International Search Engine Optimization – Fernando Maciá at the SEO Professional Congress

Written by Fernando Maciá

Although Miguel Pascual has already published on his Web site a summary of my presentation in the Professional SEO Congress 2012I wanted to complete the chronicle that I have made of this excellent Congress with the essential content of my talk on International Search Engine Optimization, which is an evolution of the one I gave at the last Walqa Virtual Shops Fair and whose presentation you can see here.

If you are interested in seeing the presentation in English, you can now view it titled Search engine optimization – Multilingual SEO on Slideshare.

This article shows working methodologies and recommendations supported by our experience in international SEO and consulting for website internationalization.

Why target international markets?

There are different reasons for seeking business in other countries:

• Traditional markets are shrinking.
• The Internet makes it possible to explore and open new markets in a low-cost and low-risk way.
• In an environment in which competitiveness is declining, IT fosters improvements in competitiveness.
• It is important to minimize risks in new business ventures, so the Internet makes it possible to contact partners globally to undertake joint ventures, maximize synergies and reduce risks.
• Traditional distribution, the usual channels are exhausted, so the risk of waiting for “times to change” is greater than the risk of venturing into new channels and new markets.

When attacking new markets from an SEO point of view, one of the main mistakes is to consider that, because English is a widely used international language, users will search in other countries in English.

No: users do most of their searches in their own languages and in the most widely used search engines in each country, which is not always Google.

On the other hand, we must take into account that if we have our Web site only in Spanish and English, we are trying to compete internationally in the language that has the highest level of competition, since Web sites oriented to an international scope are predominantly developed in the local language and, as a first option for foreign countries, in English. So in English, we compete not only with English content originally developed in English by English-speaking countries, but also with translations of Web sites from other countries around the world.

For example, the number of results for the search in English “apartment in Benidorm” is 3.2M. If we search in French, the results are close to 900,000. And if we do it in Norwegian, they do not reach 150,000. So if we have an opportunity to compete in markets that speak minority languages, it is advisable to develop our Web in those languages where it is easier to position ourselves well.

When considering an international presence of our Web, we must consider the peculiarities in the use of the language in each country. For example, in Brazil, mandarins are called by three different names depending on the region. Spanish is used differently in Spain and different Latin American countries and, for example, Japanese allows key words to be written using three different alphabets.

When selecting the search engines to target in each country, we should look for periodic reports published by consulting firms such as Forrester Research, ComScore, Experian or China Internet Watch on the market share of each search engine in each country.

Domain types

When working in an international environment, we can choose between generic Top Level Domains, country ccode Top Level Domains and IDN (Internationalized Domain Names).

Internationalized Domain Names – or IDNs – are Internet domain names that (potentially) contain non-ASCII characters. Such domains may contain characters with diacritical accents, as required in many European languages (including Spanish), or characters from non-Latin scripts such as Arabic and Chinese.
Punycode translates any non-ASCII character into standardized Unicode strings.

These are examples of this type of domains:

• http://www.españaescultura.es (Spanish): http://www.xn--espaaescultura-tnb.es/
• http://مثال.إختبار (Arabic): http://xn--mgbh0fb.xn--kgbechtv/
• http://παράδειγμα.δοκιμή (Greek): http://xn--hxajbheg2az3al.xn--jxalpdlp
• http://пример.испытание (Russian): http://xn--e1afmkfd.xn--80akhbyknj4f/

The use of IDN domains poses multiple problems when configuring tools such as Google Analytics, Webmaster Tools, etc. and in general it is not recommended to use them due to the difficulty of entering these special characters from computer keyboards that are not properly configured or do not have the necessary keys.

Key recommendations on domain types:

• If we believe that we have market potential in other countries or we will sell the domain later on, it is better to start with a TLD: .com.
• If we are clear about the country where we want to do business, a ccTLD (country code Top Level Domain) is better.
• We must, at least, register the .com and .es versions of any domain in which we are going to invest seriously in its promotion (pay attention to trademarks).

Subdomains or subdirectories?

To decide to develop the international versions of your Web sites in subdirectories or subdomainsthe following aspects, keep in mind that it depends on aspects such as the transmission of the popularity juice (best in subdirectories), amount of pages to be indexed (best in subdomains), need for geolocation in a given country (both subdomains and subdirectories are possible). Ask yourself the following questions:

• Does your traffic depend a lot on a single keyword?
• Can you generate popularity for the subdomain?
• What is the PR of your competitors? Your domains?
• Can you generate content on the subdomain?

If the answer is NO, a subdirectory is better:

• Because it will inherit the PR of the domain root first
• Because you can also geolocate from Google Webmaster Tools.

Key domain recommendations:

• Domains with a keyword rank well for that search
• Domains with more links from a given country rank better in that country.
• ccTLD domains are better positioned in the country concerned
• It is advisable to host the domain on a geolocated server in the country.

Key recommendations in the implementation of international domains:

• Buy the ccTLD domains of the countries where you think you may have business interests.
• Redirect with Redirect 301 all unused domains to your main domain.
• Develop sections or subdirectories with other versions of your Web.
• Develop microsites to test business potential in a country.
• Fully develops Web sites for countries with high business potential.

Geolocation of domains

Search engines determine the geographic orientation of a domain based on various signals:

• ccTLD domains: if we have a domain with a country extension, this is the default domain. For TLD domains, search engines look at other signals.
• Server IP: this is a factor that is lost for websites hosted on CDNs such as Akamai.
• Geolocation of inbound links.
• Country of the domain registrar’s address in Whois

Key recommendations for international domain hosting

• Host your ccTLD domains in each corresponding country.
• You will get better download times and better user experience.
• Generate inbound links and social media links from in-country domains to your domain content.

International versions of a Web site

We have different options:

Machine Translation

It’s better than nothing, but the content is created “on the fly” and is not indexable; the quality of the translation leaves much to be desired; we can set up an “event” in GA to track where the translator is being used to get clues as to what our next version should be.

Translation of content

Advantages:

• Indexable content: can be crawled by search engines
• Its quality can be very good

Disadvantages:

• It is a simple translation of content: perhaps the product argumentation, presentation, etc. should be different.

Content localization

• It is indexable.
• It is not a translation: it adapts the content to the idiosyncrasy of the audience, taking into account cultural, social, etc. references.

SEO-focused content localization

• All of the above advantages and, in addition, always include search terms that have been proven to have a high potential for generating quality search engine traffic in the right places (title, H1, anchor text, alt, etc.).

When using special characters in texts, we must take into account that Google treats searches with and without tilde (also with “ñ”) slightly differently:

• It works: include the incorrectly spelled word in the anchor text of the incoming link, title, hidden text layers with CSS, <noscript areas>, Alt tags (with and without link), URL (great because it is justified that there are no hyphens or accents.
• Does not work: include the incorrectly spelled word in meta tags keywords, dc, object tag, code comments.

Recommendations on inclusion of special characters to improve relevance

• It is neither professional nor advisable to include incorrect text in areas visible to the user, but other areas can be used: alt text, URLs, inbound links, <noscript>, etc. (with care).
• Avoid using zones that can trigger cloaking detection alarms: do not add many more keywords, just the possibilities of error.
• Be careful with the use of the wrong words in the anchor text: they will be visible on the pages containing the links.

Key recommendations in content localization

• Study the target audience profiles in each country and what are the sales arguments.
• Research the keywords used in that country for search categories related to the Web site.
• Adapt the texts to the target audiences of each country taking into account keyword research.
• Don’t ONLY translate content.

Multi-language and multi-country Web site structure

Languages separated by hidden variables for the search enginesuch as content customization by variables passed by post method, environment variables (language of the browser or the user’s operating system), IP of the page request or variables stored in cookies are very unfriendly to search engines and generally prevent the correct indexing of the different versions of the Web.

Languages separated by variables passed in URLs by the GET method: they are more user-friendly because there is a different URL for each version of the content. The disadvantages are that we do not have translated URLs and we cannot geolocate versions from Google Webmaster tools.

Separate languages in subdomains or subdirectories: this is the recommended option because it allows you to have semantic URLs in each language, geolocate content from Google Webmaster tools, etc.

In any case, it is very important to check the default content that a search engine can actually see.

We can simulate this using the text version of the search engine cacheYou can use the “show code” option in the W3C HTML code validator, browse with Firefox by disabling cookies, Javascript, stylesheets and images, or use the “crawl as Google” tool in Google Webmaster Tools.

It is also highly recommended to try to avoid the inclusion of the language selector in the header of the pages, since it is a section shared by all of them and the first one to be downloaded, can end up affecting the relevance of our website. (making the most relevant words precisely the names of languages).

The geolocation of a subdomain or subdirectory from the Google Webmaster tools prevents that when the user applies the filter “Webs of this country” to the SERPs, our pages are excluded from the results.

New international branding of alternative versions of international content

In this section, I made reference to the new international version markup recommended by Google and which is extensively developed in this post of mine, with the only exception that now Google also allows to include this markup directly in the sitemap files. In any case, the lines of code to be added do not represent an excessive additional weight, especially if GZip compression is used.

Key recommendations in language versions

• Create a subdirectory for each language version.
• If your positioning depends a lot on a keyword, generate a subdomain for each new language.
• Adjust the geographic orientation of each section of your Web to each country from Google Tools.
• Use the new markup for international Web sites.
• Achieve localized popularity.

Video optimization for international searches

Video optimization involves applying the usual recommendations regarding the wording of the title, description, etc. relevant. Additionally, we recommend generating video text transcript files with applications such as DivXLand Media Subtitler which allows you to generate and add subtitle files in multiple languages for a video. This text transcript functions as a kind of “alt text” for video content and allows the video to rank for searches related to this text on YouTube.

Additional recommendations for international SEO

• Configure the geographic targeting in Google Webmaster Tools: select in Settings > Geographic targeting the country to which we direct our domain, subdomain or subdirectory. If we have different language versions in subdirectories or subdomains, we add each one separately in Google Webmaster Tools and in Settings > Geographic targeting we set each subdomain or subdirectory to the country we want.
• SEMRush: in SEMRush we must also adjust the local version of Google consulted in order to compare regional visibility data.
• Google Insights: select the correct geographic area. In Google Insights we must filter by geographic area to correctly identify seasonal patterns and search trends in different countries. In countries located in different hemispheres, seasonal patterns are reversed.
• Google Trends: in Google Trends we must filter by geographic area to correctly identify seasonal patterns and search trends in different countries.
• Google Keyword Tool: in Google Keyword Tool we must also configure the language and country to obtain relevant results.
• Übersuggest also allows you to study keywords for different markets
• Links from the selector must link to the equivalent content in the selected language, not to the home page of the other version of the portal.

Web analytics in international portals

The implementation of a GA footprint for each version is simpler but involves multiple errors: unique visitors, referrers, bounce rate, average time of visit, page views, source…

Implementing a multi-domain GA footprint with appropriate filters generates more accurate results.

The presentation concluded with a mini-tutorial on Yandex and Baidu, which can be found in the presentation on international SEO mentioned at the beginning of the post.