Fable
It's not about keeping accounts only when the bridge collapses, but about every beam having footprints
The North Gate workshop built wooden bridges for the city. In the past, acceptance was simple: if the bridge could carry people, it was complete; if it could not, it was redone. The master craftsmen were experienced, and for many years they got by on this method.
Later, as orders grew larger, the bridge was no longer a single plank but consisted of surveying, material preparation, cutting, painting, transport, installation, and review. Each step could change the next. Yet the ledger still recorded only the final result: it worked, or it did not.
The first major rework came suddenly. After the bridge deck was installed, they found the middle section was off by two inches. The carpenter said he cut according to the drawing, the surveyor said the drawing was correct, and the materials transporter said nothing was bumped on the road. Everyone argued around the crooked bridge for a day, and no one could find where the deviation had started.
The manager's remedy was to have everyone sign at the end. There were more signatures, but the problem was still not solved. Because a signature only shows that the person was present, not what he did, which drawing he used, what anomaly he found, or whether the anomaly was later handled.
The second incident was smaller, but more expensive. A bucket of waterproof paint was swapped for cheaper paint, no one noticed during installation, and after the rainy season the bridge surface blistered. The accountant saw only lower cost, the inspector saw only that day's smooth finish, and when the problem appeared, it was no longer clear which purchase, which confirmation, or which judgment had gone wrong.
The new steward switched to a step-by-step ledger. Each bridge no longer recorded only the result, but also every intermediate action: who generated a new diagram, who called the warehouse, who retrieved materials, which step waited for approval, which step failed and retried, which step produced a result for the next step to use.
At first, the workers felt these little records were too fragmentary. But the next time a beam dimension was wrong, they only had to flip three pages to find the cause: a post-measurement drawing change had not been passed to the cutting table. The problem no longer required mutual guessing, and the fix did not require tearing down the whole bridge.
At the year-end review, the workshop found another kind of value in the step-by-step log. It could show which steps waited most often, which tools failed most often, and which human approvals always slowed delivery. The quality of the bridge no longer depended only on final acceptance; it kept improving through the evidence left at every step.