One of the major benefits in running a multilingual site is appearing high on Google search results in different languages. The process of promoting your site to the top is called SEO and when it’s happening on a multilingual site, we call it “multilingual SEO”.
Just to get a few things out of the way before we get started. There is no magic here, no tricks and no shortcuts. We can’t pay our way to the top and we can’t take shortcuts. We can follow good practices and use good tools to achieve our goals and get visitors from all over the world to visit our site and buy our things.
In this tutorial, we’ll review the objectives, tools and best practices for multilingual SEO.
Imagine for a second that you’re the all mighty Google. You get billions of people looking for different things 24hs per day. You have to pick 10 sites from the entire Internet and show them on the first page for every search.
There are hundreds of millions of web pages for you to pick just 10 items for and you need to do that in a split second.
When we think about it this way, we understand that everything is automated and there’s nobody to talk with. We cannot pay to have our site appear higher. So, any SEO or web development first that guarantees #1 place on Google search results is selling you nothing but air. At best.
So, back to Google. If you were Google, how would you tell what 10 pages to show for a search results? The answer is surprisingly simple and also explains to us what we need to do in our SEO work.
- Look for pages that talk about the subject of the search – obviously, the page must focus on the specific subject. If we are looking for ‘cat food’, a good page will include a wealth of information about ‘cat food’. If it talked about ‘pet food’ and briefly mentions ‘cat food’, that page is a lot less interesting to Google than a page that talks specifically about ‘cat food’.
- Look for pages that belong to websites that talk about the topic – if you have a page about ‘cat food’, but it’s part of a site that talks about ‘auto parts’, Google will assume that you are not the world’s biggest authority on cat food. However, if the page belongs to a site that talks about pet food and mentions cat food in a dozen other places, linking to that ‘cat food’ page, Google knows that this site may be relevant for ‘cat food’ searches.
- Look for pages that other sites consider as authority – this is the tricky part in all this game, but it’s also the one that comes very naturally, in time. Google cannot go and interview a billion people, asking their opinion about hundreds of millions of websites. But it can ‘interview’ other websites and see which sites they are referring to. When Google sees a link from one site to the other, it considers this link as a vote of trust. It assumes that placing a link means you recommend people to read quality information in another site. Of course, many people try to exploit this by buying links or creating misleading content on their other sites, but Google’s algorithms are pretty smart and are not easily fooled.
So, if this is what Google does to find those best 10 pages to show as search results, what should we do to be one of those pages?
- Produce high quality content.
- Stay focused. The narrower, the better.
- Promote our content to others, so that they know about it and will link to it, if they truly appreciate it.
Is this simple, or what?
Now that we know what we need to do, let’s talk about how we do it. Let’s agree that we are writing great content. Our purpose when we talk about SEO is to make sure that Google also understands it.
When Google reads our site, it looks at it a little differently than we see the site in a browser.
Google doesn’t care about fancy design, fonts, colours, animation and images. When Google reads your site, all it sees is the content and the HTML that goes around the text.
Here is our homepage, as we see it in a browser:
And here is the same page, as Google sees it:
Google doesn’t get impressed by the grey background or the nice gradient that we worked so hard on. What Google does see how we use standard HTML tags to emphasize important parts of the page.
Do you read newspapers? Imagine yourself skimming through a long weekend newspaper, looking for interesting stuff. Here is what I do:
- Flip the pages and read through the titles
- When I find a nice title, I take another few seconds to read the subtitle.
- If the subtitle is also interesting (I found what I’m looking for), I would read the entire content.
Google assumes that we all do the same when we read web pages. The most important part of any web page is its title (the <title> tag). Then, there are the headings, starting with the <h1> heading and going down in importance as we go through H2, H3 and on.
So, you should always use short, descriptive and clear texts inside the <title> tag of every page. Never ignore that. Never have empty title tags or meaningless titles. Never have the site name appear in all title tags.
In our ‘cat food’ example page, a good title tag would be <title>Cat Food</title>. Of course, that’s a really broad title and the chances of your appearing in the top 10 results would be slim. If you provide organic cat food in San Antonio, there’s a much better way to appear in the top ten, or possibly in the first position, if your page’s title tag would be <title>Organic Cat Food in San Antonio</title>. Now, you’re competing only with others who sell into a very narrow niche, instead of everyone in the world who has anything to say about cat food. Makes sense?
After you’ve written good title tags and good header tags, you can get to the content. If you’re not sure how to do this, think about yourself as a visitor, looking for your things. Don’t think about what you want to say but about what they want to learn. Write for them and answer the questions that they are most likely interested in.
When you use WordPress to build your site, a lot of the SEO work is already done for you. WordPress produces simple HTML, which Google easily understands. Many themes come with fields for SEO. If your theme doesn’t include that, you can always use a plugin that will add that information to your pages.
Like many other optimizations, you will get the biggest benefits by doing the simplest things. Of course, there’s a whole depth of other activities that you can do to tweak and improve, but if you are just starting, you will discover that by adding the basic SEO information, you get the biggest jump in results.
We use and recommend an SEO plugin called WordPress SEO by Yoast. It’s simple to understand and use, does its job and works great with WPML for multilingual sites. It makes the basic features easy to access and also includes the more complex tools, for when you need them.
To learn how to translate everything that the WordPress SEO plugin creates, you should read our tutorial on using WordPress SEO with WPML.
The WordPress SEO box starts with a preview of how your page will look in Google search results if it showed there. The plugin cannot guarantee that your content will show up on Google, but it can help you achieve it yourself.
The next part is the ‘focus keyword‘. This is where you can check if your target keywords appear in important parts of your page. If it says ‘yes’ for everything, you’re in excellent shape. Otherwise, you know where you need to edit.
Next come the parts that you actually control, so we’ll expand on them a bit more.
The SEO Title is, by far, the most valuable text in your page for SEO purposes. This is the title of the page, as visitors and search engines see it. We already talked before about good title tags. The WordPress SEO plugin lets you easily control page titles by entering them into that box, per page.
The Meta Description is what you suggest for search engines to show as the page summary, in search results. Keep in mind that it’s only a suggestion. In most cases, Google will accept your suggestion and display that sentence if Google thinks that it’s close to the content of the page. If your page is about plastic coffee cups and the page title is about organic cat food, you can forget about Google ever displaying it. In fact, you can count on Google to disregard your entire page, or site, considering you to be a spammer.
Good meta description will provide a short and clear summary of the content of the page.
Finally, we get to the good stuff and we can explain about achieving good SEO for multilingual sites. Let’s start with a few basic facts and then see what you need to do:
- Google needs to understand what your site is about
- Google needs to understand the structure of your site
- Google needs to understand what each page is about
- Google wants to know the language of every page in your site
- Google wants to know what’s translation of what
Since WPML keeps different languages in different pages, it makes it easy for Google to detect the language. You can help that too, by making sure that everything in your site is translated. This should include the theme and any texts that other plugins output.
Google likes to know how translations are linked together in your site and WPML provides this information.
Scroll down the page source and you will see something like:
<link rel="alternate" hreflang="es-ES" href="http://wpml.org/es/documentation-2/soporte/" />
<link rel="alternate" hreflang="de-DE" href="http://wpml.org/de/documentation-3/support-2/" />
<link rel="alternate" hreflang="fr-FR" href="http://wpml.org/fr/documentation-6/informations-pour-les-developpeurs/" />
<link rel="alternate" hreflang="pt-BR" href="http://wpml.org/pt-br/documentacao/informacoes-para-desenvolvedores/" />
These are the hreflang links that WPML generates (see the complete explanation of hreflang tags for multilingual sites) Your content has a lot of links going to different pages and off-site. This list of links tells Google what translations exist for any page. With that list, Google can distinguish between ‘regular’ links to other content and links to translation.
If you were Google, you should be very happy to receive this information. Now, you know how what’s translation of what. When WPML outputs this information in a consistent manner for the entire site, Google has a complete picture of the entire structure of the site, languages included.
From this moment, Google doesn’t need to guess anything. It knows the language every page is written in and how translations are connected together. Google will quickly understand that it’s looking at one website with translated content in several languages.
Google offers free machine translation. It’s often accurate and a fantastic tool for understanding texts in other languages. The only thing you are never supposed to do with it is publish it, as your own content, on your websites.
Why? Because this throws Google off and it pisses their algorithm. Let me explain the logic.
Google indexes billions of web pages to get its knowledge. Part of that knowledge is how to translate texts. One of the things that Google likes best is multilingual websites, translated by real people (preferably professional translators). With these multilingual websites, Google creates a huge dictionary of how to translate.
It looks at the same content in different languages, breaks it into paragraphs and sentences, matches between them and uses that information as reference for its machine translation engine.
This means that what you translate yourself has some tiny impact on what others get when they use machine translation. In effect, when we translate, we all contribute a little bit of knowledge to the great global pool of ‘correct translation’ reference.
So, what happens when you inject machine translation into this? You skew the reference and make this entire system collapse. At least you would, if Google allowed it.
Google cannot use its own machine translation as reference for what’s ‘correct translation’. It must get real human translation for that. So, when you take the machine translation from Google and present it as your content, the immediate result is the Google drops that content from the search index. If you do it in just a few places, these pages will drop alone. If Google sees it in all your site, the entire site will drop. After that, good luck getting indexed again.
Google’s SEO team explained this process and the reasoning in different occasions. If you’re brave enough and want proof, go ahead and experiment yourself 🙂
Please note that it’s completely legitimate to offer machine translation on your sites. The way to do it is by adding a Google Translation widget. This passes your content via Google’s translation system, but doesn’t tell Google that it’s your translation. It’s legitimate and encouraged by Google. If you don’t have the need or resources to build true multilingual sites, you can also offer free machine translation, done directly by Google. Just don’t store it in your database and offer as your own content.
Now that we understand better what SEO means and how it connects with multilingual sites, let’s recap and list what you need to do to achieve good SEO for your sites.
- Write focused content
- Explain to Google what different pages mean by setting good titles, using good headings and good meta description.
- Translate pages accurately, yourself or with ‘real’ translators
- Remember to translate SEO attributes