**TL;DR.** Tokyo Tea House sells ceremonial-grade Japanese tea internationally. BigCommerce Pro's multi-storefront feature couldn't handle their per-entity locale gating cleanly, creating 1,240 hreflang errors in Google Search Console. Migrating to Ordiko cleaned the cluster in 2 weeks; international organic sales grew 2.3x in 6 months.

## Background

Tokyo Tea House was founded in 2012 in Asakusa by Hiroko Tanaka. The catalog spans:

- Single-origin matcha (Uji, Nishio, Yame).
- Sencha, hojicha, genmaicha leaf teas.
- Tea ware (chawan, chasen, chashaku) sourced from specific kilns.
- Books and educational content in Japanese, English, and translated editions.

International expansion accelerated 2022–2024. By 2025 international (non-Japan) revenue was ~$48k/month spread across US (60%), UK (15%), France (10%), Germany (8%), and other (7%).

## The challenge

Tokyo Tea House had been on BigCommerce Pro since 2020. The plan supported their Japan domestic store + four international storefronts (US, UK, FR, DE) under multi-storefront. But:

1. **Hreflang correctness was broken.** Many products were translated only into JA + EN, but BigCommerce was advertising all five locale variants as if they existed. Google Search Console showed 1,240 hreflang errors growing month over month.

2. **Performance varied by storefront.** The JA storefront was tuned; the international ones were afterthoughts. LCP on FR and DE PDPs ran 3.2–4.0s, hurting organic rankings in those markets.

3. **2026 Product schema fields missing.** Like other ecommerce platforms, BigCommerce hadn't shipped `hasMerchantReturnPolicy` natively. Tokyo Tea House lost rich-result eligibility on most PDPs starting January 2026.

4. **Revenue cap pressure.** As international sales grew, they neared BigCommerce Pro's $400k/year cap; the next tier was custom Enterprise pricing.

## Why Ordiko

The team evaluated Shopify Plus (priced out), commercetools + custom Next.js (rejected due to engineering investment), and Ordiko.

Ordiko's deciding factors:

- **Per-entity `availableLocales`**. A product translated only to JA + EN advertises only those two hreflang values, regardless of the store's full locale set.
- **Native 2026 Product schema** including `hasMerchantReturnPolicy` and `shippingDetails`.
- **8 supported locales** including Japanese as a default-eligible language (one of the rare SaaS platforms with proper JA support).
- **No revenue cap**.
- **Next.js 16 + Cache Components + PPR** for consistent performance across all storefronts.

## The migration

Migration ran over 3 weeks.

**Week 1: catalog and infrastructure**

- BigCommerce bulk-edit CSV export (820 SKUs).
- Asset migration (~6,200 product images including alternate angles).
- Ordiko account + 5 stores provisioned (JP, US, UK, FR, DE).
- Catalog import per store, mapping BC's option sets to Ordiko variants.

**Week 2: translation and SEO**

- Per-product `availableLocales` flagging. Roughly:
  - 40% of products: JA + EN (most matcha and ware).
  - 30%: JA only (Japan-exclusive items).
  - 20%: JA + EN + FR + DE (popular international items).
  - 10%: JA + EN + UK-specific (UK-only beverages and tax-coded items).
- Per-locale meta titles and descriptions.
- Hreflang cluster validation per entity using Ordiko's preview tool.
- 2026 Product schema configuration (30-day return for international, 14-day for Japan domestic per Japanese consumer law).

**Week 3: cutover and tuning**

- Parallel running on Ordiko subdomains.
- QA across 5 storefronts and 6 currencies (JPY, USD, GBP, EUR for FR+DE).
- DNS cutover Saturday 22:00 JST.
- Submission of sitemaps to GSC, Bing, Yandex.

## Results

**Hreflang health** (Google Search Console → International Targeting → Language):

- Day 0 (pre-migration): 1,240 hreflang errors.
- Day 14: 0 errors. Resolution confirmed via crawl.

**International organic sales** (monthly, 6 months post-migration):

| Market | Before (BC)  | After (Ordiko)  | Delta  |
| ------ | ------------ | --------------- | ------ |
| US     | $28k         | $58k            | +107%  |
| UK     | $7k          | $19k            | +171%  |
| France | $5k          | $14k            | +180%  |
| Germany| $4k          | $12k            | +200%  |
| Other  | $4k          | $7k             | +75%   |
| **Total non-JP** | **$48k** | **$110k**   | **+129%** |

Domestic JP sales held flat with a small ~5% lift, attributed to performance improvements.

**SEO surface** (Ordiko-verified post-migration):

| Surface                                       | Before  | After                       |
| --------------------------------------------- | ------- | --------------------------- |
| Hreflang errors (GSC)                          | 1,240   | 0                            |
| PDPs with valid 2026 schema                    | ~10%    | 100%                         |
| PDPs indexed in non-JP locales                  | 180     | 640                          |
| llms.txt published                              | No      | Yes (per store, per locale)  |
| AI crawler allow rules                          | No      | Yes                          |
| Per-store sitemaps with proper lastmod           | Partial | Yes                          |

**Performance** (CrUX p75, FR storefront mobile):

| Metric | Before (BC) | After (Ordiko) | Delta |
| ------ | ----------- | -------------- | ----- |
| LCP    | 3.6s        | 1.4s           | -61%  |
| INP    | 290ms       | 130ms          | -55%  |
| CLS    | 0.12        | 0.04           | -67%  |

**Cost**:

| Line item                           | Before        | After          |
| ----------------------------------- | ------------- | -------------- |
| BigCommerce Pro                     | $399/mo       | $0             |
| Multi-currency / translation extensions | $80/mo     | $0             |
| Schema app                          | $30/mo        | $0             |
| **Ordiko Scale**                    | $0            | **$149/mo**    |
| **Total**                           | **$509/mo**   | **$149/mo** (-71%) |

## What we learned

Hiroko's three takeaways:

1. **Hreflang correctness is invisible until you fix it.** We spent two years thinking we had a multi-locale problem when we actually had a hreflang-correctness problem. International rankings recovered fast once the cluster was clean.

2. **The 2026 schema enforcement was a fortuitous forcing function.** It made us audit the stack we'd inherited from 2020. Migrating now beat patching three more times.

3. **PPR makes a real difference.** A consistent sub-1.5s LCP across all 5 storefronts is a thing we never had on BigCommerce. The static shell + dynamic islands pattern is the right architecture for international ecommerce.

## Quote

> "We sell ceremonial-grade matcha and Japanese tea ware to customers in Japan, the US, the UK, France, and Germany. BigCommerce's Pro plan supported multi-storefront but the hreflang was a constant headache — Google Search Console showed over a thousand hreflang errors that we never fully cleaned up. Ordiko's per-entity availableLocales filtering was the single best technical decision we made. Hreflang errors went from 1,240 to zero in two weeks. International organic sales more than doubled."
>
> — **Hiroko Tanaka**, International Director, Tokyo Tea House

---

*Tokyo Tea House runs on Ordiko Scale at tokyo-tea-house.example.com (JP, US, UK, FR, DE storefronts).*