An agency that grows a brand it does not own is working inside a termination clause, and the better the work goes, the more valuable the clause becomes to the other side.
This is the standing condition of most brand businesses in Asia. Distribution rights, market representation, licensed IP: the operator supplies the growth, the name supplies the leverage, and the contract supplies the exit. Everybody in the arrangement knows how it ends and most operators plan as though they will be the exception.
Sonic Song plans as though he will not be.
Song says he holds sole handling for Taiwan of the Japanese designer Naoki Takizawa, who spent some twenty-four years inside Issey Miyake, took over as creative director of its men’s and women’s lines in 1999, and set up his own house in 2006. Takizawa opened his Fitting Room shop in Daikanyama in 2018, and since 2024 he has run planning and design at MUJI.
That last fact is the structural one. The creator’s attention has moved, the name has not, and somebody in Taiwan is doing the work of keeping it alive in a market the creator does not watch.
I control the IP of these designs, or they pay royalty, or they buy over the whole IP.
Song’s move is to stop reaching for the name.
I control the IP of these designs, or they pay royalty, or they buy over the whole IP.
Read it as three exits priced in advance rather than as one demand. The luxury side is design-led, so Song develops the brand’s DNA into specific pieces, and those pieces are where he takes title.
He does not ask for the house. He asks for the drawings he caused to exist, and he settles which of the three endings applies before the work starts rather than when the relationship sours.
The asymmetry is worth naming. A brand cannot be taken from its owner, but a mandate can be taken from an operator in a single letter, and a design an operator holds outright survives both.
The rule has a boundary and Song states it without prompting. Collaborations break, in his account, when one party cannot control or protect their own personal branding, and he offers politicians as the category where that failure is most common. He will not put names to it.
What he cannot ring-fence is reputation. Controversy attaching to a brand he is carrying reaches him through the work regardless of who holds the drawings, and against that the three exits do nothing at all.
They overthink their own brand and over perfect the thought in their minds.
He also declines the usual origin story about what the rule cost him. He says he learnt everything during his salaried years, so the lesson arrived with no bill attached when he went out on his own.
The tuition was paid. It was paid on somebody else’s payroll, which is the cheapest way anyone has found to buy it, and it is available to any operator still inside a company and paying attention.
The last thing worth taking from him is what he thinks brand owners get wrong, which turns out to be the same error in a different chair.
They overthink their own brand and over perfect the thought in their minds.
His counter is two words long: practicality, and already successful cases as ammunition. The owner perfecting a brand in their head and the operator hoping to be handed one are making the same bet on a future nobody has signed for.1
Song is betting on the drawings, which are already in a drawer with his name on them.
Footnotes
-
The counter-argument Song does not make: a design portfolio ring-fenced piece by piece is also a portfolio nobody can market as a whole. He is trading the upside of owning a house for the certainty of owning a room in it. Song sat with CÈ’s AI interviewer on 16 August 2026, by text; quotes lightly normalised for print, wording untouched, transcripts on file. He declined to name the brands and the collaboration that broke, and the Takizawa mandate is his own account. ↩