Every prospective unit — a business we build from scratch, a domain we are considering acquiring, an idea that came in over the transom — has to pass three questions before it earns a slot. The questions are simple. They are also ruthless. The portfolio cap of ten means every slot is expensive, and the three questions are the filter that keeps the expensive slots from being wasted.
Question one: does it solve a real, searchable problem? Not a problem we imagined. A problem people are typing into Google this week. BCPrivateLoans solves the question 'how do I get a private mortgage in BC.' BCMedicalAccess solves 'private MRI BC.' ZeroLawyer solves 'NDA template Ontario.' If the prospect cannot be anchored to a real search query with real volume, it does not pass. Real search volume is the market telling us the problem exists before we build anything.
Question two: can it reach $1,000 per month within 12 months? This is not an ambition target — it is a reality check. A unit that cannot reach $1k/mo in twelve months is telling us something about the market, the unit economics, or our own ability to execute. We would rather know that on day 30 than on day 400. If the path to $1k/mo is not plausible with the operating stack we have, we walk.
Question three: can it operate without full-time employees? Every unit in the portfolio runs on the three-layer stack plus a founder who reviews. If the unit requires a full-time operator to survive, we do not take it on. Not because operators are bad — because a hire is a decade-long obligation and we will only make it when the unit has already proven it can run without one. This is the question that kills the most prospects. Most good businesses still require people. We do not want good businesses. We want good businesses that compound without employees.
Three yes answers and the prospect earns a slot evaluation. Any single no and it does not. This keeps us from the biggest failure mode a small holdco can have, which is filling the portfolio with decent businesses that do not fit the operating model. Decent plus misfit still costs a slot, and we only have ten.