🎙️ Unscripted SEO Podcast
International SEO: Unlocking Global Markets with Christina Spaulding
Multilingual SEO, hreflang tags, DeepL vs. ChatGPT for translation, and why 85% of buyers prefer to purchase in their first language.
Key Takeaways
- Multilingual SEO doesn’t require international shipping. Target Spanish-speaking customers in the US (25-30% of the population) with Spanish content for top-level keywords at a fraction of the competition and cost.
- Translation tools like DeepL offer deterministic LLM capabilities. Unlike probabilistic LLMs such as ChatGPT, DeepL translates content consistently every time, making it more reliable for translation work—though native speaker review is still essential.
- International links carry more weight when they match your target language and come from country-specific domains. Getting German links from .de or .ch domains is more valuable than German links from .com sites.
- Audit your internal resources before outsourcing multilingual content. Many companies don’t realize they already have employees fluent in target languages who could serve those customer segments directly.
Interview with International SEO Christina Spaulding
Jeremy Rivera: Hello, I’m Jeremy Rivera and this is the Unscripted SEO Podcast. I’m here with Christina Spaulding. We’re going to be talking international SEO this time. I’m super excited about it.
We were talking before the drop a little bit about some of your international experience. Why don’t you give us a recap on where you’ve been in the world and what led you into international SEO. Tell us a little bit about your expertise and what’s given you the authority that you have in this area.
Christina Spaulding: Well, Jeremy, first of all, thank you so much for having me on the show. Really appreciate it.
As we were talking about pre-start, my focus is international SEO, especially. So I do German and French SEO myself. I have colleagues who cover the other languages. I started learning languages in grade school. I did languages in high school, majored in them in college, and then got an internship that took me over to Germany. So I lived there for six years.
While in college, I also did study abroad programs in Quebec City, where they speak Quebecois, and in Strasbourg, France, where they speak not Parisian French—they speak a local version of French, but they also speak Alsatian. So I didn’t get to learn Alsatian, but it was really helpful when I moved to Germany because I moved to the Swiss German border where they spoke Swissdeutsch, which is probably closer to Alsatian than it is to German.
I did jobs in marketing in Germany. It was super fun. I came back to the US for personal reasons, moved out to Las Vegas and started with a German startup. They had a good business in Europe and then expanded to the US. I was their first employee. Now that company is called TextBroker—it’s a content creation company.
So I had to learn about SEO really fast because that was our target client—SEO folks. This was in 2009. So the SEO world looked very, very different than it does now.
The Evolution of SEO: From 2009 to AI
Jeremy: Yes, it certainly did.
Christina: The thing that’s the same between 2009 and now is that there was massive changes. In 2009, you had Panda, you had Penguin. So you had to shift your marketing and SEO strategies to get away from keyword stuffing and really move into quality content.
And now you’ve got AI. We’ve seen a few examples where keyword stuffing is back—low quality blog posts that mention your topic a lot are back. But my perspective on it is that if you are delivering content that meets your customers’ needs and questions, then you will get found. And I think AI does a really surprisingly good job of that.
And it’s sometimes those tiny little details. Maybe you’re the only person that actually mentions a price. You win. Maybe you’re the only person who’s got a certain size of something. You win.
Jeremy: Yeah, someone I was talking to, Michael McDougall, pointed out that with rankings with Google, you’re topically identified in a RAG process of the 10 sites that are most relevant that have the answer.
But really, LLMs are a selection model. They’re choosing one item out and featuring it—which is a different end goal.
LLMs and the Language Data Gap
Christina: One thing I’ll say is that LLMs… Okay, so half of the content on the web is in English. Full 50% of content on the web is in English. The next largest languages have 5% of content on the web.
So what does that mean? That means our database for LLMs is massive—10 times more than any other language.
I’m going to be honest, I haven’t done a lot of experimentation in interacting with LLMs in my target languages yet. But one thing that I think is important for English speaking marketers who are looking at going international is that we should talk about DeepL, which is a fantastic tool for translation work.
Deterministic vs. Probabilistic LLMs
Christina: DeepL is an LLM, but it is a deterministic LLM instead of a probabilistic one.
So if you put in, “I want to go to bed, please translate that for me,” it will translate it the same way every time you ask it to do that. Whereas probabilistic LLMs—like ChatGPT, like Gemini—are calculating in a slightly different way, and they’re always generating a different result. Same prompt, different result.
So whenever you’re working with LLMs, my personal mantra is verify. Usually you say “trust but verify.” I don’t even trust them. Verify. Look at the stuff that you’re putting out before you put it out.
“AI search is entity search. LLMs don’t rank pages — they surface entities. The language you publish in matters less than whether your brand entity is clearly and consistently defined across every platform where the LLM might encounter it.”
— Jason Wade, Ninja AI · Jason Wade on why entity consistency matters more than content volume in AI search →
The Opportunity in Going International
Jeremy: Let’s talk about that use case you mentioned of English language marketers looking to go international.
Christina: So I’m going to talk about multilingual as well as international because here in the US, we have Spanish speakers everywhere. Between 20 and 30% of the US speak Spanish as their first language. And there is search volume for Spanish terms in the US.
So if you are a US-based company, you can serve local customers in Spanish. And you can get top level keywords that have a teeny tiny fraction of the cost and a teeny tiny fraction of the competitiveness.
The 85% Purchase Preference
Christina: Here’s the thing. People prefer by 85% to purchase in their first language.
“Thirteen years at REVOLVE. No link building program. A live A/B test showing 25% organic lift from long-tail category page content.”
— Paul Baterina, REVOLVE | 13 Years at REVOLVE: What Paul Baterina Learned →
And that’s not from me. That’s a study from Weglot. They did a study because they have all of this data on regions and languages and purchasing. They’re like, “Hey, by the way—85% more likely to buy in their native language versus a secondary language.”
Jeremy: That’s a pretty convincing number.
Demystifying Hreflang Tags
Christina: People are very concerned about hreflang tags. So scary, so complicated. I don’t find them complicated. It’s in a different language. It should be marked as such.
Region tags are different. They are incredibly helpful, even if you’re working in the same language. You see that in US versus Canada. Language tags and region tags are both two letters. Usually your language tag is lowercase, comes first. Your region tag is uppercase, usually comes second.
So you would have EN for English and then US for the United States. But if you’re also selling into Canada, then you have EN-CA for Canada.
International Link Building and Server Location
Christina: There are some things that are very similar all across the world. It is almost always better to have links coming from your target language. It is even better if they are coming from pages that end in a country-specific top-level domain.
So if you are getting German language links from a .de or a .at or a .ch domain, that’s better than getting German language links from a .com.
When Server Location Becomes Critical: A China Case Study
Christina: As to server location, server location can be critical in certain countries. I did a project for a company that we shifted from 10 languages down to three—and those languages were German, Chinese, and Japanese. The product was online gambling.
China has a very restrictive internet policy and environment. They also have an anti-gambling law. So if you really want to rank in China, you don’t care about Google. You care about Baidu.
It was all about understanding our customer, understanding the basics of the target market you’re approaching. Very, very, very basic: Is it legal? Can we promote that here?
It is always worth it if you are looking at going international to do that due diligence first.
Connect with Christina Spaulding
- Website: manzanitamktg.com
- LinkedIn: Christina Spaulding
The Unscripted SEO Podcast features candid, unscripted conversations with SEOs and digital marketers. Hosted by Jeremy Rivera.
