How it works

How we find luxury that's actually worth your money.

Merino is not a search engine with better filters. The pieces that make it into our catalog have been through a five-stage evaluation — and the vast majority of what we look at doesn't make it. Here's exactly what that looks like.

Five-stage evaluation — click any stage to explore

Hundreds of searches.
Around the clock.

We monitor eBay continuously, running queries across every brand, garment type, and size combination we track. The goal at this stage is simple: breadth. We want to see everything before we start dismissing most of it.

Each query returns up to 250 listings. Across all garment types, that's tens of thousands of candidates entering the pipeline every sweep cycle.

Listings entering the pipeline
Coats & Jackets8,400+
Blazers6,200+
Sweaters5,100+
Dresses4,800+
Tops3,600+

Most listings fail
right here.

Before anything reaches our AI, every listing runs through a set of hard filters. Brand signal must be confirmed — in the title and item specifics, not just inferred. Condition must be pre-owned good or excellent. Size must fall within the accepted range for the garment type.

Roughly 90% of listings are eliminated at this stage. No AI calls, no false positives, no wasted time on your end.

Filter gates
Brand confirmed in title or item specifics — not just inferred
Condition gate — pre-owned good or excellent only
Size matched — US, EU/IT, and letter sizes all checked
Price window — calibrated per garment type and tier
Brand keyword in title only, not confirmed in specifics — rejected
Wrong size detected — rejected
Already evaluated this cycle — skipped

Valuation first.
Price second.

The first thing our model does is set the listed price aside completely. Its only job is to answer one question: what is this piece actually worth on the secondary market?

It works through a strict signal hierarchy, from most to least reliable. If it can't find a real comparable, it says so — and the confidence score reflects the gap. A null is always preferable to an invented number.

Signal hierarchy — in order of reliability
1
Product line + channel
Mainline vs. wholesale exclusive vs. collab. Diffusion lines valued separately.
2
Fabric + material spec
Vicuña, Storm System, baby cashmere — confirmed in item specifics.
3
Construction details
Country of origin, canvassing, hand-finishing signals.
4
Comparable sold listings
Specific historical transactions on eBay or Grailed. Never fabricated.
5
Brand MSRP — last resort
Used only when nothing else is available. Confidence score widened accordingly.

Scored against
where things actually trade.

With valuation locked, a second pass asks: how good is this deal for a buyer who follows this market closely?

A 40% discount off retail sounds impressive until you know that piece consistently trades used at 55% below retail. You'd actually be overpaying. Our score reflects the secondary market reality — not the sticker.

The same coat — two benchmarks
Brunello Cucinelli cashmere blazer
Original retail
$4,800
Secondhand avg
$2,400
This listing
$1,600
33% below where comparable pieces have sold → Deal score 8/10 — Exceptional

Everything you need.
Nothing you don't.

Every deal that makes it through gets four things alongside the score. No urgency copy. No countdown timers. If a deal scores an 8, you'll understand exactly why — and whether it's actually right for you.

Deal score
1–10, benchmarked to secondary market transactions. Not retail.
Deal context
One sentence: what this piece is and where it sits relative to similar listings.
Score explanation
2–3 sentences written like a tip from a friend who follows this market — the value, made plain.
!
Buyer callout
One practical thing worth knowing before you buy. Left blank if there's nothing genuine to add.

Hundreds of searches. Around the clock.

We monitor eBay continuously across every brand, garment type, and size we track — tens of thousands of candidates entering the pipeline every cycle.

Listings entering the pipeline
Coats & Jackets8,400+
Blazers6,200+
Sweaters5,100+
Dresses4,800+
Tops3,600+

Most listings fail right here.

Brand confirmed, condition verified, size matched. Roughly 90% of listings are cut at this stage — before any AI is involved.

Brand confirmed in title and specifics
Condition: pre-owned good or excellent
Size matched across US, EU/IT, letter
Price window calibrated per garment type
Brand keyword only in title — rejected
Already evaluated this cycle — skipped

Valuation first. Price second.

The listed price is set aside. The model works through a hierarchy of signals to determine what the piece is actually worth on the secondary market — and flags honestly when it can't find a confident answer.

Signal hierarchy
1
Product line + channel
Mainline vs. diffusion vs. collab.
2
Fabric + material spec
Vicuña, Storm System, baby cashmere.
3
Construction details
Origin, canvassing, finishing signals.
4
Comparable sold listings
Real transactions. Never fabricated.
5
Brand MSRP — last resort
Confidence score widened accordingly.

Scored against where things actually trade.

A 40% discount off retail means nothing if that piece consistently trades used at 55% below retail. The score reflects secondary market reality, not the sticker.

The same coat — two benchmarks
Brunello Cucinelli cashmere blazer
Original retail
$4,800
Secondhand avg
$2,400
This listing
$1,600
33% below where comparable pieces have sold → Deal score 8/10 — Exceptional

Everything you need. Nothing you don't.

Every deal that makes it through gets four things. No urgency copy. No countdown timers. If it scores well, you'll understand exactly why.

Deal score
1–10, benchmarked to the secondary market.
Deal context
What it is and where it sits relative to recent sales.
Score explanation
Written like a tip from a friend who follows this market.
!
Buyer callout
One practical thing worth knowing. Blank if nothing genuine to add.
Proof it works

The deals that
didn't wait.

The best measure of whether something is worth owning is whether someone bought it. Every week, Merino users claim deals that scored 6 or higher — real buyers, real prices, real secondhand luxury — before they leave the pipeline.

These aren't page views. Browse what's sold to see what the pipeline catches.

Loading…
deals sold in the last 7 days
See every recently sold item
What we never do
Fabricate comparable sold prices
Write urgency copy — "won't last," "act fast"
Score against retail when secondary data exists
Surface a listing we can't confidently value
Let brand prestige substitute for deal quality