For years, when I thought about going multilingual, I thought about translation.
Translate the website into French, German, Spanish, Italian. Maybe spin up a few extra blogs. Hope a slice of new traffic shows up in each language.
That made sense when a website existed mostly for two audiences: human visitors, and the search engine ranking the page for them.
It doesn't make as much sense anymore.
The audience for your content isn't just people anymore
Today, the things I publish get read by people, by Google, by ChatGPT and Claude and Gemini, by recommendation systems, by knowledge graphs, and by crawlers I've never heard of that show up in my server logs at three in the morning.
Some of that traffic is based in the US. A lot of it isn't. I see crawler activity from Japan, Singapore, South Korea, China, and across Europe ... places I'm not actively marketing to in any language.
Most of that isn't a person sitting at a desk reading my article in English. It's a system. A research crawler. A model pulling content in for later use. Whether I planned for it or not, my content is already part of a global, machine-readable ecosystem.
So the real question isn't "should I translate my site." It's "how do I build assets that travel across language boundaries on their own?"
Most operators are still set up in single-language silos
Almost everyone I talk to is set up the same way: English website, English YouTube channel, English everything. Then they wonder why their audience never grows past one language market.
Meanwhile the systems reading and redistributing that content operate globally, in every language, all day, without asking permission.
That mismatch is the opportunity.
Start with the asset, not the language
Old-school multilingual marketing starts with translation: take a finished thing and make four more versions of it.
I'd rather flip that around. Build the asset first ... the article, the video, the glossary entry, the dataset, the structured page, the AI endpoint ... and treat language as a distribution layer you add on top, not a production requirement you carry from day one.
Building five separate websites in five languages is expensive, and now you've got five things to maintain. Building one strong asset and distributing it in five languages is leverage. Same asset, more reach, without multiplying the work.
That's the whole shift in one sentence: translation is an activity, leverage is an outcome.
YouTube quietly solved part of this
One of the more useful things to land in the last year or so, and most people I know haven't touched it yet, is YouTube's support for multiple audio tracks on a single video.
Record a video once. Add a Spanish audio track. A French one. German. Italian. All under the same video, same URL, same view count.
For tutorial content, screencasts, build-in-public videos like the ones I make walking through the Krisada properties ... that's enormous. The visual stays exactly the same. Only the language layer changes. Someone in Mexico City or Madrid or Berlin can watch the same screen recording I made, with audio in their own language, and never know it started life as an English video.
I haven't fully built this out yet, but it's on my list. The cost of adding one extra audio track is nowhere close to the cost of producing a second video.
Why Spanish is probably the first move, not German
When people hear "go multilingual," they usually picture Europe first. France, Germany, Italy.
For most of what I write about ... AI, SEO, website systems, digital assets, owning your own platform ... the bigger and closer opportunity is Spanish. United States, Mexico, Puerto Rico, Colombia, Argentina, Chile, Peru. That's a massive audience, and in most of these topics the competition in Spanish is nowhere near what it is in English.
One translated audio track, or one well-structured Spanish version of a core asset, can put existing work in front of millions of people who were never going to find the English version anyway.
Don't start by translating your whole library
If you've got fifty articles, the instinct is to translate all fifty. Don't.
That's a maintenance project, not a strategy. Every translated article is now a thing you have to keep updated in two languages, forever.
Instead, look at the core knowledge architecture underneath the content: definitions, glossary entries, entity descriptions, frameworks, the structured stuff. On this site, that's the glossary entries ... short, structured, definitional pieces that describe one concept cleanly.
Those are easier to maintain, easier to translate, and easier for AI systems to consume than a 2,000-word article. A multilingual glossary, where a handful of core concepts exist in English, Spanish, German, French, and Italian, often produces more long-term reach than a multilingual blog ever will.
This is where it connects to the federation work
This is part of why the Digital Karma Federation work matters to me beyond being a fun side project.
The federation files ... the manifest, the catalog, the role and relationship declarations across properties ... are already a structured, machine-readable layer sitting underneath the human-facing content. Concepts like Digital Asset, Leverage Building, AI Visibility, and Federation Architecture already exist as defined entities in that layer.
The next step isn't "build more sites." It's making sure those core entities can exist in more than one language while still pointing back to the same underlying concept. One node, multiple language surfaces. That's a multilingual knowledge graph, and it's a different animal than a multilingual website.
What I'd actually do next
If I were starting this for my own properties tomorrow, I wouldn't start with a translation budget. I'd start here:
- Pick one or two cornerstone concepts per site and write clean, structured definitions for them.
- Translate those definitions into Spanish first ... it's the lowest-competition, highest-reach move for most of what I write about.
- Add a Spanish audio track to a handful of existing tutorial or screencast videos instead of producing new ones.
- Make sure the structured data for those entities references the same underlying concept across languages, so the connection is machine-readable, not just a translated page sitting next to the original.
None of that requires a second website, a second content calendar, or a second production pipeline. It requires looking at what already exists and asking where else it can travel.
The future of multilingual marketing probably isn't thousands of translated pages. It's a smaller number of well-built, structured, multilingual assets ... articles, glossary entries, videos, knowledge graph nodes ... created once and left to keep working in every language a person, or a machine, happens to read in.
That's the kind of asset I want more of in 2026.
