Fable
The bakery’s two rulers
A bakery opened in the old neighborhood. The owner spent half a year researching recipes, tried more than twenty kinds of flour and fermentation methods, and finally baked a sourdough he was satisfied with. In the first two weeks after opening, he watched two numbers: how many people walked in each day, and how long it took on average from a customer entering to leaving with bread.
The first month's data looked decent. On average, it took only seven minutes from entering the shop to leaving, and many people bought something after looking around. But in the second month, the repurchase rate collapsed. Fewer than 10% came back to buy a second time. The shopkeeper stopped a few people on the street who had bought in the first week and asked why. The answers were almost identical: the bread was too sour, and their families could not get used to it.
Only then did the shopkeeper realize he had been optimizing the wrong ruler. The speed from entering to leaving was because bread was placed at the door, packaging was pretty, and checkout had no queue. But that ruler measured only 'the speed of getting bread,' not 'whether the bread they got was what they wanted.' The time he spent optimizing the checkout process should first have been spent figuring out what bread the old neighborhood actually loved.
The shopkeeper closed the shop for two weeks and changed strategy. Every day he baked four flavors of bread, only six loaves of each, and put them at the door for free tasting. After tasting, he asked customers to fill out a small card: which one would you be willing to pay for tomorrow. In two weeks he collected over three hundred cards, and the red bean and milk flavors accounted for more than seventy percent of purchase intent. Only eight people checked the original sourdough.
After reopening, the shop owner focused on red bean paste and creamy flavors as the main products. This time, the time from entering to leaving actually became longer—because more people were queuing. But the repurchase rate rose from less than 10% to nearly 60%, and three months later a second store opened across the street.
Later, a peer came to learn from the shopkeeper, who drew a line behind the counter. On the left side of the line he wrote "before making bread": you must go out on the street and ask clearly whether anyone is hungry and looking for bread, and what kind of bread they are hungry for. On the right side he wrote "after selling bread": you must make every second between a buyer pushing the door open and taking the first bite shorter.
The owner later carved this logic into the wall of the new shop: if the answer on the left is "no one is looking," then no matter how beautifully the number on the right is optimized, it is useless. It is not that the ruler is useless; it measured the wrong object.