Finance.

The framework

1. 01

   Do the division: the volume you can afford to lose.

   Here is the sum. Take your contribution margin as a percentage — price minus true variable cost, the figure you built on Day 35 — and call the increase you're considering p. The volume you can afford to lose and still be no worse off is p divided by margin plus p. On a forty per cent margin, a ten per cent rise lets one customer in five walk before profit falls. On a twenty per cent margin it lets a third of them walk. Most operators never run it, decide the idea feels dangerous and hold the price; the feeling is nearly always more pessimistic than the arithmetic.

2. 02

   Run it backwards before you approve another discount.

   The same division reversed is why discounting costs so much. To stand still, a cut of d must win d divided by margin minus d in extra volume: ten per cent off a forty per cent margin needs a third more units, and ten per cent off a twenty per cent margin needs twice as many. Say that out loud the next time someone asks for a little room on a deal, and print the required volume gain beside the discount on the approval form. What you're signing isn't a concession. It's a volume promise nobody in the room has made.

3. 03

   One contained corner, held still for ninety days.

   The sum tells you what you can afford to lose. Only a test tells you what you'll actually lose. Move the price where the damage is contained and the signal stays clean — new customers only, one region, one line, one channel — and hold everything else still for sixty to ninety days so you're reading the price and not the season. The fencing discipline from Day 46 applies here as well: a test that lets existing buyers quietly slide back onto the old terms isn't a test, it's a discount with extra steps. Decide the result that would make you reverse it before you begin.

4. 04

   The loss you can afford is not the loss you can survive.

   Two questions decide whether an affordable loss is a survivable one. First, who leaves. The buyers most likely to go at a higher price tend to be the ones with the lowest willingness to pay and the highest cost to serve, so the profit effect usually beats the volume effect — unless a single account is a fifth of your revenue, in which case the average is meaningless and you model that account by name. Second, what stays behind. Volume you shed doesn't take rent, salaries or platform costs with it, and a heavy fixed base feels an empty week far more sharply than the percentages imply.