How these numbers are built

We read the filing itself: annual reports, interim reports, quarterly investor releases, SEC filings (10-K, 20-F, financial supplements) and HKEX announcements. Segment tables, geographic notes and operational announcements are transcribed by hand and checked against the printed figure.

Where a study reports a common currency, conversion uses published yearly average exchange rates and the rate table is listed as a source. Growth rates are never computed on converted values — they are taken or computed in the company's own reporting currency, so a currency move never shows up as growth.

Scope is stated for every row: which segment, which geography, which fiscal calendar. A "Greater China segment" and a "China Mainland reportable segment" are different things and we never silently pool them.

A five-year compound growth rate tells you who accumulated. A quarterly year-on-year tells you who is growing now. These two frequently rank brands in opposite orders, and collapsing them into a single "growth" number is the most common way an industry chart misleads. We always show both, side by side, and never average them.

Every study includes a base-year sensitivity check: the same compound rate recomputed from several start years. If the sign flips with the base year, the row is flagged. This check exists because it caught a real error.

Sell-in is what the brand charges the channel. A shoe sold to a distributor or a franchisee enters brand revenue at wholesale price, and whatever the shopper finally pays sits outside the brand's accounts.

Sell-out is what the shopper pays in a store the brand operates itself. The full retail price enters brand revenue. The same pair of shoes can therefore land as roughly twice the revenue at a DTC-led brand as at a distribution-led one.

Sell-through is a third thing entirely: the share of stock the end retailer actually sold. It measures inventory health and has nothing to do with revenue. It is never used here as a revenue figure.

One hard rule: no retail-sales-value-to-revenue multiplier appears anywhere on this site.

No company in this set discloses retail sales value and consolidated revenue for the same brand in the same period. ANTA, Li Ning, Xtep and 361 give retail growth rates only, in wording like "high single-digit growth", which is not a number. The closest case is ANTA's "DESCENTE retail sales exceeded RMB 10bn for the first time" in the FY2025 release, but DESCENTE's consolidated revenue is not disclosed separately, sitting inside the RMB 17.0bn "all other brands" segment with KOLON and others. The pair cannot be assembled, so the multiplier is not derivable and the industry rule of thumb of 1.8x to 2.5x does not go into any table here.