Use cases
Find the assumptions the decision is standing on
A plan can look solid while resting on two or three sentences nobody has checked. This workflow finds those sentences, labels them for what they are, and shows which ones would change the answer if they turned out to be wrong.
Facts, estimates, and assumptions
Every plan mixes kinds of statement. Some are facts with a source, some are estimates built from data, some are predictions about the future, and some are assumptions nobody has tested at all. They look identical in a spreadsheet cell, which is exactly the problem.
Label each one honestly. A figure you supplied yourself is useful and often the best available, but it remains your figure until something else supports it, and the record should say so.
Find the assumptions that move the answer
Not every assumption matters. Vary each across a plausible range, one at a time, and watch how far the result moves; the few that swing it widely are load-bearing, and the rest are decoration. For each load-bearing assumption, work out its break-even value, the level at which the plan stops working.
Then set the break-even value beside the evidence. An assumption with a wide margin of safety can simply be watched, while one sitting close to its break-even value needs testing before you commit. The most common mistake is testing the assumptions that are easy to check and leaving the uncomfortable ones untouched.
How the room tests assumptions
The room draws the assumptions out of your plan and treats each as a claim. Inspector marks its status, the Challenger presses on the ones the plan leans on most, and the Champion must defend or revise that exact claim rather than the plan in general. Each break-even value is worked exactly from your own figures.
Red Teamer runs the scenario in which a load-bearing assumption fails, and the Strategist asks which plan would survive it. The Verdict names the assumptions that would reverse it and the evidence that would settle them, and the Human Decision stays in your hands.
A subscription plan under test
A subscription box business plans on $35 boxes earning $14 of margin each, with 6 percent of customers cancelling every month and $40 spent to win each new customer. On those figures a customer earns about $233 of margin over their time with the business, nearly six times the cost of winning them. Varying each assumption shows that the cancellation rate moves the answer most: at 10 percent, lifetime margin falls to $140. Using the common rule of thumb that lifetime margin should be at least three times the cost of winning a customer, the plan holds while cancellations stay below about 11.7 percent a month. The 6 percent figure is the founder's estimate, so the Verdict calls for measuring how many of the first 200 customers remain after three months before spending on growth.
Questions
- What counts as an assumption?
- Any statement the plan relies on that has not been checked against evidence. Forecasts, conversion rates, costs at a larger scale, and the behavior of rivals are the usual examples.
- How many assumptions should we test?
- Usually two or three: the ones that move the result most and sit nearest their break-even values. Testing everything spreads the effort too thin to settle anything.
- What if an assumption cannot be tested before we decide?
- Then design the decision around it. Stage the commitment, keep a way back, and set a date to check the assumption once real data arrives.
Further reading
Bring Argumont a Decision worth challenging.
Start with a paid 60 day pilot on your own real Decisions.