Tutus, in five years: Pricing second-hand fashion

A 2022 Jacquemus wool blazer listed on a peer marketplace sits stagnant at €480 for eight months. Two miles away, an identical jacket sells in nine minutes for €190. Neither price reflects asset value; both reflect human guesswork. The resale market for apparel, projected by ThreadUp to hit $350 billion globally by 2027, operates on the economics of a flea market: high friction, opaque information, and extreme bid-ask spreads.

Most platforms attempt to solve liquidity by spamming notifications or forcing sellers into automated price drops. This treats the symptom rather than the systemic flaw. Apparel pricing fails because peer platforms lack structural data on garment lineage, material composition, and local seasonal demand velocity.

The Failure of Historical Averages

Existing platforms suggest prices by averaging historical ask prices. This creates a feedback loop of bad data. If ten sellers list ruined fast-fashion coats for €80 because they paid €120 six months ago, the platform algorithmically recommends an €80 anchor to the eleventh seller. The item inevitably rots in catalog search results.

To price second-hand clothing efficiently, a platform must treat a garment as a yield-bearing physical asset subject to continuous depreciation curves. A 100% heavyweight cotton hoodie from a niche Japanese brand holds value along an entirely different mathematical trajectory than a polyester-blend sweater from a high-street chain, even if both retailed originally at €150.

Within Tutus, WEVONE's second-hand fashion universe, our ambition—currently running as a restricted beta across select European markets—is to convert unstructured listing images into structured financial assets with automated valuation corridors.

How Mia Calculates Garment Valuation

Consider a concrete scenario. A seller in Lyon uploads three smartphone photos of a navy Dries Van Noten wool coat purchased in 2020.

Instead of asking the seller to fill out ten drop-down menus, Mia processes the raw input against five distinct vectors:

1. Material Lineage: The system parses label photography, recognizing a 90% virgin wool, 10% polyamide mix. Polyamide content reduces natural pill resistance over four years by a calculated 14% compared to pure worsted wool.
2. Defect Quantification: Computer vision flags micro-fraying along the inner cuff edge and minor creasing on the lapel, assigning a physical integrity score of 0.78.
3. Macro Demand Velocity: Historical clearance data indicates that tailored menswear from this specific brand retains 42% of original retail value in Northern Europe during autumn months, dropping to 27% in spring.
4. Local Micro-Market Liquidity: WEVONE’s cross-universe ledger identifies active search queries for structured outerwear within a 300-kilometer radius.
5. Seller Accuracy History: The seller's historical margin of error when reporting item conditions adjusts the internal confidence interval.

Mia outputs a non-negotiable liquidity corridor: €210 for a guaranteed sale within 72 hours, or €245 for an estimated 21-day holding period. The seller chooses their velocity curve rather than pulling a random number out of thin air.

The WEVONE Infrastructure Advantage

The valuation engine does not operate in isolation. It relies directly on WEVONE’s core architectural mechanics. When a transaction occurs in Tutus, funds enter the platform's transactional escrow, held securely until the buyer verifies the garment's condition against Mia's original vision analysis within a 48-hour inspection window.

Seller reliability is tracked through a unified platform contribution score rather than siloed star ratings. A seller who accurately describes garment wear in Tutus earns reputational weight that directly lowers transaction fees when they offer services in Mission or rent out equipment in Tools. Reputation is cross-universe capital, making accurate listing behavior rationally self-interested.

Present Realities and Technical Limits

Strategic honesty demands acknowledging where this system currently breaks. Computer vision cannot smell. It cannot evaluate texture elasticity, lining odors, or micro-tears hidden beneath heavy ironed seams. A seller using strategic studio lighting can artificially inflate a garment’s visual integrity score by up to 20%.

Furthermore, our training sets for vintage garments produced prior to standardized digital care tags remain sparse. When handling unbranded vintage or altered garments, Mia’s valuation corridor widens significantly—sometimes offering a variance of €40 to €120—rendering the automated recommendation far less useful than human intuition.

We are clear about what is live versus what remains a bet. Automated condition grading from single-angle photography is in early testing. The dynamic pricing corridor for high-volume, modern designer categories is live in our Western European test cohorts. Fully automated cross-border liquidity matching remains a three-year ambition.

Beyond Closet Clearance

The long-term objective of Tutus is not to act as a digital attic sale. It is to reduce the transaction costs of second-hand commerce until pre-owned apparel acts with the fluid pricing precision of financial commodities. When buyers know precisely what an item is worth to the exact Euro—and can liquidate it six months later at a predictable decay rate—the distinction between owning clothes and leasing them disappears.