DETERMINANT China market read · July 2026

The market moved.
We are on the wrong side of it.

Youngor asks
¥780
People pay
¥310

The brand our own staff cross-shop most does not sell a shirt at its own tag price. It has no non-event price at all. That is not a promotion. It is what the tier below RMB 350 has become.

Prepared by Alfred Ho · Operating partner, Brands & Retailing
What we collected

Forty brands, priced by hand, every number dated and rated.

No brand tracker, no syndicated report. We priced the market ourselves, garment by garment, on the channels where the money actually changes hands, and recorded how much we trust each reading.

40brands on the China and APAC maps, chosen from a staff nomination survey and Dee's own brief
204price observations, each with the SKU, the channel, the date it was seen and a confidence rating
5garments per brand where we could get them: shirt, polo, trousers, jacket, tee

Two prices per garment, never one. The ticket is what the tag says. The take-home is what you actually pay after every stacked coupon. The gap between them turned out to be the finding.

Sources: hand collection 19-20 July · web sweeps 27, 28 and 29 July · in-app verification 28 July · staff survey n=158 · DET CN financial report 202606 · nine brand deep dives
Who our own people shop

Our customer's consideration set is the collapsed tier, plus one aspiration.

We asked 158 colleagues which menswear brands they and their customers actually buy. The answer is not the brands we benchmark against in strategy documents.

Youngor 雅戈尔78
Uniqlo 优衣库75
HLA 海澜之家53
Ralph Lauren41
JoeOne 九牧王34
G200032
Nice Rice 好饭1

Mentions across 158 respondents. Nice Rice's single mention came from APAC.

Four of the top six
sell their shirts below RMB 350, in the tier where the tag is fiction.
Ralph Lauren at 41
is the aspiration pole. Esquel manufactures for them. They sit at 1,790.
Zero of 121
mainland colleagues named Nice Rice, the design-led brand we had been asked to study. Our China organisation cannot see it at all.
Zero APAC mentions
for any mainland domestic brand. The two markets do not share a consideration set.

This is what sets the roster for everything that follows. We studied the brands our customer actually buys, not the brands we would prefer to be compared with.

Source: DET brand nomination survey, two waves, 9-14 July 2026, n=158 · 121 DET China including frontline, 13 APAC, 13 support, 5 PYE, 5 global branding, 1 product
The finding

The market split at RMB 350.

Below that price, almost nobody charges their own ticket. Above it, almost nobody discounts. This is not a segmentation we chose. It is where the behaviour changes.

77%average share of ticket actually paid, across the 15 China brands whose shirt sells below RMB 350. Five are under 70%.
94%average share paid, across the 15 whose shirt sells at or above 350. Not one is under 80%.

Below 350 the event price is the price. Youngor's shirt moved from 355 to 310 across two consecutive event windows in one week. There was no week in between at full price.

Source: resolved anchor cells, 30 brands with a China shirt price · hold = take-home divided by ticket · observed 20, 27 and 28 July 2026
Why it happens

Price integrity is a channel decision before it is a brand decision.

The same brand, the same season, the same garment holds its price in one channel and collapses in another. Choosing a channel mix is choosing a price integrity.

BrandHolds hereHoldCollapses hereHold
Youngor 雅戈尔商场同款 department line95%Tmall flagship40%
LILANZ 利郎JD line~100%Tmall flagship38%
Arc'teryx 始祖鸟官方旗舰店85-100%百亿补贴 channel67%
比音勒芬core SKUs, three channels95%百亿补贴 tier51%
DETERMINANTJD 常卖92%JD 真5折 events50%

The nationals run synchronised weekly event windows. On 28 July all four read the same dates, 26 to 31 July.

This is the answer to the question Dee asked in July about which channels serve the gap. The channel does not just reach the customer. It sets what the customer will pay.

Source: own-channel and competitor cells verified 27 and 28 July 2026
A play we are not running

Some brands hold the hero and dump the tail. Our ladder is flat.

We expected price integrity to decay steadily across the wardrobe. It does not. What happens instead is a deliberate split: one garment defended at full price, another used to move volume.

BrandShirtTrousersGap
霞湖世家100%50%50 pts
Lululemon100%51%49 pts
比音勒芬87%51%36 pts
LILANZ 利郎80%47%33 pts
DETERMINANT80%82%2 pts

Across 20 China brands holding all three of shirt, polo and trousers, the shirt holds higher in 11 and lower in 9. There is no general decay. There is a strategy that some brands run and others do not.

80 / 73 / 82 / 76 / 100DETERMINANT's own hold across shirt, polo, trousers, jacket and tee. Unusually flat.

We discount everything a little rather than defending one thing completely. That is either discipline or a lever we have never pulled, and the room should decide which.

This is the category cut Keith asked for in July, answered. It is also the first time we can see our own wardrobe priced against the market garment by garment.

Source: resolved anchor cells per brand per garment, China map, observed 20 to 30 July 2026
Where we actually are

Our anchor price is a clearance price.

The 318 we quote as our position is the 301 清仓 line on Tmall. It is the shirt that sells, at 900 units a month, inside a subsidy window that has been running since June 2025 and is scheduled to run to July 2028.

¥358-398what the tag says
¥268-318what people actually pay day to day, Tmall
¥159the floor reached in a 真5折 event inside the last 30 days

Dee's instinct in June was that DETERMINANT will not go cheaper than it already is. The channel has already made that decision for us. We are leaking downward while talking upward.

The good news
our JD 503 premium shirt holds 598 at 100% of ticket. The upward move already exists inside the house, in one line.
The asset
the JD 301 shirt carries more than 20,000 reviews at 99% positive. Our strongest social proof sits on the clearance anchor.
Volume truth
the polo at 157 and the tee at 238 each move 3,000+ a month. The woven shirt moves 55 to 900.
So
the shirt is our story and the knits are our volume. Any plan that says "own the shirt category" has to say which of the two it means.
Source: DETERMINANT own-channel cells, Tmall and JD, verified 28 July 2026
What the margin pays for

Our margin improved. What we spend it on is the problem.

Gross margin is up four points in eighteen months. But almost all of it goes back out through the channel, and it goes to the one mechanism the market says does not hold a price.

66.9%gross margin, China, first half 2026. Up from 62.9% for full year 2025.
55.1%total channel expenses as a share of sales, same period
11.8points of spread between the two. What is left to run the business on.
 Gross marginSelling costSpreadMediaPhysical presence
比音勒芬, FY202477.0%40.3%36.74.9%29.2%
DETERMINANT China, H1 202666.9%55.1%11.8~24%~19.5%

The most disciplined premium player in our category spends 4.9% on media and 29.2% on being physically present. We spend roughly the inverse. And we are closing stores.

Sources: DET CN financial report 202606, computed from monthly lines, Jan to Jun 2026 · 比音勒芬 FY2024 annual report, selling expense note
The in-house proof

This house can hold a price. It has not shown the economics scale.

PYE sits on the same Esquel cost base and does not discount. That answers one question and immediately raises a harder one.

Tmall
shirt at 1,380, flat, 100% of ticket
JD
1,073 to 1,173 through a standing 15% reduction, not a flash event
Knit trousers
1,180 with no discount history at all: the everyday price is the lowest price it has ever been
¥44.1MPYE revenue, 2025 actual
-¥30.8MEBITDA, 2025 actual, on 16 stores falling to 11

So the question is not whether we can hold a price. We demonstrably can. It is what makes that discipline profitable at scale, and PYE has not answered it yet.

Source: PYE own-channel cells verified 28 July 2026 · 2025 actuals from the B&R consolidated financials
The lens

Price is not held one way. It is held seven ways, each paid in a different currency.

This is the reframe the study turned on. The original question was how much marketing a price tier needs. The market says marketing is only one of seven answers, and it is the most expensive one.

How they hold itWho runs itShirt priceTypical holdWhat it costs
Channel scaleYoungor, HLA, JoeOne, LILANZ168-31073%Distribution capital, franchise mechanics
Product authority比音勒芬, ICICLE 之禾769-2,84691%Fabric R&D and a physical retail machine
Function proof龙牙, 白小T, On, Descente169-2,19096%Product engineering, hero-SKU discipline
Design voiceNice Rice, Pane, 蕉内598-659100%Content engine and a community
Factory trust衬衫老罗, 霞湖世家, 欧定214-790100%Cost discipline and a transparency story
Borrowed-luxury designGremade, KKRS ERIC353-36880%Design authority, fabric buying, no media at all
Brand worldRalph Lauren, Hugo Boss1,250-1,79096%Decades of media, flagships, ambassadors

Marketing spend appears in exactly one of these seven rows, and it is the row that needs decades and a balance sheet we do not have. Every other mechanism is bought with something else.

Source: 40-brand roster, mechanism classified per brand · median hold computed from resolved anchor cells
The gap

Between 350 and 620, real money changes hands and almost nobody holds a position.

¥90-170Btransacts in this band every year in China menswear, on four independent readings of market size
¥1-3Bis the visible revenue of brands that actually anchor here, and COS is most of it

Read those as two facts, not one ratio. The band is not empty of transactions. It is empty of positions. The money there is being taken by premium brands discounting down into it and mass brands eventing up into it, and by almost nobody who lives there.

Sources: band sizing from four independent China menswear reports read 28 July · resident prices verified in-app 28 July · held-supply estimate built bottom up from the residents
The entry price

The brands living there bought no media to get in.

This is the single most relevant fact for a business whose margin cannot fund a media machine. The residents of the band we want did not buy their way in.

0physical stores across the two 打版 storefronts holding the band
0brand ambassadors or celebrity partnerships
>46%shop repurchase rate at Gremade, approaching 龙牙's 52% on a fraction of the age

What they bought instead was design authority they rent and fabric they name. Every SKU title says the block and the cloth: 雨露麻, 三醋酸, 康赛妮. The mechanism is visible in the merchandising language.

They rent, at a monthly fee, exactly what Esquel owns outright: the mill, the fabric provenance, and a client list that includes two of the brands sitting in the aspirational band on this map.

Source: 打版 storefront verification, Taobao, 28 July 2026 · deep dive 02, Gremade
The options

Six plays. Two are live, two are closed, two are honest but not the answer.

Each play is screened on what it demands, not on what it promises: the margin it needs against the 66.9% we have, what it costs in its own currency, how long its proof took, and whether we could structurally run it.

PlayMargin it needsWhat it costsTime to proofVerdict
The national volume game44-52%National distribution capitalDecadesClosed. Wrong margin, wrong clock.
The premium discipline play77%Retail machine, fabric R&D13 yearsThe destination of margin repair, not a three-year move.
The function hero play55-68%Product engineering, communityFitsLive. Runnable at our margin today.
The design-voice playNot the issueContent engineNot the issueClosed by our own survey: 0 of 121.
The honest-factory playLow by designCost discipline, transparency3 yearsSpends the margin down, and sells forward the identity we are leaving.
The white-space factory playCompatibleDesign authority, no mediaFitsLive. The pointed one.
Screen run 28 July, revised 30 July · margin band per play from listed-company filings · time to proof from the nine deep dives
Where that leaves us

Two plays we can afford, and one uncomfortable thing about each.

The function hero play

Proven at scale in our own market: 龙牙 reached RMB 1.8 billion with 52% repurchase, on fabric narrative and a community, with no ambassador until its fourteenth year. It sits inside our margin today and is the least media-dependent option on the board.

The uncomfortable part: 龙牙 spent twenty years accruing a community before it sold anything. We would be starting that clock now, and we do not yet have the hero product it is built on.

The white-space factory play

Move up into 350-620 on factory economics, the way the 打版 storefronts did, monetising the mill and the provenance we already own rather than renting them.

The uncomfortable part: that band is an online storefront game, Tmall and JD. Our China engine is Douyin, which carried about two thirds of online sales last year. We would be moving up a price band and sideways into a channel at the same time.

And one correction to our own earlier reading. We had written off the honest-factory play on volume grounds. 衬衫老罗 went from zero to more than a billion renminbi in three years at 179 to 214, through a Douyin programme Esquel qualifies for today. The volume is reachable. The reasons to decline it are that it spends margin downward and sells forward the white-label identity we are trying to leave, which are strategy reasons, not arithmetic ones.

Sources: deep dive 01 龙牙 · deep dive 02 Gremade · deep dive 06 衬衫老罗 · channel mix from DET CN 2025 actuals
What we need from this room

Three questions we cannot answer from the market alone.

  1. Do we stop the downward leak first, or move up first? Our anchor is a clearance price inside a subsidy window running to 2028. Repairing that is a channel decision available now. Moving into the band is a positioning decision that takes longer. They are not the same project and they compete for the same attention.
  2. Do we have a hero product, or do we have to build one? Both live plays rest on one garment carrying the brand. Today our shirt carries the story and our knits carry the volume. The function play needs those to be the same object.
  3. What is this business supposed to deliver, and by when? We deliberately screened these plays on what they demand rather than against a growth number, because the market cannot tell us what to commit to. Tell us the number and we will tell you which of these can carry it.
Next execution point: the SS27 buy, where pricing moves off cost-plus. Core running meeting, 7 to 9 September.
Stated plainly

What this study does not yet know.

Everything above carries a source and a date. These are the places where we do not have one.

Who buys the band
age and city tier for the 350-620 buyer. Neither the Tmall nor the Douyin route returned the data. We have not filled it with judgement, because in a record a guessed demographic looks identical to a measured one.
Private economics
three of the seven mechanisms are run almost entirely by companies that file nothing. We can price them and count their sales. We cannot see their margin.
Our own repurchase
we do not have a comparable China repeat rate to set against 龙牙's 52% or Gremade's 46%. It is the single most useful missing number for judging the function play.
Trousers and jackets
thinner evidence than shirts. Of 21 China trouser readings, 12 are proxy or weak. Autumn stock is not fully on sale until September, so jacket prices need a rerun then.

The interactive map behind this deck carries every one of the 204 price observations with its source, date and confidence rating. Nothing here is asserted without one.

Live evidence base: det-market-study.pages.dev/map/
Appendix

Nine brands, studied properly.

Each one proves or kills a play. The question in every case is never what the brand is, but what its model costs and whether we could structurally run it.