Ines Madrid
Ines Madrid
A direct-to-consumer women's fashion store — built, customised, marketed, and pivoted across two markets.
Context
Ines Madrid is a Shopify-based D2C boutique I run end-to-end: theme development, product curation, paid acquisition, fulfilment ops. It started targeting Hungary and pivoted to Spain mid-2025 after reading the data.
The pivot
I launched into Hungary first — translations done, currency localised, ads tuned. CPMs were cheap but conversion was thin. After a month of data, the diagnosis was clear: the SKU mix didn't match Hungarian buying patterns, and market depth for premium women's fashion at our price point was smaller than my model suggested.
I pivoted to Spain. Same products, different market positioning. Restructured the catalogue around what Spanish competitors (Elara Madrid being the closest reference) were ranking for. Re-translated all copy from PT to ES with cultural adaptation, not just literal translation.
What I own
- Theme customisation — Impulse theme, Liquid edits to product page, footer, navigation, plus a fixed multi-language menu bug via the Translate & Adapt app
- Product curation — sourcing through Kungfubuy, with selection driven by competitor research and ad-test data
- Paid acquisition — Meta Ads campaigns (resolved pixel conflicts and placement asset errors during ramp-up); Google Shopping planned as second channel
- Conversion infrastructure — Meta Pixel integration, abandoned-cart flows, post-purchase email
- Strategic decisions — market selection, pricing, when to kill a SKU
Key decisions
Shopify over a custom Next.js storefront. I have the skills to build it from scratch. I deliberately didn't. Storefront-as-code only matters once you have product-market fit — until then, every hour on theme infrastructure is an hour not spent on what actually moves the needle (ads, SKUs, copy).
Pivot fast on weak data. A month of underperformance in Hungary was enough. Sunk cost is the most expensive cost in e-commerce. Re-translating to ES took a weekend. Continuing to grind on HU would have taken months to break even.
Meta + Google Shopping as the channel mix. Meta for discovery, Google Shopping for intent. TikTok is on the roadmap once unit economics stabilise.
Lessons
The technical skills I bring (debugging the language menu bug, pixel troubleshooting) are useful but not the bottleneck. The bottleneck is product-market fit, and that's a marketing and curation problem, not a code problem. I find that perspective valuable for any engineering role I take — most engineers don't think this way.