This document summarizes a presentation about website translation. It discusses why companies hesitate to translate websites, then provides reasons they should reconsider. It outlines best practices for website translation, including externalizing text and using translation memory. It also covers multilingual SEO practices like determining URL structures and using hreflang tags. International UX best practices discussed redirecting users to the same page in their preferred language and ensuring imagery is culturally appropriate.
Gutentag, Bonjour, Dzien Dobry, Hola, Konichwa, Ni-how. Thank you all for attending. We have some great stuff to talk about today regarding website translation, multilingual SEO and International UX. So, let’s get started with an introduction.
My name Is Robert J. Forloine. I have worked in localization for the last 10 years both in Europe and the US. I am now based in San Francisco working for Venga, which started as an in-house localization department at PeopleSoft and then Oracle. Sometime after the purchase we separated from Oracle and started Venga about 5 years ago. Enough about me and Venga, let’s go through what we will be talking about today.
One caveat: SEO strategy is changing all of the time, so just a forewarning to check yourselves for the most current information, but the majority if not everything I am presenting today should still be more than relevant.
Before we dive, please feel free to interrupt me at anytime, but let me at least finish the slide before you do. I prefer to have questions answered in context, so I want to extend that same courtesy to you. I will be using localization and translation fairly interchangeably. Localization is generally used in our industry to mean translation that requires some technical assistance from tools or specialists such as the case with websites.
Who in this room has translated their organization’s website or been involved in a translation project before? Ok, great.
So many organization and companies still don’t translate their web content, and I will be going over why they will eventually reconsider. Taking your website international with some translation help, can actually be a fairly smooth process with some good background knowledge about best practices, some insight on how you can prepare for localization and make the most out of your localized content. We are going to not only cover translation, but how you ensure that your website is found with proper multilingual SEO tips. Finally we will go over consideration to make to ensure that when people do visit your site in another language that their experience is not compromised.
Now we have all seen inaccurate translations. They can even provide some entertainment, but not if they are on your website. One reason organizations don’t translate is the fear of ending up with messaging that is going to negatively affect their brand or image in that target country. What are some of the reasons?
After talking to hundreds of potential and existing clients, I have come with a list of the most common reasons given for not translating a website. Let’s go through these with some response questions:
Have you given visitors of your site a means for providing feedback about what languages they would like to see available on your site or do you have a way of tracking this information?
Ask for references from a language service providers existing customers and if you have someone internally who can validate their translations this would help? Also, you will want to have some verification that the translators working on your content are familiar with the subject matter and that the same linguists and back up linguists are used consistently.
In order to have your website that uses a CMS translated, you need to provide the source files which tend to be in XML format. Great additional information to provide are an English glossary of key term to your business, all graphics with editable text boxes, a testing environment where linguists can review translations in context before they go live and maybe a few other elements. We will get into this topic in further detail later on in the presentation.
Quantifying ROI is one of the most difficult stats to calculate for an outsider of your business and industry. You have the means to start quantifying ROI against your English site based on however you calculate the success or ROI of your current site. It can be as simple as that and it can be certainly more complex than that depending on the ROI information that you would like to obtain.
Handling the translation of updates can be a very easy process if you integrate your Ingeniux CMS via APIs with a translation management system. This way you can automatically push and pull content for translations However, in general if you translate a file with a language service provider, they should be creating a translation memory that stores all of your existing translations. When you provide an updated file version of the same content you had translated before, your language translation service provider will leverage the translation memory or TM to reuse all of the initial translations and you will only pay for changes from the original.
You should translate into key languages for your business. Do you already have a strong number of visitors from specific countries or is there a blue ocean in certain markets where your competitors are not? If you are unsure, here is a great slide highlighting the internet’s top ten languages. There is another stat regarding if you translate into 5 languages you can cover a good percentage of the world and if you translate into 23 languages, you can cover 90% of the world according to CSA.
This graphic is already out of date especially with the growth of Chinese internet use being well over 500 million with a 30% growth between 2009 and 2011. If you are translating into Chinese you generally want to start with Simplified and then Traditional. They are both Mandarin with varied character sets. Cantonese is not typically a written language, occasionally in Hong Kong it is, so when people say Chinese they are referring to Mandarin.
For the translator’s never get it right thought, I would say ensure that your source content is not the problem, try to have an internal reviewer involved early on in validating a multilingual glossary and general translation style preferences and potentially help picking the main linguists.
Ensure you build a glossary and are involved in evaluating the style of the linguists.
If you don’t have anyone to validate the content, try to find someone even a client or user. Alternatively you can ask your language service provider to provide technical reviewers.
If you don’t have many internal resources to help you manage the process, set up an integration between your CMS and your language service provider’s translation management system. You can also ask your language service provider to provide more project management support and integrate them into your internal processes. We have done this for a few clients quite successfully.
In regard to no budget, let’s look at the average cost of translation on this next slide.
As you can see the cost of localization is probably less than your organization spends on janitorial services and bathroom products.
And you are not alone in being understaffed in supporting translation efforts which provides all the more reason to automate as much of the process as possible.
I will of course be sharing this presentation with anyone who wants a copy, so no need to write down every stat.
You can actually cover most of the online community worldwide with anywhere between 5-25 languages. With 5 languages you can cover 50% of the online community and with 25 you can cover 98% of the online population. CSA
So maybe the fashion was not to translate, but I feel that since the Panda update and the increase in good content having a positive effect on SEO, even smaller organizations are now implementing localization initiatives to increase their publically consumable content.
Even before the Panda update, there were plenty of researched and documented reasons to translate your web content into different languages for increase web success internationally.
In particular, I think that increasing the time visitors stay on your site is probably one of the more compelling arguments. 2x increase in that time is a fairly sizable positive shift.
The third stat related to searches is a fact that most people don’t necessarily think about in detail, which we will later on in this presentation. I will leave you with one idea in the meantime. Native language is also regional. For instance, in the US we might search for “weekend getaways” where as in the UK they would search for “weekend holiday” or “city-break”. So even in this example related to travel, how you configure your SEO for different markets can dramatically effect the online response in visitors and the success of your international web efforts.
Here are some additional stats worth perusing. I think that introspective questions are a great reflection in terms of evaluating how likely you would be to buy a product from a site where the entire site was in French or Japanese. Probably not very likely. The same holds true for other markets though there is a higher tolerance for English only sites than vice versa. A 4x increase in the probability of site visitors purchasing a good or service online if the site I localized in their language can dramatically shift revenue generation and site visitor engagement.
I had a client miss out on a multimillion dollar opportunity because they didn’t have their content translated into Turkish where one of their potential clients had a huge implementation.
There are 200 million more people online in China than there are people in the US. That is equivalent to the entire population of Brazil. Do you think that if you want to obtain more business from people outside of the US that Chinese would be a good language with which to start?
I think that by now I have convinced some of you about the benefits of translating your site into different languages. Now let’s go over some of the ways you can make the entire process easier and more effective.
Many of the points that I will be covering related to website localization best practices have to do with time savings, effort reduction and quality enhancement. String externalization has all of these benefits. It is much easier to localize externalized strings and HTML, because it is much easier to build filters for to focus only on the strings of text. In JavaScript for example, not everything in quotation marks is a string that you want to translate. For example a value stored in an object may be written as a string and it may not have effect on what the user sees, but may have several dependencies built into the code.
A similar example are images with text very common in web and marketing content. If you send over a bunch of un-editable images to your language service provider that are banners or just images you have in your site, they will either need to manually edit the image, send the text to a translator, and then paste the translations back into the image which creates another cost and increases the time required for the additional step of localizing the images. If you utilize text boxes not only can that information be read for SEO purposes, but you cut out the need for all of those manual image editing steps. This will in turn lower your project costs and decrease project turn around.
To increase consistency and quality in your translation efforts, you and your language service provider will need to build a glossary of key terms and a translation memory database. A language service provider can produce both, but I recommend you start now with the glossary. Have information about what should not be translated such a product or service offerings that are proprietary or spell out any acronyms that are specific to your business. Both a glossary and TM will ensure that approved translations are used consistently moving forward.
The translation memory will also enable you to reuse any translations you have completed before for future updates. The cost for any matches can be between 45-100% discount in cost for translating those matches.
Generally when we write we tend to write from a our cultural perspective using colloquialism and combinations of phrasal as well as modal verbs. The modal verbs can/could, may/might, must/should, will/would and phrasal verbs are for example, put up, put down, put up with, put in, put on, etc.
You don’t want translators to be native speakers of English. You want them to be a native speaker of the target language into which you would like to translate. Translators are proficient in English, but you want to ensure that they accurately and easily understand the source. Also, we had one client include many references to small towns in the US on their site that were only used as example locations.
Most people outside of the US don’t know where Murfreesboro Arkansas is and they wouldn’t know what the equivalent is in their country. In this case it is better to use well known international cities.
Venga can process a series of formats for translation such as HTML, XML, YAML and others. XML and HTML are probably the most common in content management systems and marketing automation. Venga can process native files with markup such as landing pages or emails with customized HTML filters that keep the HTML tags and attributes intact as well protected during localization. Therefore, there is no need to copy and paste text in and out of the CMS for localization. Venga has a variety of methods to pull HTML files including automated Curl requests.
Utilizing Ingeniux’ CMS alone is a great way to manage content, but if you plan on easily controlling the translation of your site and those translated versions Ingeniux has great features to help you. You will want to make sure that you have both WorldView and Translation Manager active. The reason being is you want to be able to easily connect your instance of Ingeniux with a translation management system that your language service provider has. That way you don’t have to copy and past content, or download and upload files, it is just a few simple clicks to push and pull content for translation which is completely under your control.
If you do set up an integration, make sure that if you modify your XML structure at all from what is provided by Ingeniux that you track all the changes that you have made. You will want to provide this to your language service provider so that they can modify their filters to automatically process your XML files for translation to ensure that they translate everything that needs to be translated without having to manually modify the files.
The automated messaging adds significant value to international communication with linguists, especially when there are substantial time zone differences such as Tokyo, Japan, where there is a +16 hour time difference with the US West Coast. In many cases, only minimal human intervention by the PM team is required.
Anything that you will want to translate in Ingeniux once you are set up with the integration, you will want to have set to the status of “Ready for translation”. This makes finding content for translation and pushing it out for translation very easy.
There is one thing that many of you who have not localized your site may be doing which is letting page properties populate the navigation name in the SEO group which create the title of the page. You will want to populate this field so that when you push your site out for translation that the title is also included in the file for translation. Otherwise, it will require manual work on your side to translate the title.
Here I have a few integration guides that I can give out outlining the steps needed to setup and integration. Also Domingo, the self proclaimed and self actualized social butterfly of Ingeniux and I will have our session tomorrow at 10am where will cover more about the integration process, configuration and workflow.
There is huge SEO opportunity to drive additional traffic to your site by implementing good multilingual SEO for each language. It’s essentially having another version of your site and additional traffic generation opportunities. However, it is important to note that SEO is not just straight translation. I gave the example earlier of people searching for “weekend getaways” in the US where as they might search for “weekend holiday” or “city-break” in the UK. If you used Google translate for “weekend holiday” you would get back “week-end de vacance” when actually "location vacance” gets more than 4 million hits according to search engine watch. We all know how important having the right keywords in our content title and description are to affecting ranking even if it has been underplayed lately. This is how we find things through search terms.
When you do decide to go after other markets, consider that they may use a different search engine than Google. I have a nice image on the next slide of where different search engines are often used. China’s Baidu for instance handles geotargeting and search ranking in a way that differently than Google does now, but may be similar in some aspects to the way Google use to handle ranking. You have Yandex in Russia and Naver in South Korea which expanded to Japan in 2009.
One consideration when you do have a multilingual site that can have an effect on your SEO is how do you handle your site structure to make those multilingual pages easily found by global search engines as well as the User and manageable by you. My recommendation is to use a generic top level domain or gTLD like a .com and then have subdirectories such as www.sharknado.com/es-es for your Spanish from Spain site. This is probably the easiest to manage and only requires you to validate the main domain in Google Webmaster tools once and not every single variation as is the case when use subdomains.
Here are some examples of the different multilingual site structures for your reference. So the one that I am recommending is the top one, the generic top level domain with a sub directory.
When you determine what your site structure for multilingual sites is going to be you want to make sure that you also tell other search engines. On this page we have the “x-default” value that should be used to tell other search engines, use this site when I don’t have the language the user wants. You will want to define in Google Webmaster tools what are your different URLs for different languages and regions for proper Geotargeting, but you will also want to list it in your HTML. The way that you do this is by using the hreflang attribute which enables you to use language code followed by dash and then the country code. If you ever have duplicate versions of a page on your site in the same language for the same region you will also want to tell search engine bots which is the copy by adding the canonical value to the rel attribute.
“The new x-default hreflang attribute value signals to our algorithms that this page doesn’t target any specific language or locale and is the default page when no other page is better suited. For example, it would be the page our algorithms try to show French-speaking searchers worldwide or English-speaking searchers on google.ca.” - Google
I am going to speak briefly about this slide since it doesn’t relate specifically to website translation, multilingual SEO and international UX. Video has extremely high seo effectiveness and user engagement.
Now we are not going to get into all of the best practices associated with initiating an intensive global UX analysis and implementation. Otherwise we would have to talk about Hofstede's cultural dimensions theory and talk about how to appropriately setup in-country market research and usability tests. However, we will talk about some basics. You should offer a global version of your site that is very readable and neutralized for an international audience. Many of our clients use this version as the localization starting point.
You will want to detect a user’s preferred language. I recommend evaluating their browser language selection as opposed to their IP. Imagine you travel with your laptop internationally, IP detection in this case would fail where browser language detection would not. Even after detecting you should confirm the user’s language preference and then store that information.
When doing redirects 301 communicate that it is a permanent redirect where as 302 is meant for temporary redirects. When redirecting for instance if you offer a way for users to select the language in the top right corner, which I recommend then, redirect the user to the same page they are not the home page for the whole site. You don’t want to force the user to expend extra energy getting back to where they wanted to be in the first place.
Depending on how colloquial your content is you may want to use international copywriters. We have done this for a few clients and it has provided some very nice SEO benefits as well.
You will also want to make sure that your imagery is culturally appropriate or neutral. For instance a thumbs up or a picture of people cheering in the office may have little cultural relevance or connection for people in different parts of the world.
We have covered the first point to some degree in terms of not sending people back to the home page when they choose their language.
If you have successfully implemented off page linking to improve search ranking, you will need to do the same in each target country.
Effective website localization does require some local knowledge depending on your line of business. For example, Russia is very much a cash on delivery country and Israelis prefer talking to someone to complete a purchase but they will perform research online beforehand when it comes to travel.