in World Cup history: The Netherlands. 

Anyway, there is an interesting discussion in the LinkedIn MT forum started by Lori Thicke that I thought was worth summarizing and highlighting. The original question she poses is: Why is MT adoption so slow, given the clear and real benefits that she has seen in her own experience with MT? I would recommend going through the whole discussion as I am only selectively paraphrasing some of the comments (IMO the best and most informative) and the comments are worth reading in their original state. This is my summary which is in no particular order.

The Reasons Behind Slow Adoption

  • Translator Resistance (could it be because many have direct experience with low quality over-hyped MT or insufficient training?).
  • Lack of Training on how and when to use MT.
  • People who think that MT is a replacement for human translators (especially those who expect the same quality as humans). Consumers need to understand that MT + Human is the construct necessary to get the best quality.
  • Too many eMpTY Promises from MT vendors who claim to have the silver bullet and deliver mud. 
  • There is a lot of low quality and erroneous advice (mostly opinion) that is outdated and obsolete on translator bulletin boards and forums that malign and present MT as a threat or a crappy technology.
  • Very few understand that customized MT engines are quite different from the free online engines at Google, Microsoft and Babelfish.
  • MT has been complex and expensive.
  • The poor quality of communication between MT vendors (many who feel offended by the quality criticism) and translators (who feel threatened by lower costs and bad quality and reduction of their value and role).  Marco Cevoli has provided a presentation on this conflict here from a slide deck presented at a ProZ conference.
Bob Donaldson points out the disconnect between the vendors and the clients in the translation production chain (Buyer > LSP > Translator) and points to three reasons for this disconnect:
  1. Different understanding and expectations of translation quality through the chain
  2. Lack of Translator (Vendor) Acceptance because of poor project management and skill mismatches in the editing and cleanup (post-editing) process.
  3. Inability to accept risk and manage it in a way to meet customer objectives, and without expecting the translators to bear the burden of poor quality MT systems. LSPs should be willing to share benefits as well as risk with translators as MT systems improve and ensure that they ACTUALLY raise productivity.
  • Translators may also not be the best post-editors and users will need to find how to find the right balance here. Sometimes less skilled people would be a much better fit for post-editing tasks. The world where MT is commonplace will have many more linguistic roles for people with linguistic skills.
  • I think that better definitions of deliverable MT linguistic quality would also go a long way to making the technology more widely accepted.
  • Many people do not realize that MT + Human Post-editing only makes sense when there is an actual measurable productivity advantage. Don’t use MT if this is not true and getting an MT engine to this state requires investment, effort and know-how.
Joachim Van den Bogaert also made some interesting points:
"Using MT and post-editing requires a paradigm shift. We see that, for now, translation jobs primarily come in project packages: the typical cycle is: engineering (pre-processing) -> translation -> engineering (post-processing) -> QA -> corrections -> QA -> final delivery. MT is at its most efficient in continuous projects: add content -> MT (with included automatic processing) -> post-editing -> delivery. Instead of focusing on delivering localized content at a specific point in time, MT allows you to consider having ALL your content localized ALL the time. Indeed, a well implemented MT chain, eliminates initialization costs, which allows you to have minor changes and additions to be translated instantly. But this requires a different mind set from (traditional) localization consumers."
  • Open source MT like Moses is difficult to use productively without a set of other tools (segmentation, alignment, corpus cleaning and analysis etc..).
  • MT systems have very different requirements and controls and thus it is difficult to integrate with production systems and have common and consistent pre and post processing.
  • In terms of mindset, I feel that MT should be considered as moving to a production line where some things can be built more efficiently as opposed to the craftsman approach that many have today.
Just because 90% of the professional translation industry makes disparaging remarks about MT on a regular basis does not make the technology irrelevant or unusable. Change often comes from the few, who solve many of the problems we have listed above and develop the new models for success. There are now a number of firms in the professional translation world who “get it” and they are likely to pave the way and build competitive advantage through informed and intelligent use of MT technology.

Alon Lavie points out that MT is the only way that new high-value dynamic content like customer support and user generated content (UGC) can be done, I believe this can be done at much higher quality levels when done in partnership with LSPs who are informed about the linguistic strategies to improve and refine MT engines. I have written about this type of content that presents a major opportunity for the professional translation industry. (I think so anyway). However, this will require new mindsets and new translation production strategies. We are in the early stages of this today.

So while many of these problems and issues remain unresolved, there are more and more initiatives that are attempting to address at least some of these problems. We are seeing much these days about how the translation production chain needs to evolve to handle the more dynamic and continuous flow of content that is increasingly valuable to translate. MT will be critical here.

It is clear that we should attempt to address all or as many of these problems listed above as possible but a few stand out for me.

  •  More education and more information about the current MT technology 
  • More information to help users assess where it makes sense and where it does not. 
  • Direct contact and hands on experience with a customized MT system. 
  • Better integration with the other translation production infrastructure. 
  • Development of win-win scenarios in MT + Post-editing work processes and compensation so that all participants share in the benefits of improved production capacity.
So while there are many difficulties there are also many new opportunities that provide new vistas for professional translation market leaders. And while it is clear that MT on it’s own is not the solution for every translation problem, it is also increasingly clear that it is a technology that will have a long-term role in making large volumes of high-value content multilingual at a cost and turnaround time that is viable and furthers business objectives.

And a parting note: for those of you think that MT is not a mainstream technology yet, consider how many people are using free online MT engines each and every day. I don’t have accurate data (please let me know if you do), but I estimate that anywhere from 10 –20 million people translate a web page, a document, a phrase or a word every single day in hundreds of language combinations. (About 2.5B to 5B words daily!) I think we can all safely assume that the thirst for information and knowledge in the growing human online population, has an energy and momentum that is far stronger than the inertia in the professional translation industry. It will drive this technology forward and large volumes of content will get translated at amazingly high quality levels. We are already seeing examples of this.

My advice to you: Wake up and smell the coffee (as they say in America) because often "good enough" is good enough. And I have to again use the 
quote that June Cohen, Executive Producer and Curator of the amazing TED Talks made at SXSW this year (even though it may seem naive and idealistic to some) when she was asked "What technology would you like invented? Or uninvented?"

"Instantaneous, accurate translation online. Nothing would do more to promote peace on this planet."