Budgets as a discipline
Money is a non-functional requirement. A system that works and cannot be afforded has failed a requirement as surely as one that is slow, and the disciplines that keep it affordable are engineering disciplines rather than accounting ones. Three of them are published here.
1. Profitability-first
“Until we know the traction, the product lines, the services, the cost lines, and what users actually buy, financial projections and future predictions are made-up, because we do not yet have the data. The path is to discover the market and ship quasi-daily; investment only accelerates the path to profitability.”
The claim is narrower and more defensible than the anti-planning position it resembles. It is not that projections are useless. It is that a projection built on hypothesis has the form of evidence without the content, and the form is the dangerous part: a spreadsheet gets treated as a measurement by everyone downstream of the person who made it up.
The corollary follows directly, and it is the operationally useful half: do not build town-planner financial artefacts on genesis-phase hypotheses. Where a thing sits on the evolution axis determines what kind of financial discipline it can actually support — an argument owned by wardley-maps.sgit.ai, which this page points at rather than reproducing.
2. Pre-approve the ladder
The most transferable idea on this page, and the one with real research behind it. At the moment a budget is approved, the overrun positions are approved too — what happens at one and a half times, at twice, at five times — and the kill point is named while it is still cheap to name.
“The person who approved the first million is, on the evidence, the worst available decider on the second.”
That is the escalation-of-commitment literature stated in one line, and it is why the ladder has to be built before the money is spent. Once a sponsor is committed, every subsequent decision is made by the participant with the strongest incentive to continue. Deciding in advance moves the decision to the only moment when the decider is disinterested.
Three components make it work as a practice rather than a good intention:
| Component | What it is | What it prevents |
|---|---|---|
| The named kill point | A stated multiple at which the project stops, agreed at approval. | The stopping conversation happening for the first time when stopping is most expensive and most embarrassing. |
| The calibration record | Value milestones and sponsor probabilities written down at approval, so predictions can be scored later. | Retrospective certainty. Nobody remembers having been unsure. |
| The public conditional pledge | Saying in advance, in public, what would make you stop. | The assumption that stopping costs trust. The preregistered finding is that a public conditional pledge makes stopping raise trust rather than lower it. |
And the genuinely original move, which is not in the literature:
The things not done because this project was funded should themselves be accepted, with an owner.
Opportunity cost is normally an argument made during approval and then discarded. Making it a first-class acceptance with a named owner turns it into something that persists: somebody is on the record as having accepted the things that will not now happen, and that acceptance can be revisited when the ladder is climbed.
3. Budget-on-the-step
Budgets attached to workflow steps rather than to projects. A project budget is a pool that any step can drain, so the containment is a matter of vigilance. A step budget is a structural limit: the step cannot exceed it, and an overrun is visible at the step that caused it rather than at the end of the quarter.
This is the same instinct as the CI pattern — verify at the smallest unit, before the name that hides it is applied — and the same instinct as the deny-by-default allowlists in the security posture on the topic map. Containment by structure beats containment by attention, which is the recurring finding of this whole site.
Counter-evidence
A second limit, which the boundary at the top of the page creates: a discipline published without its figures cannot be fully audited. A reader cannot check whether these disciplines are actually followed here, because the evidence that would show it is exactly the material that is not publishable. That is a real weakness of this page and it is not resolvable — stating it is the most this site can do.