Website translation services – why do you need them?

Today we have another Guest Post, this time from Thomas Chow. Enjoy!

Professional translation services for websites are an important part of maintaining search engine visibility. In addition to allowing potential customers and buyers who speak other languages access your site, good translation is important for ranking high in search engine results. For example, in Google for English-speakers, Korean or Hebrew pages will rank low. Having a translated version of your site available for each language market that you want to target will solve this problem. However, translation for search is about more than just language. Keyword usage and placement have to be correct as well.SEO Translation vs. Standard TranslationIf you are selecting a translation company to translate a document, your primary focus should be on the skill of the translator. You must ask if the translator has experience in both of the relevant languages, as well as seek references and perhaps examples of prior work. While these attributes are important for selecting a translator for search, they are not the only important considerations. For website translation, you must also seek a company that specializes in search engine optimization and understands the ramifications of different word choices on search engine visibility. A company with specialists in translation as well as specialists in SEO is ideal. Just because a translation is technically correct does not mean that it is a good match for valuable keywords or phrases.

Locating Translation Services

The Internet abounds with providers of translation services. To narrow down your choices, select a few websites that have successfully completed translation projects that are similar to your goals for your own website. If your site is in German and you want an English version available, select German sites that have high-ranking English versions of their pages. Another alternative is to hire a translator to complete the translation work, and then an SEO expert to optimize the translated page.  The SEO expert should work primarily in the language to which your site is translated. If you find this to be too expensive, choose a company that bundles translation and optimization services.

The Difference Between Keyword Translating and Website Translating

Website translation is necessary for ensuring that all of a site’s textual content is correctly translated into the second language. Keyword translation is more complicated than website translation, because the correct translation of one of your keywords may result in a word that does not rank well in the alternate-language search engine. Alternatively, you might find that if you have several keywords that are all synonyms for the same thing, after translation all of the keywords are identical. If you are working with English, you should be aware that not all languages have the number of synonyms for individual words as English, and you may need to use secondary SEO methods or different phrases to boost your site’s rank after translation.Why Not Use a Translation Tool?Translation tools may seem like a good cost-saving alternative to hiring a qualified and expert translator, but you may find your site suffering from unintended visibility consequences if you use one. Translation software often will not account for words with multiple meanings. Sometimes translation software makes very strange word choices that will be off-putting to readers of the translated page. If your translated text includes nonsense words, Google will likely penalize your site, since often nonsense words appear on spam pages. Most of all, if you use translation software, you will have no way of measuring the quality and accuracy of the finished product.

Author’s Bio: Thomas Chow is part of the operations team at Servio where he is in-charge of translating web content to help people scale their content strategy development

Website Translation – How Much Does It Cost?

Website translators have one very difficult task before even starting the work: quoting for it. And if the customer wants multinational SEO thrown in –even if it’s just on-page SEO- the job gets even more complicated. And that is even before the web page translation work is assigned!

The “traditional” translator has it pretty easy. He obtains the total word count, be it by the Word statistics, using a CAT tool or using some other word counting tool. I personally prefer CompleteWordCount from Shauna Kelly for this purpose. And, if the original is not in electronic format, then simply count the number of pages, count some representative samples of the text, and calcúlate the total. Apply your word rate, if necessary adjusting for customer or text difficulty and that’s it.

For a website it’s a little bit more complicated. First, usually you do not get the translation in a nice little package up front, so that you can count the words. Typically, you’re told that they want to translate site-so-and-do. Go, have a look and tell us how much it is. Yes, sure, especially if the site has a few hundred –or thousand– pages. Looking at them one by one? That would be madness. Count them? You must be kidding.

A little trick that I know some fellow website translators use is to use the “site:” search in a search engine, such as “site:www.seo-translator.com”. I get (as of today) about 152 results in Google. Is this reliable? To be honest, probably not. When I search in Yahoo I get 46 results, that is, less tan 1/3 or the Google results. Why? Because the search engines show up the pages that they have indexed, which may not be necessarily all site pages. And even so, some of the pages are dropped from the search results because they seem to be “too similar”.  So the result can be quite inaccurate.

I personally use a website downloader, and download the whole web site to local storage. This works pretty well, even with active server pages or php pages, as what is downloaded is what you see in your browser. You can run an analysis tool on those web pages and get a proper word count. Easy, will you say? Boy, are you wrong!

The very first problem is that those pages are *not* the original pages. Especially in the case of server-side scripting, a lot of code will have been executed that you will never see in the page itself. And even if you have nothing to translate in the code –which, believe me, you usually do– it’s something that you will see when you get the original pages, and review carefully so as not to mess with the code. That takes time, and effort, and you will not be counting that. Your price is too low.

And even so, not everything in the page can be translated. Forget a moment about tags, as any reasonably good CAT tool will filter those out. But what about things such as java code? You’ve counted the words, even if you’re not going to translate them. Your price is too high.

Oh, you thought you were going to use a CAT tool? You might have messed up with the cost if you made that assumption, my dear. What if the translation needs to be done on-line, using some kind of CRM? Your productivity will drop like a rock. It’s irrelevant if the text is repeated in every single page, you will still need to type it in for every single page. Worst of all, you may not see the context, and need to flip over to the original page to find out what it is all about. The two weeks that you forecasted for the whole website could easily become a month of painful translation. And you might also get penalized for not meeting the deadline.

And what about the links? You may only need to translate the anchor text, but believe me, it’s really difficult with many CAT tools. And that assuming that you’re not being requested to change the file names, for improved SEO. Remember, translated file names and liks provide you additional SEO juice. Even if you do not do that, you may still need to change the links because the original links point to pages in the original language, and you need to link to the translated pages. But changing the links across the whole site is pretty complicated. How many cross-links do you have? One hundred? One thousand? Or several tens of thousands? If you did not consider that, your price is far too low.

Keyword research? That is a pretty common request when somebody asks you to localize his website. Because if he knew the language so well that he could do the research himself then probably he would have it also translated himself. And yes, I forgot: for proper SEO, the keywords should be different on each page. More work, my dear. Higher price.

And yes, we have the translation itself. Easy, you say? Not if your customer wants to score well in the search engines. You need on-page SEO, stressing the keywords you want to score well for on each page, yet making the text flow properly. Anybody that has performed that kind of work in his native language knows how difficult that is. Why do then people think that it is any easier in a different language? Oops.

And of course, the checking. A professional SEO translator provides a “ready to use” localized website. And that means testing the java or PHP scripts, each individual link, etc. I usually put it up on my own server, for proper testing. Sometimes I’ve even had to develop some code or database to test it properly. Not quick. Not easy. Not cheap.

Every time that I make a quotation for a website translation I have to sweat through all this. Yes, I have developed some tools to help me out with this (good thing, being also a computer scientist), and even so I have to cross my fingers that I won’t be burned. Only to have the customer coming back and saying I am far too expensive. Sure.

A well-done website translation, depending on the page complexity, number of links, internal scripts, etc., may cost anywhere between 5 and 25 times what it would cost to simply translate the text. Sometimes even more. So it’s not cheap. It is the price for a properly working website and adequate ranking in the search engines. And if somebody thinks that he can cut off corners and not spend a significant percentage of what the original website has cost him, then he’s in for a bad awakening.

Keyword Translation – How Effective Is It For Your Multilingual Website?

How good is keyword translation for the localization of a website? Assuming you have a good translator. The response is simple: In the vast majority of the cases, plainly lousy.

Last week I had some international keyword research request for a customer. I got 45 keywords from the customer, and was requested to find the corresponding keywords in French and German. Now, while I do not usually offer the website translation into these two languages for quality reasons (though I have a degree in German & finished the French high school), keyword research is a different story. I performed the translation of the keywords, then just for double-checking I paid some native language colleagues to also translate the same keywords. No big surprises, my translation was correct, but they came up with 3 different keyword translations for German and 4 for French. Good, so I had some more. Unfortunately, 5 of the English keywords had the same translation in German, so I had to discard them.

So I had the plain translation of the keywords, and I went to see how suitable they were for the search engines. I checked first the German keyword translation. Disaster! Of the 40 translated keywords, not less than 22 (over 50%!) had absolutely NO searches. The rest amounted to some 255,000 searches/month, but unfortunately 99% of those searches were based on just 11 of the translated keyword. I attach below the table, obviously with the keywords removed so as not to impact my customer.

German keyword translation and associated search results.

Now, I was paid for keyword research, not for translation. My research found some 86 additional keywords, with a total local search traffic of 309,000 searches/month. Disregarding the low traffic keywords (as this customer is not a newcomer that needs to scrape the bottom of the barrel), I ended up recommending a total of 46 keywords for main targeting, with a total local traffic of 458,000 searches/month. Of these, just 24% were the direct translation of customer-provided keywords.

Are these keyword translation results an anomaly? Well, let’s take the second set, this time translated into French. Of the 44 translated keywords (some were repetitions, but we got alternative translations), fully 30 of them (68%!) had ZERO searches! Even worse, for the remaining 14 keywords there were a measly 40,000 searches/month, with 5 keywords taking 91% of those searches! I show the table below, again with the keywords in question removed.

Search results of direct French keyword translation

The results were truly amazing. Perhaps there was no market for the products of this company in France? Of course I had to do the keyword research to find out. I eventually found additional 147 keywords directly related to this market, with around 400,000 local searches/month. After removing the non-commercial keywords and low search traffic ones, I eventually recommended a total of 58 keywords, with a total local search volume of 351,000 searches/month. Of these, only 5 keywords with around 10% of the total traffic came from the translation of the original customer keywords.

Summarizing the used translated keywords:

Keyword translation search results

The usefulness (or lack of it) of the translated keywords becomes more evident in the following table:

Translated vs. used keywords

It becomes evident that the translation of keywords in the original language does not necessarily provide useful keywords in the target language. But it is more interesting to see the local search results for the different localized keywords:

Localized keyword search volume

Thus, the results of this exercise show that the direct translation of keywords is usually not providing the necessary SEO juice. This is consistent with my professional experience, but I am slightly surprised that the German translation actually achieved such high results.

Basically, what happens is that a translator is not necessarily knowledgeable in SEO, and may not be capable to identify high-search terms that one could target in his localized web pages. And to target barely 10% of the possible keywords is -to say the least- a waste of money if you are going to translate your website. Thus, you need someone that is capable not only of translating the website, but also who can do the necessary keyword research that will allow you to target the whole destination market and make your website translation a success.

Two Piracy Bills That Might Impact Your Website Content and Rankings

NASHVILLE, TN – If you’ve been on the Internet at all in the past few weeks, chances are you’ve seen posts or news articles about two piracy bills that are swarming around Congress: SOPA and PIPA. SOPA is the Stop Online Piracy Act, and PIPA is the Protect IP Act. While these bills may have good intentions – stopping copyright infringement – they could be detrimental to information sharing and the Internet as we know it today.

In a 24-hour protest on Jan. 18, many websites blacked out their content, including probably the biggest and most influential participant Wikipedia. Fight for the Future, whose tagline reads “a non-profit working to defend online freedom”, reports that more than 115,000 sites participated in the Jan. 18 blackout. Their numbers show the progression of Senators publicly opposing PIPA, which grew from one lone Senator on Nov. 16, 2011 to six on the morning of the protest. On the evening of the Jan. 18 protest, 34 Senators publicly opposed the anti-piracy bill.

You may have also heard that, for now, these bills have been postponed in Congress; they were scheduled for a Jan. 24 vote. It’s hard to tell if Congress has heard the pleas of many Americans to stop these censorship bills, or if the postponement is a way to shut us up temporarily. Will the bill slide back into Congress with revised text? Will they try to slip it by us?

Clay Shirky, in a very telling TED Talk (which I highly recommend watching), speaks about the history of copyright infringement legislation and its chilling future. He begins by explaining the Audio Home Recording Act of 1992, which basically said copying content such as songs from the radio to share with friends is okay (i.e, making personal mix tapes). It also said that taping high-quality versions, making multiple copies, and selling these copied tapes is not okay. The media companies wanted Congress to stop all copying, but Congress didn’t go that far. Now media companies, especially those who produce music and films, are spearheading these anti-piracy bills once again with SOPA and PIPA.

Shirky states: “The real effects of SOPA and PIPA are going to be different than the proposed effects. The threat, in fact, is this inversion of the burden of proof, where we suddenly are all treated like thieves at every moment we’re given the freedom to create, to produce, or to share. And the people who provide those capabilities to us – the YouTubes and Facebooks, the Twitters and TEDs – are in the business of having to police us or being on the hook for contributory infringement.” Some businesses may not want the job of policing their content, which could result in the businesses walking out, leaving us no place to voice our opinions or share our findings.

In the same talk, Shirky also enlightens listeners to an even deeper-rooted problem that SOPA and PIPA propose. He explains how the bills threaten our freedom and an inherent human right: “And because the biggest producers of content on the Internet are not Google and Yahoo, they’re us; we’re the people getting policed. Because in the end, the real threat to the enactment of PIPA and SOPA is our ability to share things with one another. So what PIPA and SOPA risk doing is taking a centuries old legal concept – ‘innocent until proven guilty’ – and reversing it –’guilty until proven innocent.’”

The real problem, as Shirky so eloquently states, is that SOPA and PIPA will make everyone a criminal. Without a proper trial, without gathering significant evidence, the government would be able to shut down your blog or website. The government could shut down your favorite websites. Forgoing the foundation of our basic rights – our Sixth Amendment right to a speedy, public trial with an impartial jury and our First Amendment right to free speech – the government will be able to censor our content seemingly without reason and without a proper check of power. Shirky goes one step further, deducing that media companies don’t want us to produce, but only to sit on our couches and consume. The producing is what threatens their livelihoods.

The hacker organization called Anonymous has shown their disdain for SOPA and PIPA by attacking major websites. Anonymous supposedly hacked the Department of Justice in retaliation to the shutdown of Megaupload.com, a major file-sharing website, and the arrest of its operators. They also hit the Motion Picture Association of America (MPAA), Universal Music, and the Recording Industry Association of America, according to a Forbes article. These organizations are supporters of SOPA and PIPA.

Facebook’s Washington, DC, office has a Facebook page where they discuss their views on these anti-piracy bills. While Facebook is sympathetic to the music and film industry on this copyright issue, they think SOPA and PIPA “are not the right solution to this problem, because of the collateral damage these overreaching bills would cause to the Internet.”

For now the bills are tabled, but Wikipedia writes, “SOPA and PIPA are not dead: they are waiting in the shadows.” This isn’t the end of the fight.

Master Google Ali Husayni has this to say about these two piracy bills and what they mean for website content sharing as we know it today:

“I’m all against piracy and hacking. But being raised in Iran, I’ve seen what censorship can do to freedom of the flow of information. You ‘censor’ only when you feel unable to verbally ‘engage’ and ‘encounter’ your opposition. Censorship by its very virtue puts you on the defense and the weaker position. And we don’t want to be portrayed as weak.”

Popularity: 2%

Website Translation – How Much Does It Cost?

Website translators have one very difficult task before even starting the work: quoting for it. And if the customer wants multinational SEO thrown in –even if it’s just on-page SEO- the job gets even more complicated. And that is even before the web page translation work is assigned!
The “traditional” translator has it pretty easy. He obtains the total word count, be it by the Word statistics, using a CAT tool or using some other word counting tool. I personally prefer CompleteWordCount from Shauna Kelly for this purpose. And, if the original is not in electronic format, then simply count the number of pages, count some representative samples of the text, and calcúlate the total. Apply your word rate, if necessary adjusting for customer or text difficulty and that’s it.
For a website it’s a little bit more complicated. First, usually you do not get the translation in a nice little package up front, so that you can count the words. Typically, you’re told that they want to translate site-so-and-do. Go, have a look and tell us how much it is. Yes, sure, especially if the site has a few hundred –or thousand– pages. Looking at them one by one? That would be madness. Count them? You must be kidding.
A little trick that I know some fellow website translators use is to use the “site:” search in a search engine, such as “site:www.seo-translator.com”. I get (as of today) about 152 results in Google. Is this reliable? To be honest, probably not. When I search in Yahoo I get 46 results, that is, less tan 1/3 or the Google results. Why? Because the search engines show up the pages that they have indexed, which may not be necessarily all site pages. And even so, some of the pages are dropped from the search results because they seem to be “too similar”.  So the result can be quite inaccurate.
I personally use a website downloader, and download the whole web site to local storage. This works pretty well, even with active server pages or php pages, as what is downloaded is what you see in your browser. You can run an analysis tool on those web pages and get a proper word count. Easy, will you say? Boy, are you wrong!
The very first problem is that those pages are *not* the original pages. Especially in the case of server-side scripting, a lot of code will have been executed that you will never see in the page itself. And even if you have nothing to translate in the code –which, believe me, you usually do– it’s something that you will see when you get the original pages, and review carefully so as not to mess with the code. That takes time, and effort, and you will not be counting that. Your price is too low.

Fuzzy Matches And Website Translation

I recently read an old article in the SpanishLegalTranslation blog about fuzzy matches. For those that do not know what this refers to, “Fuzzy matches” are those phrases recognized by a CAT (Computer-Aided Translation) tool as being similar to but not the same as a previously translated sentence that is stored in a translation memory. As Reed D. James points out in his post, fuzzy matches may greatly improve your productivity., specially in his field, which is legal translation.

Fuzzy Match

Now, I’ve used CAT tools for technical translations – a whole bunch of them. Fuzzy matches can indeed greatly help your performance when you translate documents. But when I finished this article I remembered that unfortunately this is not true for website translation.

Why is this? For one simple reason: You want the translated website to score well in the search engines, and therefore you will need to perform on-page SEO. This means therefore that each page needs to be different, and targeted for specific keywords. You cannot just reuse text from other pages, because you need to tweak it sufficiently to score in the search engines for a set of selected keywords. What is a benefit in traditional translation is a handicap in website translation – actually it goes so far that you could be penalized by the search engines for duplicate content!

This does not mean that you cannot use CAT tools for website translation – only that it is not a good idea, as later you will have to change the translated text later on, so as to ensure good ranking on the web pages. So I am always amazed when translation agencies ask we about “repetitions” and “fuzzy matches” for web pages. But then, most of the time they don’t have a clue, and believe that website translation is “just” another translation…

Where Web design tips and SEO begins