Culture Is Not Decorative: What Mamanfari Taught Me About Product Work

When I first started building mamanfari, I described it too simply. A recipe site with bilingual content and a strong visual identity. That framing missed the actual problem almost entirely.

The hard part was not translating recipes from Persian to English. It was building a product that could carry cultural meaning without flattening it — and figuring out what that even meant in terms of architecture, content structure, and interface decisions.

The translation problem was the wrong frame

Take advieh — the Persian spice blend. You could translate it as "Persian spice mix" and move on. But advieh is not one thing. There is advieh for rice (lighter, more floral — rose petals, cardamom) and advieh for stews (warmer, deeper — cinnamon, coriander, black pepper). Every family has their own ratio, passed down without a written recipe. Some regions use saffron as a base note; others don't touch it.

A parenthetical translation doesn't carry any of that. Neither does a simple ingredient list.

Or take piaz dagh — caramelized onions. Every khoresh starts with them. Rushed onions produce flat flavor; the process takes 15–20 minutes minimum. That's not a tip or a note at the bottom of the recipe. It's a structural fact about how Persian stews work. If the product presents it as an optional step, it's quietly teaching the wrong thing.

These are not translation problems. They are product decisions about how much context survives and what gets assumed.

Regional variation broke my first information model

The bigger surprise was how much the same dish differs across regions.

Ghormeh sabzi — probably the most well-known Persian stew — has a completely different character depending on where you are and who you learned it from. Some cooks use tomato paste as a base; others reject it entirely and double down on dried lime (limoo amani) and verjuice (ab-ghooreh) for the sour depth. The herb ratio shifts. Bone-in cuts versus boneless changes the cooking time and the final texture.

None of these are wrong versions. They are all real, all credible, and all specific to a family or a region.

That broke my first model immediately. I had assumed a recipe was a single authoritative document. It isn't — at least not in this cuisine. A dish like ghormeh sabzi is more like a schema with regional and family-level variations, each carrying its own logic. The product needed to represent that without forcing a single "correct" version.

That is a data modeling problem, not a copy problem.

The glossary and technique layer changed the architecture

Once I accepted that context was part of the product, the architecture started changing.

A glossary became necessary — not as a reference appendix, but as a linked layer inside the recipe experience. When a user encounters limoo amani or shanbalileh (Persian fenugreek), they need to understand not just what it is but why it's there and what it does to the dish. That understanding changes whether they'll actually cook it.

Technique pages became necessary for the same reason. The process of building a khoresh follows a consistent structure across dozens of stews: fat, onions, paste, meat, liquid, sour element, long simmer. A user who understands that pattern can navigate any new stew. A user who doesn't has to relearn the process from scratch every time. Surfacing that shared structure — not just in a single recipe but as a navigable piece of content — was a product decision that changed how the whole content system worked.

Shopping and pantry flows followed from there. An ingredient like dried Persian limes or saffron is not available at every grocery store. If the product doesn't help users source what they need, the cultural authenticity becomes inaccessible in practice.

Implementation choices became editorial choices

This is the part I think about most now.

Every structural decision on mamanfari had an editorial consequence. How recipes relate to techniques. Whether variations are presented as alternatives or as divergent traditions. How the glossary surfaces during a recipe — inline, on demand, or as a parallel layer. Whether a search returns a single result or a family of dishes.

If the structure is wrong, the product quietly teaches the wrong thing.

If the glossary is missing, the product assumes too much.

If variations are collapsed into a single recipe, the cuisine feels more uniform than it is — which is its own kind of distortion.

That is why I don't think of mamanfari as "a recipe website I built." It was a lesson in what happens when the software itself becomes part of how culture is interpreted. The system design is part of the editorial voice. You are not only deciding how the interface behaves — you are deciding what context survives, what gets simplified, and what kind of trust the product is able to earn.

That is the kind of product engineering work I want to keep doing.