Episode 5
The Test
Trying something is not testing it. Mira builds the smallest honest comparison there is — and holds still everything she isn't measuring.
The board was unanimous: go try it.
Trying something is not the same as testing it, and the whole episode lives in that gap.
The easy version goes like this. On Saturday Mira walks over to the meadow by the dock, has a good day, packs up and stays. What she has afterward is a good day at a new spot. What she does not have is a comparison. Saturday is not Monday, the sun was a different sun, and last week’s memory is not a measurement. That was Episode 3.
Two letters
She draws the lake on a sheet of paper and writes two letters on it. A is the meadow by the dock. B is her stretch of the shore path.
Not “new” and “old.” Not “the better one” and “my usual.” She has heard enough names with the answer already baked in — that was Episode 3 as well.
Four days, one variable
Monday and Wednesday she stands at B. Tuesday and Thursday at A.
Why not a week at one and a week at the other? Because then the weather moves with her. Alternating puts both pitches inside the same weather window, so a hot day hits them both.
Why no weekend? Because Saturday is a different animal. The comparison is supposed to run between the pitches, not between Saturday and Tuesday.
Everything else holds still. The same six hours, eleven to five. The same price. The same cooler. Just her, because the classmate has the week off: two changes at once produce two explanations afterward and no evidence.
No sign, no second flavor, no longer day. All of those are on her list of ideas. All of them wait.
What she is setting up is an An A/B test compares two variants under otherwise equal conditions — two pitches, two days of the week, one result. Without that comparison, any claim that one option 'does better' stays a guess. full card in the smallest honest form the thing has.
And B is not the boring half. B is the half that tells her whether the week is worth reading at all. If both pitches collapse, it was the week and not the place. A comparison always needs the side that doesn’t change. In Mira’s words: “B is the spot where I know what normal looks like.”
What she counts is ice creams per hour, on every test day. Not the profit at the end of the month. One of those she can still steer; the other she can only be informed about — A leading indicator tells you early where things are headed — ice creams per hour in the morning. A lagging indicator only confirms it afterward — the profit at day's end. Measure only the lagging kind, and you learn about the problem after it's too late to fix. full card is what the textbook calls that pair.
The numbers
Monday at B: 51 ice creams. Wednesday at B: 57. Together 108 in twelve hours, so nine an hour.
Tuesday at A: 72. Thursday at A: 84. Together 156 in twelve hours, so thirteen an hour.
Four ice creams an hour apart. In money: €6.50 of margin an hour at A, €4.50 at B.
Now hold Episode 4’s line up against that. It sits at €5. At her old pitch a bought hour did not pay over these four days, however well the classmate sells. Pitch A clears the line with €1.50 an hour left over.
The question from Episode 2 was never whether the new girl was worse. It was whether the stand was standing in the right place.
Incidentally, the test cost her no selling days at all. She sold on all four of them — 264 ice creams, €132 of margin. The only extra thing she did was write down where she was standing.
Ice cream and drowning
Two curves rise together every summer: ice cream sales and drownings. The link is real, it is measurable, and it turns up in every introductory course. Does ice cream cause drowning? Of course not. It is hot out. The heat drives both — more ice cream, more people in the water.
Mira gets a moment of pride here: her own product line is the standard example of the field. Specifically, it is on the page explaining how people fool themselves.
The hidden third factor is called a A confounder is a hidden factor that skews both sides of a comparison — more ice cream sales and more drownings move together not because one causes the other, but because hot days drive up both. Miss the confounder, and you mistake correlation for cause. full card and it sits inside her test week too. What if Tuesday and Thursday were simply warmer than Monday and Wednesday? Then she compared two weather patterns rather than two pitches.
She cannot rule that out. Alternating every other day makes it less likely, because the weather would have had to conspire twice in the same direction. From now on she writes the temperature down as well; it costs one line a day.
And she does the cheapest thing anyone can do with a suspicion: she runs the week again with the days swapped. Spot A on Monday and Wednesday, spot B on Tuesday and Thursday. A comes in at twelve, B at nine. It wasn’t the day of the week.
Four ice creams on top of nine is a bit over forty percent. A gap that size survives four days. Had it been 10.2 against 9.8, four days would have left her holding nothing — and she would have to say so. Small differences need many days; large ones need few. Reverse that rule and you simply measure until the result fits.
Now turn the mirror around. Your most important number went up after you changed something. What else was running at the same time?
Over to the board
Mira moves. She has a pitch, a line, and a procedure for finding out the next thing she doesn’t know yet.
One question stays open: whether two sellers at A can jointly hold above ten an hour. Two people at one table may just be serving each other’s line. That is the next test, and this time she knows what one looks like.
What should she improve next?
Write your ideas down before you read on. In Episode 6 ten of them are on the table — and she has time for six.