One Site, Two Readers: A Bilingual Portfolio Structure That Survived My Japanese Blog
Photo: bilingual website wireframe on laptop, Japanese and English navigation labels, designer desk, notebook, via www.newyorkfamily.com
For a long time, takajun.net was a Japanese blog with an English problem. Japanese readers arrived at the homepage, browsed the categories, and stayed for a while. English readers arrived from search, landed on a single project page, looked around for a sign in their own language, and left. The content was fine. The navigation was not.
The obvious fix is to run two sites: a Japanese blog and an English portfolio. I tried that in my head for about a week. Two sites mean two publishing queues, two sets of URLs, two analytics dashboards, and two places for a broken link to hide. For a solo creative, that is not bilingual publishing. That is a part-time job you invented.
In 2026, I rebuilt Takajun as one site with two reading paths. It is not a perfect system, but it is a survivable one. Here is what I learned, and a simple template you can adapt if you are a bilingual creative who does not want to maintain two separate homes on the internet.
The problem was never the Japanese
English-speaking visitors do not need every paragraph translated. They need to know three things within a few seconds: what this site is, what they can do here, and how to read more in their language. The Japanese body text can stay Japanese. The frame around it has to be legible.
The first analytics surprise was that English visitors were not bouncing because they could not read Japanese. They were bouncing because they could not tell whether the page was a blog post, a portfolio piece, or a store. The Japanese typography looked like content, but the navigation looked like decoration. They had no idea where the 'work' was.
That changed how I thought about the project. It was not a translation problem. It was a labeling problem.
Keep one site, add a language layer
One URL structure, one content system, one search index. Language becomes a layer on top of the content, not a separate building next door. That means every page has a language attribute, a visible language toggle, and a short summary in the other language near the top. It does not mean every page is fully translated.
My rule is simple: translate the frame, summarize the content, and leave the voice alone. The frame includes navigation labels, headings, dates, category names, image alt text, and the one-line description. The voice includes the actual writing. I would rather have a clear English summary of a Japanese essay than a stiff machine translation of the whole thing.
Navigation labels that survive a quick skim
Bilingual labels should be short and parallel. Long translations create visual noise. Here is the set I use:
- Work / 仕事
- Notes / 記事
- About / プロフィール
- Contact / 連絡
Notice that the Japanese is not a literal translation of the English in every case. 'Notes' becomes '記事,' which is closer to 'articles.' 'About' becomes 'プロフィール,' which is closer to 'profile.' That is fine. The goal is recognition, not dictionary accuracy. A visitor should be able to guess where the link goes without reading both languages.
The language toggle is not a flag icon. Flags are country shortcuts, and this site is not a country. I use plain text: EN / JP. It sits in the same place on every page. If a page is only in Japanese, the English summary at the top says so. If it is only in English, the Japanese summary does the same. No dead ends.
The article template I actually use
Every post, in either language, follows a simple top-to-bottom pattern. It keeps the site from turning into a pile of one-off layouts.
- Title in the original language.
- A one-line subtitle in the other language.
- Date, category, and reading time.
- A three-sentence summary box in the other language. This is the most important part for search visitors.
- The body in the original language.
- A short 'what this covers' list at the end, in the other language.
- Previous and next links.
That is it. The summary box does the heavy lifting. It tells an English reader whether the Japanese post is worth their time, and it gives search engines a clean description. It also forces me to write a real summary instead of pasting the first paragraph.
What the analytics actually showed
Three surprises, in order of usefulness.
First, English readers spent more time on process posts than on finished-project posts. They wanted to see the messy middle: the sketches, the failed layout, the note that says 'try again.' Japanese readers were more likely to click through category pages and read several posts in a row. That told me the English path needed more cross-links, not more translation.
Second, mobile readers used the language toggle far more than desktop readers. That makes sense. On a phone, you cannot open a second tab and puzzle out the navigation. The toggle has to be visible without scrolling. I moved it into the sticky header and the toggle use doubled.
Third, the search queries were not what I expected. I assumed English visitors wanted 'Japanese design.' A lot of them wanted 'Tokyo desk setup,' 'small apartment storage,' and 'Japanese stationery under 1000 yen.' Those are specific, practical searches. The lesson was not to write more about Japan in general. It was to write clearer titles for the posts I already had.
Translation choices that keep the voice intact
I do not translate puns. I do not translate rhythm. I do translate dates, prices, and measurements into a format the other audience can use. I keep Japanese terms when they are the point—'genkan,' 'shotengai,' 'washi'—and add a short gloss the first time. I avoid honorifics in English summaries unless they carry meaning.
The biggest rule is: do not let the translation flatten the tone. If the Japanese post is casual, the English summary should be casual. If the Japanese post is a straight listicle, the English summary should not sound like a museum plaque. Readers can tell when a voice has been sanded down.
A template for bilingual creatives
If you are building this from scratch, start with one site and three page types: a home page, a work index, and a post template. Put the language toggle in the header. Write the navigation labels in both languages. Add a summary box to every post. Use one analytics tool and look at landing pages, not just pageviews.
Do not build two sites unless you have a real reason. Two sites double the maintenance and split your search authority. One site with a clear frame is easier to keep alive, and alive is the only design principle that matters for a personal site.
The bilingual portfolio problem does not get solved once. It gets managed. The site is a translation of a translation of a person. The least you can do is make the door easy to open.