Why Larry Page killed Google Translate API and other assorted thoughts

Image representing Larry Page as depicted in C...

Image via CrunchBase

My first post on Google’s decision to shut down its immensely popular Translate API was the most retweeted and most highly read GTS Blog post ever. I should also mention that I started a discussion on Linkedin which received comments by some pretty important people in the MT industry like Alon Lavie, Manuel Heranz and Jeff Allen. It’s worth checking out.

Now that the dust is starting to settle, here are a few more thoughts on this topic.

  • Translate API wasn’t the only language API to be deprecated. Google depracated its entire Language API family.  This includes Transliterate, Virtual Keyboard and Diacrtitize.
  • The order to kibosh Translate API must have come from the top: Google’s CEO Larry Page. Since Page took over CEO duties from Eric Schmidt about one month ago, reports have it that he is much more focused on products and technology. I’m guessing that Larry did not feel that Translate API is part of that focus.
  • Google shut down Translate API because of the economic burden caused by extensive abuseHow heavy was the economic burden? Nobody can say for sure (and Google is keeping quiet about it), but I have a few theories. Consensus seems to be that Translate API was used for massive translation of websites for manipulation of SEO. If that is true, then the worldwide web became a much bigger place. Which means that Google needs to crawl and index many more pages than ever before. This not only means more storage capacity, but would also result in slower response times to search queries. Not to mention the number of servers Google needs to maintain to field all of those translation queries.
  • Extensive abuse. People were flagrantly abusing Google’s generosity and not even giving them credit for it. I mean companies did not put a ‘Powered by Google’ in their applications, did not link back to Google. But it was even worse. Companies were taking translations from Google for free and charging money for it. We have a term for that in Yiddish: it is called being a ‘chazer.’
  • Preempting the competition. It is no secret that Google is competing with Facebook in the online advertising space. Google Android is also competing with some other smartphones on the market. By providing a free Translate API, Google was allowing third-party developers to make translation applications for competitive products. It really makes no sense for Google if you think about it. Why give the competition an edge? By cancelling Translate API, Google can take control and keep the best translation applications for their own use.
  • The large LSPs ain’t dumb. SDL bought Language Weaver last year. Lionbridge entered into an agreement with IBM (Geofluent). These companies will restore the dominance in the MT segment to where it belonged in the first place. Leadership in translation technology should be maintained by translation companies, not by a search engine company.
  • Google Translate wasn’t the only free translation API. Apertium, Ta with you and Microsoft Translator offer free APIs. But if there is anything that Google’s shutdown of Translate API has taught us, it is not to build a product plan around a free API which may not be free or may even be gone in the near future.
  • Is Google Translate Toolkit API next? The GTT API is currently restricted, and Google says it has no plans to remove it. Does that sound familar?

That’s all for now. If you have anything to add, please leave a comment.

Related articles

Enhanced by Zemanta

9 thoughts on “Why Larry Page killed Google Translate API and other assorted thoughts

  1. Good summary. Not so sure that the processing of all this data is an issue for Google. The Google webmaster guidelines did warn about using Google Translate to populate large volumes of translated data (http://www.seroundtable.com/archives/022816.html), but this move seems to go further. There is far too much valuable data involved for it to be suddenly ‘turned off’ permanently as a service for developers, especially as there as indications that developers would pay for its use. Given the importance of translated, organized information in the world, I’d say we’ll see the return of GT as an service for developers in a different but paid form in the not too distant future.

  2. Let us not naive. Its not about greediness of users or abusers.

    Thousands of people are making living on Google API and products. This is exactly what makes Google so strong.

    And this is exactly what is going to make Google very weak once its competitors will start pushing even more into the developers market.

    I for one think that Lawrence Page just committed his first major mistake since he took over. Microsoft already took over Skype – and did not mind losing 4-5 milliards there. They have been very strong on the non major languages market during the last year. I suspect that they would be launching some surprises…. very soon. Watch and see :-)

  3. “…(don’t) build a product plan around a free API which may not be free or may even be gone in the near future.”

    Well said. It’s difficult to sympathize with developers who choose to build a business value offering based upon free and “beta” (Google’s favorite word) technology.

    I would disagree that “Leadership in translation technology should be maintained by translation companies.” In my opinion, leadership in any technology, process or innovation should be driven by any organization having sufficient foresight to position their value proposition in the direction where they see their market evolving.

    History is full of companies – and entire industries – that have ‘dropped the ball’ by bullishly assuming that a particular corner of the market is their own. Sony clung on to the concept of the cassette-based Walkman, while Apple developed the iPod and iTunes. Just five years ago no-one would have thought that the biggest distributor of music in North America would be a computer company.

    “Should” doesn’t make it so. The only way to prevent an “Apple” coming in and implementing disruptive technology that makes great swathes of the translation industry irrelevant is to implement those advances ourselves.

  4. Thanks Gee, your point re: leadership is well taken.

    On anothe topic, would you care to write a guest post on my blog? If so please contact me offline at davidg at gts-translation.com.

    Best, Dave

  5. Are we really that sad? The translations are horrible, sorry but they are. Outside of the core European languages the translations are bad, really bad. Sure its a wonderful tool for gist translation, and you can still do that by plugging in their widgets, but for what people seem to want to use it for they are awful.

    I’m not sure if data load is an issue. But since Google released an Afrikaans MT we’ve seen a spike in Afrikaans pages. Clearly people are dumping MT Afrikaans onto the web. The web as a valuable resource for Afrikaans text harvesting has dropped to zero. So I’m glad we’ll see no more free Afrikaans MT from Google.

  6. Inspirational read and very helpful. Your blog always helps me in my projects and thesis work. Regular reader of your blog. Thanks for the posting all the good reads.