HARMAY sold luxury samples at pocket-money prices — the traffic came, the profits didn't

HARMAY (话梅) opened its first stores around 2017 and became China's most photographed beauty retailer: warehouse-style spaces, one-store-one-design interiors, and a draw that made it famous — the 小样, luxury-brand samples sold cheap, letting young shoppers buy a piece of a counter brand for pocket money. In early 2022 it raised nearly $200 million across its C and D rounds, at a valuation later reported near ¥6.9 billion.

The model's contradictions showed within two years of the raise. The samples brought crowds but not always purchases — reviews of closed stores concentrated on 'expensive' and 'only browse, don't buy'; the big international brands were traffic sold at zero margin or a loss, while the small brands that carried the profit sat in the corners. In early 2022, Shanghai regulators fined and confiscated more than ¥887,000 over sample products with incomplete Chinese labelling — the gray area of reselling items marked 'not for sale'.

The closures began: HARMAY's only Hangzhou store shut in July 2023, its Xi'an store in October, each about two years old. The company talked of strategic adjustment; the heavy one-store-one-design build-out could not be standardised or scaled, and the full-size prices sat awkwardly between e-commerce and brand counters. A store that was a destination people visited once could not carry a unit economy that needed them to buy.

The samples drew crowds but not purchases; closures began within two years of the raise.

Samples are a loss leader by nature — when the loss leader is what customers come for and the full-price product is what they compare online, the traffic monetises for someone else.

Selling items marked 'not for sale' borrows the brand owners' marketing budget without their permission — the compliance risk and the brand-relationship risk are the same risk.

A one-store-one-design flagship is a marketing asset with a retailer's cost base; it photographs well, and it does not replicate.

A destination store wins the visit and loses the repeat — if the draw is an experience and the margin is somewhere else in the store, measure what the visit converts, not how many people came.