Sonic Song sent back NTD 20 million because he had nobody free to run it.
The mandate arrived in the middle of a year already full. His agency had the brief and the relationships; it did not have the hands. Song counted the loss in the plainest unit available to him, which was the money.
He did not conclude that he needed to hire.
He concluded that he needed to be able to hire, which is a different problem, and it took him half a year to solve.
Asked what he actually did next, Song describes something closer to an audit than a search. “My own capabilities and energies would attract the right people with good capabilities,” he says, and he asked the upper end of his market for recommendations on top of that. Then the correction, which is the useful part: “I didn’t realise the right people were already around me.”
If the person is able to face his own failure positively, this person is not too far from success.
Once he stopped recruiting strangers, the team assembled in six months.
The filter he runs on the way in is short enough to repeat. One minute of introduction, then “three achievements in your life,” then “three failures in your life.”
“It takes me five minutes to know if the person is suitable,” Song says, “and if the person could last thirty minutes in a meaningful conversation with me, that’s the basics.” What follows is not a trial period. The new person goes “straight into the deep end,” with safeguards Song keeps in place around them.
Most interviews stop at the achievements, which is why most interviews learn nothing.
If the person is able to face his own failure positively, this person is not too far from success.
He names one pass and one refusal. The pass was a KOL he reads as a community-attractor and a builder, strong on leadership. The refusal was a former boss from his salaried years, whose state of mind and working style did not fit the room he was building; that conversation did not continue past the thirty minutes.
I do not usually name them as staff, they are partners to me.
What happens to the people who clear the test says more than the test does. Song puts them on salary and then declines the word for it: “I do not usually name them as staff, they are partners to me.”
The test has one recorded miss, and his account of it is worth the price of the rest. Nobody has raised a flag in five minutes and then turned bad at work. One partner did underperform, which Song attributes to personal issues rather than to a signal he misread, and he is direct about what he concluded: “I still trust my own judgement.”
A filter that fast will be wrong eventually. Song has decided to carry that cost himself rather than slow the filter down, which is a trade an operator running four companies across four markets can defend and a first-time founder probably cannot.1
The line he opened the conversation with, before any of this came up, was eight characters long: 低调行事，高调成就. Act quietly, achieve loudly. He glosses it as a rule about pace rather than volume, and the gloss is the same one the hiring test runs on.
“You can be fast, not hasty.”
The half-year he spent formalising what was already standing around him is the reason the next NTD 20 million does not have to go back.
Footnotes
-
The four companies run from luxury brand IP and legacy-brand handover to a restaurant IP builder and a pet cremation service, across Taiwan, Hong Kong, Thailand and Japan. A five-minute filter gets stress-tested faster in that spread than in one industry. Song sat with CÈ’s AI interviewer on 16 August 2026, by text; quotes lightly normalised for print, wording untouched, transcripts on file. ↩