Origin + cocoa% storytelling, how do you make bean-to-bar PDPs actually sell?

Single-origin Madagascar 70% reads differently than “dark chocolate.” The structured-data approach beats the marketing-prose approach 2:1 on conversion. The Magento setup:

• Custom product attributes for origin country (Madagascar / Venezuela / Ecuador / Ghana / Peru / Tanzania), cocoa percentage (60% / 70% / 80% / 85%), processing style (stone-ground / conched / raw), farmer cooperative (e.g. Akesson’s Estate, Hacienda San José), harvest year, fermentation days, roast profile.
• Hyvä PDP layout, origin map (Mapbox or static SVG of the bean origin), tasting notes (citrus / cherry / earth / tobacco / molasses), pairing suggestions (red wine / espresso / aged cheese), farmer story + photo. Mast Brothers built the visual language; everyone else copies.
• Layered nav filters at category, cocoa % (60-70 / 70-80 / 80+), origin (Madagascar / Venezuela / Ecuador / Ghana), certification (Fair Trade / Rainforest Alliance / Organic / Direct Trade). The data is in the EAV attribute, the filter is automatic.
• Schema.org Product with custom additionalProperty entries for origin + cocoa %, gets picked up by Google rich results for “Madagascar 70% chocolate” head terms.

Conversion lift on bean-to-bar PDPs runs 1.4-2.1x when origin storytelling is structured (filterable, schema-marked, visually rendered) vs. buried in prose. Mast Brothers, Dandelion, TCHO, and Vosges all ship this pattern. Hu Kitchen (CPG-scale) uses the same attribute model.