How we work

How Worthmore prices things

Worthmore is a pricing tool for people selling second-hand on UK platforms. We help you figure out what something might be worth before you list it. This page explains how we get to the prices we show — what informs them, what doesn't, how confident we are in any given number, and how the data improves over time.

The platforms you sell on will often suggest a price too. Those suggestions might feel geared toward a quick sale, which can leave you with less than you might otherwise have got. Worthmore exists to offer a second opinion — your item might be worth more than the platform is telling you.

We're putting this in plain English because you should be able to read it without a methodology degree, and because if a journalist or a regulator asked us how we sourced our pricing data, this is the page we'd point them to.

What Worthmore is — and isn't

Worthmore is a guide, not a guarantee. When we tell you a Levi's 501 in good condition might sell for £15–25, that's our best read of the UK second-hand market based on the research we've done. It isn't a contract, a valuation, or a promise that anyone will pay that price. Markets move. Your photos matter. Your timing matters. Your description matters. The number we give you is a starting point — informed, considered, and yours to use or override.

We're not a financial product, an investment tool, or a tax advice service. If you're trying to work out whether your selling activity crosses an HMRC threshold, please talk to someone qualified. That isn't our lane.

How prices are researched

Worthmore's database is supported by research using a large language model, following a methodology we wrote and maintain. For each combination of category, brand, condition, and — where it matters — size, the model searches the public web for evidence of what items like yours have actually been selling for in the UK, then combines that evidence with a methodology-based sense-check before producing a price band.

What we deliberately don't do is query any single resale platform's sold-listing pages directly. We don't scrape Vinted, Depop, eBay, or Vestiaire. We don't pull data from any platform's API. We don't take feeds from third-party data brokers who do those things on our behalf. Beyond the terms-of-service problem with most platforms, single-platform extraction would be a worse approach in its own right: it would give us a snapshot of one market, not a read on what something is actually worth across the UK resale economy.

What sources inform our pricing

The research draws on a diffuse mix of public web sources, including:

Open web discussions — public threads, communities, and forums across the open internet where UK sellers and buyers talk openly about what they paid and what they sold for.
Trade press and editorial coverage — fashion and retail industry outlets covering the resale market.
Brand resale programmes — official pre-owned channels run by brands themselves, which publish their own prices.
Fashion forums and review sites — communities that discuss specific items and how their resale value moves over time.
Charity shop online listings — for high-street and unbranded items, where charity shops have become a meaningful comparable.
Retailer outlets — to anchor the "new" end of the range against current RRP and ordinary discount levels.

No single source dictates a price. The model is instructed to weigh multiple independent source types, and we record what source types informed each estimate.

How confidence is calibrated

Every price band carries a confidence score between 0 and 1. We calibrate it honestly. The ceiling for any LLM-researched row is 0.90, not 1.0 — nothing we produce gets to claim certainty.

In broad terms:

0.82–0.90 — four or more independent source types converge on a tight band, with recent references (within the last six months).
0.68–0.81 — two or three independent source types converge, or a single strong recent source.
0.50–0.67 — single source type only, wide variance across sources, or limited recent references.

When a Worthmore price comes back lower-confidence, that isn't us being modest — it means the research surfaced less to go on, and you should weight your own judgement accordingly.

The methodology sense-check

After researching the market, the model runs a constraint check against what the item should be worth based on its retail price and brand tier. As a rule of thumb:

Luxury items typically retain 35–60% of RRP at the "new with tags" point on the second-hand market.
Premium items typically retain 25–45% of RRP.
High-street items typically retain 15–30% of RRP.

If the researched price band sits more than 25% outside this methodology range, the model is required to reconsider — tightening the band to methodology, widening it to acknowledge both signals, or lowering its confidence score. The point is to stop a noisy patch of search results producing a price band that doesn't make sense against the basic economics of the item.

How the dataset improves over time

This feature isn't active in the current version of the tool. When it launches, sellers will be able to tell us what they actually sold an item for after using Worthmore. When enough sellers report back on the same kind of item — our threshold is five submissions — we'll blend their real sold prices into the recorded band, weighted 60% toward what sellers actually achieved and 40% toward the original research. Prices updated this way will be marked distinctly from purely research-derived estimates.

This is the part of Worthmore that gets more useful the more people use it. Every "I sold mine for £18" is data the next person benefits from.

What we don't do

To be explicit:

We do not scrape Vinted, Depop, eBay, Vestiaire, StockX, GOAT, or any other resale platform.
We do not access any platform's API or any paid commercial data feed for sold-listing data.
We do not present a single platform's sold listings as our source.
We do not give financial, investment, or tax advice.
We do not guarantee any price.

If a price looks wrong

If we've got something materially off — a Mulberry Bayswater priced like a high-street tote, a pair of Levi's underpriced for the size band — please tell us. The gap detection picks up at-scale problems on its own, but a direct flag from a real seller is the fastest route to a fix. You can reach us at hello@worthmore.uk.

We'd rather know, and we'd rather get it right.

This methodology document will be updated as the approach evolves. Updated May 2026.