Google Says Using Google Translate Can Be Against Google's Webmaster Guidelines

Aug 30, 2010 • 8:35 am | comments (15) by twitter Google+ | Filed Under Google Search Engine Optimization
 

I spotted an interesting comment from Google's JohnMu at the Google Webmaster Help forums. John said that in some cases, using Google Translate to auto-translate content for your web site might be seen as "creating auto-generated content, which would be against our Webmaster Guidelines."

I believe this is the first time I have seen a Googler say this, at least in response to using Google Translate to translate your content. Maybe John is saying this is a no-no when it comes to taking other people's content, running Google Translate on it and publishing it on your own site. I am not sure if he makes a distinction here. Let me quote JohnMu fully:

I just want to add a word of warning here -- using automated translation tools to directly create content for your site could be seen as creating auto-generated content, which would be against our Webmaster Guidelines. Instead of just taking the output of a program like Google Translate, I'd strongly recommend at least having it corrected before putting it online. While Googlebot may initially fall for some Spanish keywords in your text, your users are not going to appreciate content that has been automatically translated and published without a review. I love Google Translate, but if you publish the results and get them indexed without having them reviewed, you're not showing a lot of respect to your users...

You see, I am not sure if John is making a distinction between using it to translate content you wrote or not. It seems the point John is making here is that when you translate content using Google Translate, you must have someone manually look it over or else Google can penalize your site somehow because it is against Google's Webmaster Guidelines.

John is saying that using Google Translate, without manual intervention of some sorts (maybe even on your own unique content), is the same as creating auto-generated content, which is not in accordance with Google's Webmaster Guidelines.

Am I reading that right?

Forum discussion at Google Webmaster Help.

Previous story: Evolve Your Long Tail SEO Strategy: Exact Fails
 

Comments:

freako

08/30/2010 01:03 pm

the google webmaster guideline states Use robots.txt to prevent crawling of search results pages or other auto-generated pages that don't add much value for users coming from search engines. it's not per se against auto generated content (there is a shitload of that in the serps and sometimes is just the best result for the qurie there is) , it's against auto generated content that does not offer any additional value. so if you auto translate german wikipedia articles into english, and then offer this as content on your site, and it's the mjority of your content, don't be suprrised if google demotes these pages / your site. i ran into this issue some time ago, but just because we did not do our math correct and suddenly 95% of our content was translated gibbish. works for some time, but not long. also the bounce rate sucks. on a small scale, mixed with other more valuable content works great, and long time. it's all about the added value you can mix into this.

Seoidiot

08/30/2010 01:23 pm

Barry - try and translate English > german then back into English and you'll see why they think that Ahemmm - not that I'd abuse them that way as you know... Paul

methode

08/30/2010 01:32 pm

Look from this point, Barry: You have SERoundtable, install a plugin which works on top of the Google Translate API (for i.e. WordPress there are several such plugins) and suddenly all your pages are translated into dozens of languages, and each new page has its own URL. Say, this article, translated into Argentinian would be at http://www.seroundtable.com/AR/archives/022816.html You got [number_of_pages]*[number_of_languages] new crawlable and perhaps indexable page. Google Translate is good, it's awesome in fact, but has its faults and the translated content is sometimes plain gibberish, or in better cases slightly just hard to understand; users won't appreciate the junk content, as freako said your bounce rate on those pages will be awful, and Google won't appreciate these new, auto-translated pages either.

Barry Schwartz

08/30/2010 01:37 pm

Guys... I am not taking sides here. Just bringing a comment I found, that I never saw before, from a Googler. I totally agree that auto-translation can be bad for users. But it is funny coming from Google that it is Google Translate.

methode

08/30/2010 01:48 pm

Well :) >> Am I reading that right? Yes, you're reading it right :)

Michael Martinez

08/30/2010 04:05 pm

His motivations are irrelevant. John's advice should be universally accepted. When you use an autotranslation tool on your Web copy, you might as well be asking a blind monkey to type with one hand behind its back. I have never seen autotranslation produce effective Web copy. It often looks like gibberish but ALWAYS looks like the person who wrote the copy dropped out of Kindergarten. Let the user decide to use a translation tool.

Bryan Coe

08/30/2010 06:55 pm

The point John makes is good advice on many levels. Google looks for quality sites with quality copy. If you auto generate text from a translation program it is quite often not that reader friendly. Plus, Google just doesn't like auto generated anything. On the SEO front simple translating text and keywords is not a good way to optimize a site because a straight translation is often not what people will commonly use in the target language. The fact is machine translation still isn't good enough to replace humans. It might work for simple unimportant translations, but I highly recommend always having a proofreader.

Alex B.

08/30/2010 07:54 pm

Google appears to be in a quandary here. The entire aim of Google Translate (as stated by Google) is to vastly widen the market for SE/SEO services ("We're an advertising company, first and foremost!"). It now appears they've discovered that indiscriminate use of Google Translate is having a negative impact on search signal-to-noise and user experience. Notice the plug for Google Translator Toolkit as a fix for this issue... (is this, in fact, precisely the business case for GTT?) Either way, many megabytes of unreviewed translationese are currently being indexed and archived as we speak. I wonder how they'll get this demon back into the bottle...

Patrick

08/30/2010 08:47 pm

Well obviously google is going to penatly you if you use content created with automated translater programs! I am lucky enough to speak 3 languages fluently and everytime I have some idiot sending me auto-translated stuff, I can't stop laughing (or crying). Seriously, computers are made to solve mathematical problems, not linguistic problems. DO NOT USE ANY AUTOMATIC TRANSLATION PRGORAM FOR ANYTHING SLIGHTLY IMPORTANT! they just don't work.

Melinda

08/31/2010 05:38 am

Will Google penalize the site if we use google translated articles to promote it?

cls_wired

08/31/2010 09:37 am

I asked there: But if it eventually happened, how Google will react? Will that reflect website rankings and the quantity of search phrases website can be found by? Would there be any penalties from GoogleBot? And John answered me -)

cls_wired

08/31/2010 09:39 am

>Will Google penalize the site if we use google translated articles to promote it? I asked there, and John answered. Read the topic

Sean Oliver

08/31/2010 05:58 pm

He seemed vague about whether the Googlebot would penalize you for doing it... but he's unambiguous about the quality issues. If your objective is to actually communicate the meaning of content, rather than a word-for-word translation (which is often not the most accurate translation), then even post-editing machine translation (like GoogleTranslate) will give deliver sub-standard results. Unless you want your content to read like vacuum cleaner instructions.

Franjo Lagani

02/12/2011 02:14 pm

How can I turn off automatic translation of web pages?

SEO Translator

05/06/2011 03:49 pm

I think you're reading it correctly, that it might be understood as auto-generated content. On the other hand, I find it very interesting that he invites for manual review. I highlighted in a recent post (http://www.seo-translator.com/website-localization-impacted-by-new-google-algorithms/) that the new Panda algorithm uses spelling and grammar errors as "quality signals", and therefore pure machine translation (including that of Google Translate, mind you!) is likely to have you penalized.

blog comments powered by Disqus