SEO, Easyling, and You

Jun 11, 2014 - Easyling.com

While browsing the most frequented linguist groups, I recently came across a discussion of an article on SEO for hotels and how it impacts customer conversion. One of the most frequent questions I get is on how the translations are stored in Easyling. Believe it or not, these two are closely related. And that prompted me to write this article on how Easyling can help your SEO efforts along.If you’ve been keeping up with this blog, you probably know already what Easyling can do, but for the sake of new readers, a quick recap: Easyling is a website translation proxy - it allows you to take an entire site, extract the content from the pages, process it in your favorite CAT tool, and re-insert the translations, then publish your site, powered by Google’s infrastructure. That penultimate bit is where the magic happens for SEO.

Easyling’s proxy operates in the same infrastructure Google Search, Gmail, Youtube, etc. uses, so one thing that’s guaranteed is uptime: as long as the source site is live, the translations will be as well. Another thing is speed: Gmail and other Google services are so fast, because they are always near you, geographically - existing beside the AppEgine infrastructure is the EdgeCache network, a network of data centers and servers all over the globe. Under the appropriate circumstances, Google will cache the translated site’s content in these EdgeNodes, making future requests to the site from the same geographical region much faster.

Given the rise of multilingual content on websites, Google already has facilities in its PageRank algorithm to accommodate these cases. Although Google only uses the visible content as a pointer towards the site language (making the lang attribute useless for “tricking” the engine), it can also match that content across domains and subdomains, and if it finds the same content in different languages, it may react differently if their relation is identified (i.e. it may not split ranking scores between the two, letting them augment one another). This relation is identified by the rel=“alternate” hreflang=“x” attributes. Unfortunately, maximizing returns from this method will require you to put in a little elbow grease in code, but the results are well worth the trouble. Google has a helpful primer on the use of these attributes. To cut a long story short (and eliminate a lot of technical, but interesting and relevant, information), you need to modify your site or your site’s template to include a number of link tags, with the appropriate attributes. Each tag needs to specify a URL, a language, and a relation, and as Google’s post warns us, these must satisfy two criteria to be taken into account when determining the final ranking: a) they must be reciprocal (the linked site must link back), and b) they must be self-referential (the site must link to itself as well). With that in mind, let’s see an example page header (Waring: heavy code incoming!)

<head>
    <title>Your fantastic vacation awaits!</title>
    <meta charset="UTF-8">
    <meta content="[various relevant keywords]" name="keywords">
    <link href="http://www.examplehotels.com" rel="alternate" hreflang="en-us">
    <link href="http://www.examplehotels.com" rel="canonical" hreflang="en-us">
    <link href="http://es.examplehotels.com" rel="alternate" hreflang="es-es">
    <link href="http://uk.examplehotels.com" rel="alternate" hreflang="en-uk">
    <link href="http://jp.examplehotels.com" rel="alternate" hreflang="jp">
</head>

Now that we got over the initial shock, let’s see what this very bare-bones header means! The title tag should be fairly obvious: it specifies the title displayed in the browser. The meta tag is starting to mix things up a bit, since it’s literally just that, a tag that contains information about the page - its character encoding (“UTF-8”, the most widespread encoding scheme these days), its search keywords, etc. Finally, we get to where the real SEO happens, the link tags. The first one is the self-referential link needed for Google to acknowledge the references. The second one defines the “default” URL - in the absence of any language preference or similar, this result will “get all the points”. The other tags reference translated versions of the site, and since Easyling replicates page headers without modifications, as long as the source site contains references to all the target language sites, the reciprocity criterion will be satisfied as well.

If you’re interested in giving Easyling a try, head on over to our website, or register a free trial, and give the proxy a trial run!