Mariam Naficy Keeps Rewriting the Way Taste Becomes a Product

The first month of Minted produced almost nothing. Mariam Naficy had raised $2.5 million, signed exclusive stationery brands and built the machinery of a respectable online store. The virtual doors opened in 2008. Then silence. Not one sale arrived from the branded assortment. Off to the side, though, an experiment she had protected with roughly $100,000 was making a faint sound. Independent artists had submitted 60 designs to a competition. Shoppers voted. A winning card sold once, then another sold the next week. A tiny, strange corner of the business was the only corner behaving like a business.

Naficy followed it. Minted stopped trying to be a digital shelf for established stationery and became a system for finding talent outside the usual gates. Creators submitted. The community judged. Minted manufactured and distributed the winners. The pivot looks clean in retrospect, as successful pivots tend to do. In the room, it required admitting that almost all the money had gone toward the wrong thesis while the useful idea was being assembled at night with a programmer in Oregon.

The signal was not loud. It was one transaction a week.

This is the recurring pattern in Naficy's work. She does not merely put a familiar category online. She changes who gets to decide what should exist there. Eve.com asked whether prestige cosmetics, then guarded by department-store counters, could be bought on the web. Minted asked whether a crowd could surface stronger design than a conventional buying team. Arcade, her current company, asks whether the shopper can begin before the product itself exists.

The mall was an education

Naficy grew up between countries, the eldest daughter of an Iranian father and a Chinese mother who met as graduate students at Georgetown. Her father's work as a development economist took the family through Africa and the Middle East. At nine, during the Iranian Revolution, she left Iran with her family and resettled in the United States. Years later, speaking to a stadium of graduates, she distilled that instability into a private promise: “I am always going to remake myself.”

The grand education ran beside a more ordinary one. As a teenager, she read fashion magazines, spent time at malls and wore a lot of makeup. She later gave the habit a sharper name: consumer psychology. At Williams College she studied political economy. At Stanford's business school, she wrote a practical career guide called The Fast Track. Between the degrees came investment banking at Goldman Sachs, an unlikely but useful apprenticeship in the language of capital.

In 1998, while finishing her MBA, Naficy co-founded Eve.com with Varsha Rao. Buying premium cosmetics online sounded implausible to many investors and brands. Eve nevertheless reached $10 million in first-year sales and was sold in 2000 in a deal reported around $100 million to $110 million. Naficy then worked in digital movies and led e-commerce at The Body Shop. By the time she started Minted, she understood both the seductive speed of venture capital and the cost of committing too early to a polished plan.

The business hiding inside the mistake

Minted's competition did more than pick cards. It offered an answer to a cultural question: how much creative ability never reaches a shelf because its maker lacks the right school, city or résumé? The company's voting system let trained and self-taught artists enter the same field. Winning designs earned production, distribution and continuing commissions. The customers were no longer a market surveyed after the buying decision. They became part of the buying decision.

That architecture made Minted unusually elastic. Stationery led to art, textiles and home goods because the underlying asset was not paper. It was a community and a selection mechanism. Naficy once said she thought she had founded a cash-flow stationery retailer, only to discover a content studio. The business showed her what it was. By 2019 it had about 400 employees and revenue in the hundreds of millions. A 2026 account put annual sales near $300 million.

The human part took longer than the model diagram. Only five people attended the first artist meetup. Naficy said this year that the community had reached 20,000. The contrast is a tidy reminder that a marketplace starts as a room, not a flywheel. People need a reason to show up before the network effects can be graphed.

Her operating choices also changed. Eve had taught her what abundant financing could buy: talent, attention and speed. Minted began more deliberately. The company reached cash-flow break-even before raising heavily to build an executive team and expand. Naficy led it for roughly 15 years, shifted into the executive-chairman role in 2022, and kept the artist community close enough to return to its gatherings and its founding argument: unknown creators can win when the work is judged directly.

From a prompt to a package

Arcade begins with a blank that ordinary e-commerce cannot tolerate. A shopper types an idea for a ring, rug, scarf, ceramic or pillow. A model turns the words into visual options. The shopper revises. Then the dreamy part meets a set of stubborn nouns: metal, wool, dimensions, labor, price, shipping and quality. A vetted maker has to produce an object that resembles the image. Arcade takes on the connective tissue, pairing generative design with specialized models, product rules, real-time pricing and a network of workshops.

Naficy calls this the end of the null search result. In the catalog era, a precise search for a chocolate-brown pillow or a pendant shaped around a private memory could end in compromise. In her proposed creation era, the absence of inventory becomes the beginning of a design session. The store does not answer “no results.” It asks what should change.

“This is not search-and-retrieval AI. This is 100% new product generated.”

The distinction keeps Arcade on the difficult side of generative AI. Producing a persuasive picture is cheap. Producing a physical good that earns a satisfied second look is expensive and operational. Price must appear before desire cools. Materials limit form. A maker needs an economically sensible order. The delivered piece has to honor the screen. The more interesting technology may therefore be “prompt to price,” the system that translates a generated design into a number a workshop can accept.

Arcade started in jewelry partly because a three-dimensional product would force the system to confront those constraints. Jewelry is fragmented, personal and naturally shareable. If the machinery worked there, flatter categories would not get harder. The company later expanded into home goods and partnered with makers and brands including Christofle, Bitossi Ceramiche, Cabana and Salam Hello. It introduced Maia, a conversational design assistant Naficy said she personally co-authored with two colleagues, to ease the anxiety of facing an empty prompt box.

What builders can steal

Naficy's career is often packaged as instinct, but instinct alone is not the transferable part. Her Minted hunch survived because she ring-fenced enough money to build it. The competition created observable behavior. When the established products failed and independent work moved, she had comparative evidence. A founder cannot copy her taste, but can copy the arrangement: reserve a small budget for the idea that sounds hard to defend, make it real enough for customers to react, then treat the reaction seriously.

There is another useful move in the sequence from Eve to Minted to Arcade: begin with a category narrow enough to teach the system. Cosmetics exposed whether shoppers trusted premium retail online. Stationery revealed whether unknown designers could beat known labels. Jewelry made AI confront the physical world in three dimensions. Each wedge looked smaller than the intended future, but contained the hard question of the larger business.

Arcade's current ambition extends beyond helping someone order a sentimental necklace. Naficy has described a marketplace where creators can design and sell product lines, artisans gain new demand, and businesses license the underlying creation tools. In early 2026, she located the company's work in a middle layer “where taste and AI meet,” arguing that broad foundation models need specialized knowledge of texture, opinion and craft. By summer, Arcade was hiring an AI product manager and a chief of staff in San Francisco for its next stage.

The bet remains open. Custom commerce can be thrilling in a demo and unforgiving after delivery. Yet the continuity in Naficy's work makes the wager legible. For nearly three decades, she has looked at a shelf and asked who has been excluded from choosing what goes on it. First the web admitted a new kind of purchase. Then a voting community admitted a new kind of designer. Now a prompt might admit a new kind of product - one that enters the world because a shopper could describe it before anyone stocked it.