DETERMINANT MARKET STUDY · APAC

The same shirt costs a quarter more in Singapore than in Japan.
We are cheaper than G2000 in Hong Kong.

Nine brands, six markets, one basic white shirt. Hand-collected 5 August 2026.

51listings in the grid
43with a price
6recorded absences
48operating models established
01

This study asks a different question from the China one

The China map holds the market constant and varies the brand: who sits where in one country. This holds the brand constant and varies the market: what one item costs in six countries, and whether the answer changes when somebody else operates the brand there.

The six words this runs on

Ticket
the list price on the tag
Take-home
the cheapest price to buy one today, after any automatic discount
Spread
a brand's dearest market divided by its cheapest, in RMB
Control
a brand that operates every market itself, so its spread is what the market does
Treatment
a brand licensed or franchised somewhere, so its spread is market plus operator
Match basis
whether the same product code was found across markets, or only the same specification

What it deliberately does not do

  • It does not price the polo yet. That sheet is empty, and polo is half our own volume.
  • It does not cover markets outside the six. South East Asia, India and the Middle East are where the plan actually expands.
  • It does not read store prices. Every figure is from a brand's own market website.
  • It does not claim a customer view. This is shelf price, not demand.
02

The plan commits to two things that may not both be possible

The price commitment
RMB 358

Take-home by 2028, reached by cutting discount from 27.8% to 12%. Gross margin walks to 77.6%.

The distribution commitment
RMB 55.8M

Distributor sales by 2029, from zero. Inside a build to 100 doors across 16 regions by 2030.

Under a wholesale or licence model, the brand does not set the take-home price. Whoever holds the market does. Both commitments sit in the same document and nobody has measured whether they fit together.

Seven comparator brands have already run the experiment. Four operate every one of these markets themselves. Four hand at least one market to somebody else.

03

Nine brands, six markets, and the same product code wherever it exists

The roster is not arbitrary. Seven of the nine appear in our own staff survey's APAC consideration set, so these are brands the organisation says our customer actually sees.

BrandRole on the evidenceMarkets pricedMatch basis
UniqloControl own company in all six6code in 2 markets, spec elsewhere
COSControl H&M Group entity in all six6same code in all six
TheoryControl Fast Retailing owned6same code in five
Ralph LaurenControl own company in all six6same code in five
ZaraTreatment franchised in Singapore only6same style in five
Brooks BrothersTreatment not direct anywhere4code varies by operator
LacosteTreatment distributor in three of six3code varies
G2000Treatment franchised in Singapore3same code in three
DETERMINANT APACOurs HK, SG, VN3same code, 301

Operating model was filled only from a quotable source: a corporate filing, a group subsidiary list, a company registry, or a statement on the brand's own site. Where none existed it is left unknown rather than inferred. 48 of 51 established.

04

The first version of this analysis was wrong, and it is worth saying how

Every cross-market number here depends on one exchange rate table. The table we started with was a set of desk estimates built off USD/CNY at 7.20. The renminbi has since strengthened to about 6.76.

CurrencyEstimate usedLive rateError
Japanese yen0.04800.04288-10.7%
Korean won0.005330.0047352-11.2%
Taiwan dollar0.22500.20860-7.3%
Hong Kong dollar0.92000.8617-6.3%
Singapore dollar5.40005.2710-2.4%
Australian dollar4.68004.7496+1.5%

On the old rates this deck would have said the market barely moves the price. On live rates it moves it by about a quarter. The finding on the next page exists only because the rates were replaced.

Rates dated 5 August 2026 from two independent providers that agree within 0.2% on the six currencies both carry. Taiwan dollar and Vietnamese dong are single-source. These are mid-market rates, not a dealt rate.

05

Japan and Korea are cheap, Singapore and Australia are dear, by about a quarter end to end

Index of the four brands that operate every market themselves, Hong Kong set to 100. Holding the brand and the product constant, this is what the market alone does to a price.

Singapore112
Australia110
Hong Kong100
Taiwan96
Korea93
Japan88

What it means for a price ladder

A single APAC take-home price is roughly 12% expensive in Japan and 12% cheap in Singapore. That is the size of the argument for or against one number across the region.

What it does not mean

Four brands, one garment, one day. The band is 24 points wide and the two single-source currencies sit inside it, so treat the ranking as firmer than the exact figures.

06

In Hong Kong we are below Zara, and a quarter below G2000

Take-home for one basic white shirt, converted to RMB at live rates. Hong Kong dollar is one of the two rates carrying an independent cross-check, so this ladder needs the fewest assumptions of anything here.

215Uniqlo
343DETERMINANT
404Zara
456G2000
595COS
1,284Brooks Brothers
1,422Theory
1,542Ralph Lauren

G2000 is the brand our own frontline named as our value block. In Hong Kong we sit 25% underneath it. In Singapore we sit level with it: G2000 421, ourselves 422.

Our own figure is the 301 shirt at its official market price. Singapore ladder in the appendix of the live grid.

07

Thirty-eight of forty-three prices are the ticket price, exactly

This is the sharpest break from China, where the event price is the price and the nationals hold 35% to 60% of ticket. In these six markets, 88% of what we observed was full price.

The five exceptions% of ticketWho operates it
Lacoste, Japan60%direct
Lacoste, Korea73%licensed
Uniqlo, Taiwan77%direct
G2000, Taiwan80%66% on a bundledirect
Theory, Korea95%not established

Two consequences

The bundle is not an APAC habit. One bundle appears in forty-three cells, at G2000 in Taiwan. The Benjamin Barker three-for-S$175 architecture we flagged in July is a Singapore local-brand pattern, not a regional norm.

Discount discipline is normal here, not distinctive. Removing our own discount from 27.8% to 12% moves us toward the market standard rather than ahead of it. It buys margin, not position.

08

Brands that hand a market to somebody else show a wider spread, by about a fifth of a turn

1.42xControl spread, 4 brands
the market alone
1.61xTreatment spread, 4 brands
market plus operator
+0.19xThe operator effect
what changes when we do not hold the market

Why I would not bank this yet

The missing prices are not random. 4 of 12 non-direct cells have no obtainable price, against 1 of 29 direct cells. Ten times the rate. Distributor and franchise markets often run no brand website, so the panel systematically under-observes the very cells this measure depends on.

And the binary is the wrong cut

The next page shows one brand where two licensed markets and two franchised markets sit 74% apart. Both are "not direct". Grouping them together hides more than it reveals.

09

It is not whether the brand is licensed. It is who the licensee is

Brooks Brothers operates none of these four markets itself. All four are somebody else. They price 74% apart, and they split cleanly by the type of partner.

MarketTake-home, RMBOperatorType of partner
Japan849Brooks Brothers (Japan) Ltd, 80.5% held by Daidoh Limitedlocal listed corporate
Korea848Artworks Korea Brooks Brothers Co Ltd, a CJ affiliatelocal conglomerate
Hong Kong1,284Walton Brown Group joint ventureluxury retail group
Singapore1,476ImagineX Groupluxury retail group

The two local corporates land within one renminbi of each other and roughly 40% below the two Lane Crawford Joyce companies. A licence does not raise or lower price by itself. The partner's own retail position does.

Lacoste points the same way from the other side: it discounts hardest in the markets it runs itself, Japan at 60% and Korea at 73% of ticket, and holds ticket exactly in Australia where a distributor runs it. Three observations, directional only.

10

One price across six markets is achievable, and COS is the proof

1.21xCOS spread across six markets, the tightest in the panel
  • The identical product code, 1307720001, in all six markets. The only brand in the panel to manage that.
  • Ticket price paid in every one. No discount anywhere.
  • RMB 545 to 659, sitting just above the band we talk about.

What it costs to do that

COS runs through an H&M Group operating company in every one of the six markets. It carries no licensee, no distributor and no franchise partner anywhere in this panel.

Read against our own plan

The tightest price discipline in the set belongs to the brand with the least outsourced distribution. Our plan proposes to tighten price discipline and outsource distribution at the same time.

Uniqlo1.66x
Zara1.54x
COS1.21x
Theory1.37x
Ralph Lauren1.45x
Brooks Brothers1.74x
Lacoste1.99x
G20001.18x
DETERMINANT1.23x

G2000 at 1.18x and our own 1.23x are measured across three markets, not six, so they are not comparable with the six-market figures. A narrower observation window produces a narrower spread.

11

Our APAC organisation sees a completely different market from our China organisation

Staff brand survey, 158 responses across both regions, closed 14 July 2026. The regional split is the cleanest result in it.

0

mainland Chinese brands named by any APAC respondent. Not one, in either wave.

0

APAC mentions of the premium outdoor cluster that dominates the China answer: Descente, Arc'teryx, Kolon, On.

7 of 9

brands in this panel appear in the APAC or bridge consideration set, which is why the roster looks like it does.

Named only by APAC respondents

Benjamin Barker · Suit Supply · Marks & Spencer · Calvin Klein · House of Axis · Club 21 · Nice Rice

Named by both regions, the bridge set

Uniqlo · MUJI · G2000 · Ralph Lauren · Hugo Boss · COS · Zara · Theory · Tommy Hilfiger · Brooks Brothers · Lululemon

Why this matters for the two plans

There is no shared competitor set between the two markets. A single positioning statement has to be argued on something other than a common rival, because the rival lists do not overlap at all on the local side.

It also validates this panel: the roster was chosen from the map, and it turns out to match what our own APAC staff say the customer shops.

Caveat carried from the survey itself: 13 of the 158 respondents were DET APAC. This is a strong signal on direction and a weak one on magnitude, and it measures staff perception rather than customer behaviour.

12

What this does not yet know

Gaps in the data

  • The polo is not priced. Forty-five empty cells, and polo is half our own volume.
  • Five prices could not be obtained, four of them in non-direct markets, which is the bias described on page 08.
  • Theory's operator is unestablished in four of six markets, so it carries the control group on partial evidence.
  • Taiwan dollar and Vietnamese dong are single-source rates.

Questions the data raises but cannot answer

  • Every price here is an online price. A brand can run its own website everywhere while licensing its stores. If so, the operator effect would only show in physical retail, and we have not looked there.
  • The six markets are not where the plan expands. Distributor growth is aimed at South East Asia, India and the Middle East, and franchising is far more common there than in these six.
  • Shelf price is not demand. Nothing here says what sells.

The single cheapest thing that would sharpen all of this: price the polo, and check two or three of these prices in an actual store.

13

The live grid, and what happens next

Every figure in this deck resolves from one dataset. The grid is live, filterable by garment, price basis, tax treatment and currency, and each cell carries its source and its date.

17 Augfirst three-year numbers, framework before detail
28 Augfirst checkpoint, China and APAC separately
7-9 SepSS27 buy, where pricing moves off cost-plus
2 Novbudget final, into the B&R process

Prices hand-collected 5 August 2026 from brands' own market websites. Operating models from corporate filings, group subsidiary lists and company registries; 48 of 51 established, the rest left unknown rather than inferred. Exchange rates 5 August 2026, mid-market, two providers. Confidential, Tessellation Group.