In this episode of the International Web Effectiveness Podcast, Bill Scully and Bill Hunt break down the details of multi-regional domain structures in international SEO, inspired by an in-depth research report on keyword rankings.
They discuss the implications of Google’s AI-powered expansion of Google Translate, the benefits and challenges of using country code top-level domains (ccTLDs) versus subdirectories, and the crucial role of HREFLang tags in managing duplicate content across different markets.
Key Takeaways
- Google Translate Update: Google announced adding 110 new languages to Google Translate, which could benefit localizers and AI translation tools. However, it’s unlikely to impact Hreflang implementation significantly.
- Domain Structure Research: A study by GA agency and SC Ranking analyzed 20,000 keywords across 15 markets and found that 56% of websites use ccTLDs (country code top-level domains), 20% use folders, and 24% use subdirectories.
- Domain Structure Considerations: The choice of domain structure should be based on several factors, including:
- Resources: ccTLDs require more resources to build, maintain, and optimize than other options.
- Brand Identity: Consider how the domain structure aligns with your brand identity in different markets.
- Infrastructure: Ensure your infrastructure supports the chosen domain structure.
- Duplicate Content: HREFLang can be used to address duplicate content concerns when using the same product page for different languages.
- Best Practices:
- Choose a domain structure that aligns with your infrastructure and resources.
- Implement Hreflang correctly to avoid duplicate content issues.
- Be consistent with your chosen domain structure.
Catch the Full Episode
YouTube: https://www.youtube.com/watch?v=MQ3nsQYrTZk&t=687s
Spotify: https://open.spotify.com/episode/6W1q3NCy3OdN9HIN1ytLVS?si=BmpuncESRqaqtNkZKqL25g
Apple: https://podcasts.apple.com/us/podcast/episode-2-domain-structures/id1755789311?i=1000661661910
Transcript
Bill Scully: Welcome back to the second episode of the International Web Effectiveness Podcast. I’m Bill Scully, joined by Bill Hunt, and today we’re focusing on strategies and techniques to ensure your web presence serves all stakeholders globally.
In our last episode, we discussed the frustrations of international SEO and Hreflang implementation. Today, we’re diving into multi-regional domain structures. But first, let’s talk about some recent news.
Google recently announced that they’re using AI to add an additional 110 languages to Google Translate. Bill, how do you see this impacting international SEO and Hreflang?
Bill Hunt: For most people, it won’t be an issue. Google now translates 240 languages, but only 107 of them have an official ISO 639 two-letter language code. That means the remaining languages won’t have usable codes for Hreflang, since Google doesn’t support three-letter codes.
Most of these new languages are lesser-spoken ones that won’t pose duplicate content risks like English or Arabic. So, from an SEO standpoint, it’s not a major concern. However, this could be great for the localization community, especially with AI-powered tools from Google. But in terms of direct Hreflang benefits, I don’t think this changes much.
Bill Scully: Got it. So, this expansion in Google Translate—does it use AI?
Bill Hunt: Yes, it’s part of Google’s AI-powered translation system. It’s integrated into website translation and also available through the Google Translate app. As a Google Pixel user, I even received a notification that these new languages are now supported natively on my phone.
Bill Scully: Interesting. And for those who don’t know, does using the Google Translate website widget provide any SEO value?
Bill Hunt: No, and that’s something a lot of people misunderstand. Many site owners add the Google Translate widget and then wonder why their translated pages don’t appear in search results.
The issue is that Google’s translation widget doesn’t create actual web pages—it simply changes the text dynamically on the screen. If you look at the URL, it doesn’t change when you switch languages, meaning Google isn’t indexing those translated versions.
If you want indexed multilingual pages, you need to create separate language versions of your site. And I always recommend having a native speaker review any machine translation before publishing it.
Bill Scully: That makes sense. Alright, let’s move on to our main topic. You and I were discussing one of our biggest pet peeves—the oversimplification of domain structures from an SEO perspective.
Everyone seems to have strong opinions on this, and it ties into why some companies don’t follow standard SEO conventions. It’s also a continuation of our last episode, where we covered infrastructure chaos as a major challenge in international SEO.
When we were planning this episode, you shared a research report from the GA Agency. We haven’t discussed it in depth yet, but can you give our listeners an overview and share your thoughts on it?
Bill Hunt: This study is actually really solid and well done. Danwood from the GA Agency partnered with SC Ranking to analyze about 20,000 keywords across 15 different markets—mostly in Europe, with Japan included as well. They looked at 1.7 million SERP positions, focusing on a mix of e-commerce-related keywords.
What they found was interesting: 56% of the ranking sites used country code top-level domains (ccTLDs), while 20% used language-based subdirectories on generic top-level domains (gTLDs) like .com, .co, or .tv. It’s probably the most comprehensive study done on this topic so far, and it got some media attention—I first saw it in Search Engine Land.
I wanted to flag it for discussion because it sparked a lot of questions. People started reaching out, asking, “Should I switch to a ccTLD since the majority of ranking sites in Europe are using them?” But what I appreciated about Dan and his team’s conclusion in Search Engine Land was that they didn’t take sides. This could have easily turned into clickbait—“Most sites use ccTLDs, so you should too!”—but instead, they provided a balanced take.
Their key point was that while ccTLDs can be beneficial, they require more resources to build, maintain, and optimize. They also highlighted an interesting case in Switzerland, where brands often use subdomains to separate content for French, Italian, and German-speaking users. While subdirectories weren’t the dominant approach, they were still widely used. Ultimately, their advice was to choose a structure that aligns with your overall strategy.
This ties into something we see often—most articles on domain strategy lack real-world context. They’re written purely from an SEO standpoint, without considering operational and business factors. That’s why I think it’s important to break down the reasons behind why companies choose different approaches.
Bill Scully: That’s a great point. And before we dive into that, I agree—the article was well done and data-focused. But I do have some questions. Since the study focused on e-commerce, did it consider user behavior? Do people prefer clicking on local e-commerce sites? Does Google prioritize local results for e-commerce queries? And were the ranking sites truly global brands, or just local businesses that excelled at SEO?
Did the article answer any of those?
Bill Hunt: Not directly. But they did touch on some regional trends. For example, in Portugal, the top-ranking sites were mostly .coms rather than local .pt domains. They spoke with a local SEO expert who suggested that this could be because there are fewer native Portuguese e-commerce companies, allowing global brands to dominate.
This raises an interesting question about local vs. multinational brands. Are local businesses inherently more visible in their home markets simply because they focus on localized content? A single-market brand naturally integrates local references and signals—“We’re in this town, serving these people”—whereas a multinational brand is trying to push the same content across multiple regions.
And speaking of e-commerce, I noticed something while researching for Hreflang Builder. One of our helmet manufacturing clients had a visibility issue in Portugal. When I searched for their brand name with a VPN, their official website ranked fifth—behind four e-commerce sites selling their helmets. That’s despite it being a branded search!
This ties into something Rand Fishkin once pointed out—there’s a phenomenon where manufacturers often rank lower than retailers selling their products. I don’t remember the exact term he used, but it’s a common issue. It seems that for product-related searches, Google often prioritizes retailers because users are looking for choice, not just the official brand page.
Bill Scully: That’s a classic SEO challenge. And when it comes to domain strategy, I always say it’s an infrastructure decision first—SEO is just one factor. Sometimes, SEO drives that decision, but there are bigger business considerations.
For example, if you’re expanding into a new market, does your brand need to change? I think there’s a mattress company called OT in the UK, but in the U.S., they go by Helix due to naming conflicts. That’s an example of how branding affects domain choices.
Logistics also play a role. If you need separate fulfillment centers, marketing teams, and customer service operations for each country, it might make sense to use ccTLDs. On the other hand, if you’re centralizing operations and just localizing content, subdirectories might be the better option.
And going back to your e-commerce example—Google’s preference for local retailers makes sense. These businesses are doing the work, creating localized content, and signaling that they serve a specific market. So in a way, Google is balancing what’s best for users with an inherent bias toward local businesses.
Bill Hunt: You make an excellent point about infrastructure. I’ve worked on a project for a 2.0 product with multiple companies, and one of the key challenges was handling ccTLDs. In global companies, subfolders are often used instead, particularly in regional markets.
For example, I recently worked with a company operating in seven different markets across North Africa, the Middle East, and India. They structured their site using subdomains, with each market having a local language subfolder. So, they had URLs like mydomain.com/english and mydomain.com/arabic, creating an added layer of complexity.
One thing the research didn’t fully cover is how these global organizations manage site structures internally. Once they built their Middle East site, a separate team handled its market structure, while local teams managed content for individual regions. The study took a macro-level view, but I’d be curious to see if it examined these nested structures as well.
Another challenge, particularly for e-commerce, is duplicate product pages. Many global companies use the same product descriptions across multiple markets. If a company has 40 versions of the same product page, is it better to place them under one ccTLD or spread them across 47 subfolders? And when dealing with multiple languages, how does that impact SEO?
Bill Scully: That’s a great point. The issue of duplicate content comes up frequently, but we’ve both heard the same answer from Google—if you use Hreflang correctly, duplicate content isn’t a problem. What’s your approach to handling this?
Bill Hunt: That’s a good question. I recently worked with a tech company where the legal and compliance teams were concerned about duplicate content. Google has clarified many times that if you implement Hreflang correctly, it distinguishes content by language.
One challenge this company faced was maintaining consistent URLs due to compliance rules. Every product page had to retain the same URL structure, which complicated Hreflang implementation. Their development team eventually realized that their issues stemmed from incorrect Hreflang setup.
Another mistake they made was trying to slightly modify product descriptions for different regions. Ironically, this performed worse than simply keeping the descriptions identical and using Hreflang.
A second major issue was incorrect Hreflang linking. Their implementation check revealed that six product pages had conflicting Hreflang tags pointing to the wrong versions. That’s an entire discussion on its own—maybe a topic for another episode.
The key takeaway? If you’re using Hreflang correctly, you don’t need to worry about duplicate content. Google understands that the same content in different languages isn’t duplicate content. Just make sure your Hreflang implementation is accurate.
Bill Scully: That’s a great explanation. I think we’ve covered the core issues well. Any final thoughts on domain structures?
Bill Hunt: The most important thing is to consider your infrastructure first. We’ve both seen cases where a company’s infrastructure doesn’t align with SEO best practices, leading to long-term challenges.
Second, if you’re using different domain structures—whether ccTLDs, subdomains, or subfolders—implement them correctly from the start.
Finally, consistency is key. If you’re dealing with product pages across multiple regions, Hreflang is your best tool for avoiding duplication issues.
Bill Scully: Great advice. Thanks, Bill.
Before we wrap up, I want to thank everyone for tuning in to Episode 2. You can find the show notes on our website, internationalwebmastery.com, along with links to the Google Analytics article and other resources on domain structures.
If you have any questions or topics you’d like us to cover in future episodes, reach out to us on social media or through our website. See you next time!