The discipline of killing a project on schedule

Nobody struggles to start a project. Starting is the fun part — a new idea, a gap in the market, a reasonable-sounding thesis for why this one will work. The struggle, consistently, is ending one that isn't working. Not because the signal is unclear — usually it's obvious well before anyone admits it out loud — but because ending a project carries a cost that starting one never does: it means admitting the thesis was wrong, writing off the time already spent, and living with the discomfort of having been the person who championed it.

Why sunk cost is worse inside a portfolio than it looks

In a single-business context, sunk cost reasoning is a well-known trap: the money and time already spent on a failing initiative shouldn't factor into whether to continue it, because that cost is gone regardless of the decision. Inside a portfolio of multiple ventures, the trap is quieter but more expensive, because a struggling project doesn't just cost the capital already sunk into it — it costs the attention that could be going to the ventures that are actually working. Every week spent propping up something that isn't clearing the bar is a week of founder attention, engineering time, and partner goodwill not spent compounding the bets that are.

A board typically forces this discipline on a business from the outside — quarterly reviews, external directors asking uncomfortable questions, a governance structure that doesn't care about anyone's feelings toward a pet project. A privately held, family-named group with no outside investors doesn't have that external pressure by default. That's a genuine advantage in most respects — it means decisions get made for the next decade rather than the next quarter, without an outside party pushing for a faster exit. But it also means the discipline of ending things has to be built deliberately, because nothing external is going to impose it.

What the discipline actually looks like

• A real review checkpoint, set in advance. Every venture gets an honest point at which it either has clearly earned continued investment or it hasn't — decided before the emotional attachment to the idea has had time to build, not improvised under pressure once it's already uncomfortable.
• The bar doesn't move to fit the outcome. The temptation, when a venture is close but not quite clearing the bar, is to quietly redefine the bar. The discipline is keeping the original bar and letting the venture fail it if that's what's actually happening.
• Ending is a decision, not a drift. A project that just quietly stops getting attention without anyone formally deciding to end it is the worst version of this — it ties up nominal ownership and reporting overhead without anyone benefiting from an honest close-out.

Telling a pause from an ending

Not every underperforming venture needs to be killed outright — some genuinely need more time, a different approach, or a smaller scope than originally attempted. The discipline isn't reflexive termination; it's an honest answer to a specific question at the review checkpoint: is the thing that's not working the execution, or the thesis itself? A venture with a sound thesis and weak execution deserves another attempt with the lesson applied. A venture whose thesis has been fairly tested and hasn't held up doesn't deserve a third or fourth attempt just because the idea still sounds appealing in the abstract.

The distinction matters because conflating the two in either direction causes real damage. Treating every stumble as proof the thesis was wrong kills ventures that would have worked with better execution. Treating every failed thesis as an execution problem keeps ventures alive indefinitely on the promise that the next attempt will finally get it right. Neither error is really about the venture in question — both are about avoiding the harder conversation of admitting which one it actually is. Building that honesty into the review checkpoint itself, rather than relying on whoever happens to be in the room that week to raise it, is what turns triage from an occasional act of courage into a routine part of running the portfolio.

What killing a project protects

The instinct is to think of ending a venture as a loss — time spent, thesis wrong, nothing to show for it. The more useful frame is that ending it on schedule is what protects everything else. Capital and attention are the scarcest resources in a group built by one operator rather than a large organisation, and every venture still being propped up past its honest expiry date is capital and attention not going to the handful of bets that are actually compounding. The discipline isn't really about the project being killed — it's about protecting the ones that deserve to keep going.

A group with no board has to be its own board. The discipline of ending things on schedule is what that looks like in practice.

None of this makes the decision easy in the moment — admitting a thesis didn't work never feels good, regardless of how sound the process around the decision was. But the alternative — letting ventures drift indefinitely because nobody wants to be the one to call it — is a slower, quieter, more expensive way to lose the same amount of ground. Killing a project on schedule, with a clear reason and a clean close-out, is one of the least visible disciplines in running a portfolio, and one of the ones that compounds the most.

It also has a quieter benefit that's easy to miss: a group that's known to end things cleanly, rather than let them limp on indefinitely, earns more trust from partners and the people working inside each venture — because everyone knows an honest signal will be acted on, rather than absorbed into a permanent grey area of half-supported projects nobody wants to name.