What’s The Real Value Of Local TLDs For SEO?
The history of using local Top Level Domains (TLDs) for multinational SEO benefit is a long one.
However, with the extensive localisation controls provided in Google’s webmaster tools these days, is there still a case for multinational strategies using local TLDs for SEO benefit?
A History Of TLD Optimisation
Back in 2005, Google’s emerging webmaster guidelines advocated the use of local TLDs for content intended for a particular country. This meant, for example, that to target French language content to France, hosting it on example.fr would be beneficial as the TLD was the major factor considered when localising content.
At the time, Google also advocated hosting on an IP in the target country and made it clear that it was also essential to tag up content with meta language tags so Google could determine your intended audience most accurately.
All of these technical requirements were important metrics for their crawler localising content to the correct country version of Google.
Largely, this was because it was, at that stage, only possible to directly locate a URL to a location using Google tools for Google Local Listings (i.e., you could set a Google Map link to point to your domain, thereby tying that domain to the map location in Google Maps).
In short, there was no way to actively set geographic targets across multiple countries with any of the major search engines at the time. It was also true that intuiting the location on their part was a hard problem. Errors were rife, most especially for multinational brands.
Invariably, issues would arise for such brands with the intuited localisation of the homepage regularly being dragged across different countries with dramatic effects for incoming organic search traffic (imagine if your UK marketplace traffic lost 80% of its natural search traffic for a whole month while you battled to wrest control of it from Google US listings on Google.com back to UK listings on Google.com.uk).
Localisation was messy, hard, and frustrating for webmasters andsearch engines alike. Doing everything you could to make that localisation easier was a positive boon for your multinational SEO campaign.
Cometh The Hour, Cometh Webmaster Tools
In June 2005, Google launched its Sitemaps Tool into Beta, allowing webmasters direct input into how Google would crawl their sites. This service then morphed into the Google Webmaster Tool in August 2006, and new features began to be regularly built in — often tested in beta with high volume site webmasters. (Off topic, but does anyone else remember the impact crawl controls had back then? SEO had definitely never been so easy!)
Then, in October 2007, Google rolled out the ‘Set Geographic Target’ tool, finally allowing webmasters to dictate the target of their website content to Google.
At that point, Google began to remove guidance that using a different TLD was essential for accurate localisation across different countries. They did, however, maintain that they would “still make geographic assumptions based on the top-level domain and the IP of the webserver.”
They have continued to advocate using a local IP to serve content, though Google’s reasoning now is more dependent on improving poor site load times for domains served from a remote country.
In December 2011 (yes, four years later!), Google rolled out their hreflang link element (albeit after an initial flirtation with canonical tags alongside) to allow webmasters to dictate to Google the intended audience for their content in more granular detail and with one, significant, change: while previously Google had treated multiple domains targeted to different countries as duplicate if they were largely the same content (or just basic translations rather than uniquely written content); now, even totally identical content would not be considered duplicate if it was intended for a different country.
This is true for site architectures that sit on a single TLD, or sites that span multiple TLDs intended for different countries. Google followed up by pointing out that there is no inherent benefit to using a local TLD if your content is accurately localised.
At this point, looking objectively at the ideal site architecture for a multinational campaign becomes very interesting. Given the decoupling of TLD from localisation in the eyes of Google, it means that for most markets where Google has dominant marketshare, there is no SEO justification for using different TLDs in your site architecture that aren’t superseded by the new flexibility and power of the hreflang tool.
And, in fact, you’d likely be actually damaging your SEO campaign with a multi-TLD strategy because you are then forcing each TLD to attempt to perform on its own merit in non-Google search engines, a strategy which would require your client to commit the same amount of linkbuilding resource to each target country as would need to be applied to a single domain for the same level of performance in a single-TLD campaign.
Of course, the argument could be made that you’d risk the same issues experienced in Google pre-hreflang tag in the likes of Baidu, Yahoo!, Naver, Bing, Ask & Yandex; but, the reality is that localisation for those engines can largely be controlled by delivering from local country IPs — meaning the use of a Content Delivery Network (CDN) with exit nodes in your target countries would cover off your localisation as well as deliver significant page speed benefits (and so therefore also boost your performance in Google: hey! It’s win win!).
Supporting your single-domain strategy with accurate HTML language tags is also best practice regardless of localisation issues, and will cement your localisation in those non-Google search markets.
In short: whether you prefer subdomains or directories, a single-TLD solution has SEO benefits that simply can’t be achieved with a multi-TLD solution, and should always be the route taken if it is a viable option for your company.
A Final Word On Local TLDs
Of course, one important element deliberately not touched on in the preceding argument, is the user benefits of local TLDs. For example, a German speaker searching from Germany using Google.de is frequently argued to be more likely to click on a .de TLD result than otherwise.
This certainly rings true, but it’s also important to consider other factors that come into play in that scenario: every element of the snippet will have an impact on the CTR here.
It’s surely much more important to ensure you have a correctly written German title attribute and meta description featuring the search term (so it is triggered) and a strong call to action (to encourage clickthrough) and to feature the search term in the URL (reinforces the page about your search term) than fixate on the difference between .com and .de (just three characters out off all the other variables).
In addition, I’ve yet to see a study which normalises for brand recognition bias whenever the assertion of local TLD preference is made. On the other hand, multiple heatmaps are available which show how little attention is given to the URL in the SERP snippet, or any part of the URL in the browser address bar.
Opinions expressed in this article are those of the guest author and not necessarily Search Engine Land. Staff authors are listed here.