Korea's recruitment market is less than 2% of the global total. A niche SaaS cannot survive on the Korean market alone.

When I first built Convince-X, I only thought about the Korean market. Korean headhunters, Korean recruiters, the Korean hiring landscape. But the numbers tell a clear story. Korea's TAM for recruitment tech is tiny compared to the global market. Building a sustainable business as a niche vertical SaaS on Korea alone is simply unrealistic.

That's why v1.8.1 marks the beginning of our global pivot.

Korea Alone Isn't Enough

Let's look at the numbers.

To generate $1M ARR in Korea's recruitment tech market, you'd need to capture a significant share of an already small pie. It's not that competition is light -- the market itself is small. Enter Japan and the market grows 7x. Reach the US and it's 50x.

Japan is strategically critical. It's geographically close, headhunting is a well-established practice, and SaaS willingness-to-pay is high. Plus, someone who understands the Korean hiring market has a relatively strong foundation for understanding Japan's.

Localization, Not Translation

The first trap in going global is the assumption that "translating to English = going global."

Translation and localization are different things. Translation changes text. Localization changes the experience.

It's not just UI text. AI-generated analysis reports need different tones per language. Korean reports can be direct. Japanese reports require more polite phrasing. English reports prize brevity.

The Weight of 550 Keys

Convince-X's i18n system manages over 550 translation keys. With Korean, English, and Japanese translations for each, that's 1,650+ text entries total.

The implementation is simple. Add a data-i18n attribute to HTML elements and load translations from JSON files. Built with vanilla JS, no framework needed.

Harder than creating 550 keys is maintaining consistency. You can't call the same concept "Analysis" on one page and "Analysis Results" on another. "Candidate" in one spot and "Applicant" in another won't do.

To solve this, I built a glossary. Thirty core terms with official translations in all three languages, and every key follows this glossary.

The Secret Behind 400 Demo Profiles

i18n isn't just about UI text. Demo data needs localization too.

Convince-X's demo mode includes sample profiles across 13 job categories, letting users experience the product without logging in. If all demo profiles feature Korean names and Korean companies, English and Japanese users immediately think: "This is Korea-only."

So I created 200 English demo profiles and 200 Japanese demo profiles separately.

Designing 400 profiles seems like grunt work, but each profile needs to look "realistic." A profile with 5 years of experience and a CTO title? Recruiters spot that as fake instantly. I validated every career trajectory using 12 years of industry intuition.

Deployment Day: v1.8.1

On March 15, 2026, v1.8.1 went live on the cloud.

Key changes in this release:

The deployment itself was straightforward thanks to cloud hosting's auto-deploy. One git push handles build and deployment automatically. At a reasonable cost, it's more than sufficient.

The first thing I did post-deployment was verify the full flow in all three languages. Log in in Korean → analyze → check results. Switch to English → repeat the same flow. Switch to Japanese → repeat again. All manually. We have 607 automated tests, but the "naturalness" of translations can only be verified by human eyes.

The Japanese Wall

Of the three languages, Japanese was the hardest. Three reasons.

First, the honorific system. Japanese has sonkeigo (respectful), kenjougo (humble), and teineigo (polite) forms. You need to decide which level of formality to use in UI text. Too formal feels stiff; too casual comes across as rude. For a SaaS tool, "polite but approachable" is ideal -- and achieving that in Japanese is genuinely difficult.

Second, text length variation. The same sentence in Japanese is often 30-50% longer than Korean. Buttons, error messages, and tooltips in constrained spaces need UI adjustments to prevent Japanese text from being truncated.

Third, kanji usage. The same kanji characters can have different readings and meanings in Japanese. Recruitment terminology particularly relies on proper kanji usage -- "面接" (interview), "採用" (hiring), "候補者" (candidate).

Japanese localization isn't perfect yet. Native speaker review is needed. But the gap between "nonexistent" and "exists but needs improvement" is enormous. Ship it first, then improve based on Japanese user feedback.

Going Global Is a Choice

Going global isn't a "nice to have" -- it's a survival strategy for niche SaaS.

Serving only Korea caps your TAM (Total Addressable Market). But supporting English and Japanese expands the market by 50x or more. With reasonable hosting costs, even a small number of global paid conversions covers server expenses.

Of course, going global comes with challenges.

But these challenges must be embraced. Staying in Korea feels comfortable, but "comfortable markets" don't exist. A small market doesn't mean less competition -- it means a smaller pie.

v1.8.1 is the first step. 550 keys, 400 demo profiles, 3 languages. On this foundation, we're preparing for a global launch and introducing Convince-X to the global recruiter community.

We started in Korea, but we won't stay there.