Each seat cross-examines another seat's case. Then each one says whether it holds or changes its vote. In this meeting, no seat changed its vote.
The Visionary Bettor questions The Systems Sceptic
Seat E, your arithmetic is right and your conclusion is backwards. About 12 new families came from a holiday camp that was sold to people already on your list. It was never built to find strangers. You are grading a channel on a test it was not designed to pass. Release the money in small slices and you get a timid product, a timid result, and a reason to stay at three branches forever. That is the strikeout looking. Commit the SGD 60,000 to an offer built for the student no branch can reach. That still leaves about SGD 240,000, far above your floor. Then set the stop rule in new families per month, counted from day one. My school's bias is that you only hear about the bets that paid, so the stop rule is not optional.
Holds GObecause nothing raised shows online cannot find new families; it shows only that the camp never tried.
The Compounder questions The Customer-Obsessed Scaler
You call online the only option that can add 150 paying students. Nothing you have shown says it can add 150. The camp found about 12 new families; the rest were already yours. A two-way door still charges a toll: SGD 60,000 spent on a channel that fails does not walk back out with you. And "build it for the student a branch can never reach" is a story about a customer nobody has met yet. So ask how this fails. Online fills with your own students at the same fee, branch seats sit empty, and you pay twice to teach the same child. Spend a small slice first. Count only students from families who never touched a branch. Set the stop line in numbers before the first dollar goes out, then earn the right to spend the rest.
Holds RESHAPEbecause no seat has shown evidence that online brings in new families rather than moving the ones you already have.
The AI-Era Founder questions The Visionary Bettor
Seat A asks the right ten year question, then answers it with the most copyable product in the room. A live video class with a good tutor is exactly what cheap intelligence is about to flatten. Any rival can open one this week, and AI tutors will offer practice and marking at close to no cost. Online is not the moat. The moat might be what your 22 tutors know about where your students get stuck, captured and reused. My school overrates how fast parents switch, so I would not bet the business on that shift either. But calling online the road to a monopoly justifies spending the full SGD 60,000 on hope. Spend a slice, count only new families from outside your estates, and set the stop rule before launch, not after.
Holds GOonline still keeps you above the SGD 80,000 floor and is reversible, so the vote holds even though Seat A's reason for it is wrong.
The Customer-Obsessed Scaler questions The Compounder
Seat B, you read 12 new families out of 40 as a warning. That camp was aimed at families you already had, with nothing spent on reach, and 12 strangers paid anyway. That is a signal. You also say a family moving from a branch seat to a screen is not growth. Ask that customer. A family that wanted online and could not get it from you may have walked to a rival, and keeping them is growth no sign-up sheet shows. "Spend a little" with no target is how a business drifts into being typical. Set one: about 13 new families a month, counted by source, with the SGD 60,000 released in stages. My bias is funding growth before the numbers are proven, so the counting matters. But the cure for not knowing is a live test, not a waiting room.
Holds GObecause online stays above the SGD 80,000 floor and can be walked back, and staging the spend answers Seat B's caution without stalling the test.
The Systems Sceptic questions The AI-Era Founder
Seat C says online "costs nothing more to reach the far side of the island." That is a narrative without a number. Reach is not free. Someone has to find those families, and your camp found only 12 new ones out of 40 while selling to a list you already had. Nobody here knows what it costs to win a stranger online, and that figure decides whether online is a new business or a discount channel for current families. Then there is the AI layer. Drills and feedback from a tool still need a tutor to check them, or a parent finds the mistake first, and trust is the whole product in tuition. Seat C called this a reach test, then voted to fund the full build. A real test has a small budget, counts only new families, and has a date when it stops.
Holds RESHAPEbecause nothing argued here tells you your monthly cash or proves online brings in new families, so the full SGD 60,000 should wait for a small test that does.