BRENNEN

Industries / Retail, hospitality and personal services / Retail and e-commerce

Thrift and resale

Thrift and resale collects used clothing and household goods, as gifts or by buying them, sorts and prices each item, and sells it again in shops, online or abroad in bales.

1897

The Salvation Army in the United States formed its first collecting brigade: residents of a men's shelter gathered castoffs in exchange for a meal and a bed, and the goods soon needed a place to be stored and sold.

salvationarmyusa.orgdaily.jstor.org

35

Milestones achieved across 6 domains.

9

To watch for: announced, not yet confirmed outside.

12

Open deliverables. Nobody has shipped them.

In one minute

Where it started

People have sold used clothes for centuries; JSTOR Daily notes that second-hand shops thrived in Elizabethan England. The thrift store came from US charities around 1900: the Salvation Army (1897) and Goodwill (1902) had poor workers collect and mend donated goods, then sold them to pay for their programs. Charity shops spread in Britain from the 1940s, followed by for-profit thrift chains, shops that buy from the public, and, from 1995, websites where anyone can sell.

Where it stands

ThredUp, a US resale company, and the research firm GlobalData put the world's second-hand clothing market at about $393 billion by 2030, roughly 10% of clothing spending; the figures were commissioned by a reseller. Much of what is donated never sells where it is given: by the European Environment Agency's count, EU exports of used textiles tripled to almost 1.7 million tonnes in 2019 while the price per kilogram fell, and large European sorting firms went bust in 2024 and 2025. Apps such as Vinted are growing fast, while UK charity chains close shops as rents and wages rise.

What is open

Many of the trade's basic numbers are private or disputed: what share of donations is sold, how long items take to sell, how much of an exported bale is waste, and whether buying used replaces buying new. The open jobs are public datasets and controlled trials that would settle those numbers, and open tests of the tools that now grade, price and list one-off items from photos.

Domains

AchievedWatch forOpen

Open in thrift and resale

3 of 12 open deliverables

Open · Donations and sourcing

Donation-to-shelf rates, compared

A free public table showing, for at least 20 large charity and thrift retailers in several countries, what share of donated goods each sells in its own shops, what share it sells to textile merchants or exporters, and what share goes to waste or burning, with the unit (weight, items or money), the year and the page cited for every figure, and blanks marked where nothing is published.

Built with what exists

Annual and impact reports; charity regulators' registers, such as the Charity Commission for England and Wales, ProPublica's Nonprofit Explorer for US Form 990 filings and Australia's ACNC register; a large language model to pull candidate figures out of PDFs, with a person checking every figure against the cited page; a shared spreadsheet.

Credit · unclaimed

Open · Sorting and grading

Test sorting models abroad

A public baseline model, trained on the Swedish Clothing Dataset for Second-Hand Fashion, that predicts each garment's route (reuse, export or recycling) and its stains and holes, plus a new test set of 1,000 garments photographed and labeled the same way at a sorting room outside Sweden. Together they show how much a sorting model loses when it meets another country's donations.

Built with what exists

The dataset from Zenodo (CC BY 4.0); open image models such as SigLIP or DINOv2 with PyTorch; a no-training baseline from a model that returns probabilities for fixed-answer questions, such as Gemini Decision 1 (see the Note "A thousand questions, one pass"); Label Studio for labeling; a phone on a stand over a gridded table, as in the Swedish set.

Credit · unclaimed

Open · Pricing

Pricing tools, tested in shops

A six-week trial in a charity shop in which alternate incoming garments are priced either by a tool or by staff as usual, comparing the share sold within four weeks, money taken per garment and seconds spent pricing each one. The tool can be an open-source helper that reads the brand label from a phone photo and looks up the shop's own past sale prices for that brand and category, or a commercial app. The team handles no money; the shop sells as usual.

Built with what exists

A phone camera; a vision-language model that reads labels, such as Claude or Gemini through its API, or open OCR software such as PaddleOCR; the shop's sales records; Python; a label printer.

Credit · unclaimed

Where it sits