Concepts Draft
Load-bearing assumptions
How to find the two or three beliefs a decision cannot survive being wrong about, and why confidence is the wrong sorter.
The decision it serves
You have a plan with a hundred things in it that you believe rather than know, and you cannot test them all. This idea tells you which two or three to test first, and it prevents the failure of testing everything equally and nothing enough.
Every plan stands on beliefs
Anything your plan treats as true that reality has not yet confirmed is an assumption. "Independent professionals are our customer" is not a fact about the world. It is a belief the business is standing on. Naming it as an assumption does not weaken it, because it makes the belief visible, and visible things can be checked.
A business has hundreds of these. They are not equally important.
Three grades of consequence
The method classifies each assumption by what happens if it turns out to be wrong, not by how confident you feel about it.
- Load-bearing: the decision fails if this is wrong. The model does not work.
- Material: the decision would need significant change.
- Supporting: relevant, but not decisive at the present stage.
Most claims feel load-bearing to their authors. Most are not. The one that is deserves your next test, not your next slide, and that single classification is what the gate reads later, so it is worth an argument now.
Why confidence is the wrong sorter
Confidence moves for the wrong reasons: it rises with repetition, with age, with mood, and with a persuasive colleague. A number written down in month one feels more certain in month six purely because you have seen it so often. The assumption you have carried longest may only be the least examined, which is exactly why it feels like knowledge.
Sorting by consequence instead catches the belief that is both quietly held and structurally critical, the one nobody thought to question. That combination is where businesses tend to break: high confidence on weak ground, with a great deal bet on it.
Finding them
A practical way in is to ask, for any belief, where it came from. Something customers did. Something they said. Or something you hoped. What people did is usually the firmest ground, though even behaviour only proves what it proves.
Then walk the decision across the areas, because a price rests on an assumption about hours, and a channel rests on an assumption about what a customer may cost. The load-bearing assumptions are often in a different column from the decision, made by different people in different years.
Consulting Practice adds a habit for every load-bearing hypothesis. Write one sentence beginning "I would be most uncomfortable to find that..." and then go and find out whether that is so, first.
At Halden Systems
The canvas course follows Halden Systems, weighing a security review subscription for its existing clients. The central claim was that clients would pay, and the basis was six clients expressing interest at quarterly reviews. The evidence was none, because interest is a sentence, not a behaviour, and nobody had yet asked anyone for money. The counter-evidence, once someone looked, was two clients declining a comparable add-on the previous year at half the price. Criticality: load-bearing, without argument. The model fails if this is wrong.
The next test followed from that classification: an offer at full price to the eleven clients whose insurance renewed inside sixty days, made in a conversation that already happened, with a threshold of four signed written before a single call.
Failure patterns
- Everything classed load-bearing, so nothing is prioritised and the gate becomes impassable.
- Criticality assigned by confidence rather than by consequence.
- A load-bearing claim left in the plan with an empty counter-evidence line, read as reassurance rather than as an unsearched room.
- The easiest assumption tested, instead of the cheapest test of the most load-bearing unknown.
Where to see it in the estate
- Wisdom, Thread 01, names assumption and dependency, and Thread 04 shows where confidence outruns evidence.
- The Validation Canvas, module 04, the evidence layer, is where criticality is assigned.
- Consulting Practice Part One, foundation 04, and its Hypothesis and Assumption Register.