**TL;DR.** Boutique Papier, a fountain-pen-paper specialist in the 6th arrondissement of Paris, migrated from Squarespace Commerce to Ordiko to gain native multi-locale support and AEO surfaces. In 120 days, AI search citations grew 3.5x and EU organic traffic doubled.

## Background

Boutique Papier is a stationery brand founded in 2020 by Camille Lefèvre and Antoine Charron. They specialize in fountain-pen-friendly paper — notebooks, loose sheets, and writing pads with weights between 80g/m² and 120g/m² and acid-free, archival construction.

The catalog is small but considered (185 SKUs). Average order value is €68. The customer base is split roughly 40% France, 25% Germany, 15% Italy, 10% Spain, 10% other EU/UK/US.

## The challenge

By Q1 2026 Boutique Papier had outgrown Squarespace Commerce on three axes:

1. **Multi-locale**. Squarespace's multi-language handling required a third-party plugin (~€40/mo) that didn't correctly emit per-entity hreflang. Pages frequently advertised translations that didn't exist, causing crawl errors and ranking suppression in Germany and Italy.

2. **AI search**. Customers in Berlin and Milan increasingly reported that Perplexity and ChatGPT recommended American or British competitors when asked about fountain-pen-friendly paper. Boutique Papier had no llms.txt, no Markdown twins, no AI crawler controls. They were invisible to AI engines.

3. **Performance**. PDP mobile LCP averaged 2.6s on Squarespace's Fluid Engine, kept above the 2.5s threshold by Squarespace's mandatory image-loading script. Lighthouse SEO scores hovered at 88; performance scores at 78.

A fourth motivation: Camille and Antoine wanted to publish a glossary of paper terminology — fiber composition, sizing methods, GSM weights, archival ratings — as a long-term AEO play. Squarespace's content tools weren't built for taxonomy-rich publishing.

## Why Ordiko

The evaluation list: Shopify (Markets pricing too high for their scale), Wix Commerce (similar limitations to Squarespace), self-hosted WooCommerce (rejected due to maintenance burden), Ordiko.

Ordiko's deciding factors:

- **8-locale native support** including French, German, Italian, Spanish — exactly Boutique Papier's market set.
- **Per-entity hreflang correctness**. A product translated only to FR + DE + IT will only advertise those three, not the full locale set.
- **llms.txt, llms-full.txt, Markdown twins** built in.
- **DefinedTerm schema for glossary**.
- **Clean editorial content tooling** via MDX collection.

## The migration

Migration window: one weekend (Friday evening → Sunday afternoon).

**Friday evening**: Squarespace catalog export, asset download, Ordiko account setup.

**Saturday**:
- Catalog import (185 SKUs in 25 minutes).
- 4-locale configuration (FR default, DE, IT, ES; EN as a secondary global).
- Per-product `availableLocales` flagging — about 40% of SKUs are FR-only; 60% are FR + DE + IT.
- Squarespace permalink → Ordiko permalink redirect mapping (32 entries needed; most slugs preserved).
- Theme customization (Camille is design-led; Antoine's brother is a frontend developer who handled the React Server Component theme work).

**Sunday**:
- DNS cutover at 14:00 CET.
- Sitemap submission to Google Search Console, Bing Webmaster, Yandex.
- Smoke testing across all four locales.

## The AEO build (post-migration)

The bigger initiative ran over the next 8 weeks:

**Glossary**: 62 terms covering paper, ink, pen, and writing topics. Each entry: definition first sentence ≤30 words, 2–3 paragraph explanation, related terms cross-links, DefinedTerm schema. Translated to FR/DE/IT.

**Guides**: 12 long-form guides — "How to choose fountain pen paper", "GSM explained", "Why bleed and ghosting happen", etc. Each ~1,500 words with TL;DR, H2 questions, tables, FAQ blocks. Markdown twins automatic.

**llms.txt**: auto-generated per locale. Apex `llms-full.txt` weighs in at 240KB and includes every product description, guide, and glossary entry.

**Customer reviews**: integrated Trustpilot for verified-purchase reviews; aggregate ratings emitted in Product schema.

## Results

**AI citations** (manual quarterly query audit, 25 fixed queries):

| Engine        | Q4 2025 citations | Q2 2026 citations | Delta  |
| ------------- | ----------------- | ----------------- | ------ |
| Perplexity    | 6                 | 21                | +250%  |
| ChatGPT       | 5                 | 17                | +240%  |
| Claude        | 1                 | 8                 | +700%  |
| Google AIO    | 3                 | 14                | +367%  |
| **Total/month estimate** | **~15** | **~52**     | **+247%** |

**EU organic traffic**:

- DE: 1,800 → 4,100 monthly sessions (+128%)
- IT: 1,400 → 3,200 monthly sessions (+129%)
- ES: 1,000 → 1,800 monthly sessions (+80%)
- **Combined**: 4,200 → 9,100 monthly sessions (+117%)

**Performance** (CrUX 60-day):

| Metric | Before (Squarespace) | After (Ordiko) | Delta |
| ------ | -------------------- | -------------- | ----- |
| LCP    | 2.6s                 | 1.2s           | -54%  |
| INP    | 235ms                | 120ms          | -49%  |
| CLS    | 0.09                 | 0.03           | -67%  |

**SEO surface coverage**:

| Surface                              | Before  | After                            |
| ------------------------------------ | ------- | -------------------------------- |
| 2026 Product schema fields            | Partial | Complete                          |
| llms.txt                              | No      | Yes (auto, per locale)            |
| Markdown twins                         | No      | Yes for all content              |
| AI crawler allow rules               | No      | Yes                              |
| Glossary with DefinedTerm schema      | No      | Yes (62 terms × 4 locales)        |
| Per-entity hreflang                   | Broken  | Correct                          |

## What we learned

Camille's three takeaways:

1. **AEO requires content, not just infrastructure.** Ordiko gave us the surfaces (llms.txt, Markdown twins). The 3x citation lift came from also publishing 62 glossary entries and 12 guides that AI engines could cite.

2. **Per-entity hreflang correctness fixed our German rankings.** We had been advertising IT and ES translations of FR-only products in Germany for years. Cleaning that up alone recovered 30% of our German organic.

3. **Performance compounds.** A 54% LCP improvement on mobile didn't just bump rankings — it lifted conversion 8% on PDPs.

## Quote

> "Squarespace served us well for five years, but as we expanded into Germany and Italy we hit two walls: their multi-locale handling required a third-party app that wasn't worth its cost, and they had nothing for AI search. Our customers in Berlin and Milan kept telling us 'Perplexity recommended your competitor when I asked for fountain pen paper'. Ordiko gave us native hreflang for four locales, llms.txt out of the box, and a clean way to publish our glossary of paper terminology that AI engines actually cite. The migration took a weekend."
>
> — **Camille Lefèvre**, Co-founder, Boutique Papier

---

*Boutique Papier runs on Ordiko Growth at boutique-papier.example.com (French default, German, Italian, Spanish locales).*