International and multilingual SEO is arguably one of the more challenging disciplines within the SEO remit.
While there are decisions to make over which languages and countries to target, taking into account wider business factors, there are technical decisions that need to be made – one of which being the international setup of your URL structures.
Anatomy of URLs
Throughout this article, I’m going to refer to URLs (Uniform Resource Locator), URIs (Uniform Resource Identifiers), and URNs (Uniform Resource Name). The difference can be explained by the diagram below:
- URL can only be used to describe the complete (or absolute), containing the scheme, host, and path.
- URN can only be described to describe the path.
- Everything is a URI.
In short, all URNs and URLs are URIs, but not all URIs are URLs.
Multilingual URL Structures
There are five primary ways in which a website can be set up in terms of URL structures for international SEO. These being:
- gTLD with subfolders
- gTLD with subdomains
- Different ccTLDs for the root domain
- gTLD with parameters
- Different domains entirely
While there are a number of articles exploring the pros and cons of each implementation method, some of which I will go into later, there are also the limitations of your technology stack.
This means you’re likely going to need to use different ccTLDs or domains while injecting hreflang through Shopify’s theme liquid.
The Merits of Each Approach
You can be successful with any of the five aforementioned implementation methods, as there are a large number of variables that go into the success of a multilingual SEO campaign – this being just one of them.
These send the strongest signals to search engines about which country is being targeted. For example, domain.fr is a strong indication for France.
This method is often more expensive to maintain (as it’s maintaining multiple domains) and each one will have its own backlink profile to build and maintain.
Over the years, some ccTLDs (such as .co and .ax) have been used for non-international purposes as the ending of words.
For instance, bron.co and beesw.ax may look clever from a marketing and brand perspective, but in reality, they are the ccTLDs for Colombia and the Åland Islands.
Subdomain & Subdirectory
Without wanting to reignite an SEO debate as old as time, using a gTLD with a subdomain or a subdirectory route are both viable implementations for international SEO. Both have been implemented “in the wild” with good success.
From experience, I’ve had success with both implementation methods – but here’s a more comprehensive analysis of the subdomain versus subfolder debate by Jenny Halasz.
From a development perspective, you can map other platforms and databases to a subdomain, or reverse proxy them to a subfolder.
This again depends on the limitations of your technology stack, and both can be achieved through edge SEO and user workers to handle either the mapping, or reverse proxy.
This is where the domain is appended with a ?lang=fr URL parameter or similar.
While this can work, this is the implementation method I recommend the least as it is not user-friendly to link to.
Different Domain Entirely
Similar to the different ccTLD approach, using different domains can be a fruitful international SEO strategy especially if your domain is not your brand name.
A good example of a business that would benefit from this approach is British retail brand B&Q which uses the domain name diy.com (as they are a DIY retail business).
DIY (Do It Yourself) may not translate as directly in other languages, so this would be a candidate for using a different domain altogether.
URL Best Practices for Certain Countries
When deciding on your initial URL structure, it’s important to also thoroughly research and take into account cultural differences and nuances not only of non-Google search engines, but also different user bases.
URL Best Practices for Japan
There is no causational proof that having a Japan-specific ccTLD (such as .co.jp) gives any advantage over ranking within Google Japan, goo.ne.jp, or Yahoo.
When putting together your URLs for Japan, and to help with localization, you should look to use Japanese characters within your URNs.
For example, if you’re a U.S. travel company targeting the Japanese market with hotels to book, your URLs might look something like this:
旅行 being the word ryokou in Kanji. Japan uses multiple alphabets including Kanji, Hiragana, and Katakana.
Users in Japan typically type in Latin before converting their search query into their preferred writing style. Hiragana is the most common alphabet used in Japan, while Katakana is often reserved for foreign words.
As long as your URLs are UTF-8 encoded, Japanese characters will display in SERPs if formatted correctly.
URL Best Practices for Russia & Yandex
Unlike Google, Yandex values keywords being included in the path – although it’s important that they’re not used in a “spammy manner”.
Another common question for Yandex is the use of Cyrillic in URLs when targeting Russia, Google’s official byline on the subject is that:
“[W]e (Google) can crawl and index non-latin URLs normally.”
A good example of this in the real-world is the Russian version of Wikipedia, who use Cyrillic paths:
However, when you copy and paste the URL outside of the browser, the Cyrillic doesn’t format correctly and instead you get a broken path:
But when you paste it back into a browser, such as Chrome or YaBrowser, it works and formats.
URL Best Practices for China & Baidu
Baidu operates differently to Google and has a number of differences to Google. How it processes URLs and what Baidu defines as “friendly” URLs also differ slightly to Googles.
With Baidu, you should avoid using URLs with parameters at all costs as Baidu can have issues in crawling them. Similar to Google, best practice would be to use a descriptive URN.
Unlike Japan, it’s also advised to not use Chinese characters within URL. This is because some Chinese characters have issues being encoded, which can lead to unreadable characters in a URI path.
This will cause issues for Baidu-Spider crawling and discovering the content on the URIs. However, you can replace the Chinese characters with Pinyin.
All screenshots taken by author, March 2019