Lessons · Lesson 7 of 8
- 01 · The six questions a record has to answer
- 02 · The prefix everybody stopped typing
- 03 · Two owners, and therefore none
- 04 · A permission with more verbs than the reason
- 05 · What a version is for, and what it has to keep
- 06 · Both edits were right; one of them is gone
- 07 · Six people, one record, and who pays for the tidying
- 08 · What a rule costs to keep
Six people, one record, and who pays for the tidying
Find the decisions that never reached the record, work out who pays the minutes and who receives the days, and change the cost instead of the instruction.
Lesson 7 of 8 · 15 min
A hundred and forty-eight decisions
Six people write into a Tanvara style record: merchandising, product development, sampling, planning, the trim store and costing. Over one season, on three styles, Hanadi Naguib read back every email thread and every message group. She counted the decisions. A decision is a moment where somebody chose between options, and the choice bound somebody else.
| Where the decision was recorded | Decisions |
|---|---|
| In the record | 58 |
| Only in email or a message group | 90 |
| Total | 148 |
39.2% of the decisions about the order were in the order's record. The rest were in a thread. That is not a filing problem. It is a decision that exists only for the people who happened to be on that thread.
Twelve of the ninety were needed later by somebody who was not on the thread. Nine were found. Each one took about forty-seven minutes of asking around: 7.05 hours in total, USD 100.11. Three were never found, so they were decided again from scratch. That is fine twice and expensive once. The third was a packing method, re-decided in a way that contradicted an approval Halverton had already given. 1,620 cartons relabelled at 3.00: USD 4,860.00.
The instruction that cannot work
Every factory has already tried the instruction. Put decisions in the system. It gets repeated in meetings and written into procedures. It fails everywhere. The reason is arithmetic, not character.
| Share of… | Merchandising | Everybody else |
|---|---|---|
| Minutes spent copying decisions in | 100% | 0% |
| Lookups that needed them later | 25% | 75% |
Copying one decision from a thread into the record took Hanadi about four minutes. Ninety decisions is six hours a quarter, twenty-four hours a year — USD 340.80, all of it hers. Of the twelve later lookups, three were hers. She pays every minute of the cost and receives a quarter of the benefit.
That is not a moral failure, and no amount of reminding will fix it. It is a plain split between who bears a cost and who gets the value. A procedure built across that split is followed while somebody is watching, and not otherwise. Before you write any teamwork discipline into a procedure, work out who spends the minutes and who receives the days. If they are different people, the procedure will fail. And the failure will be read as a bad attitude.
What worked was making it cheaper, not making it compulsory
Tanvara did not tighten the instruction. They moved the cost.
The style record was split into nine sections: specification, materials, costing, sampling, production, packing, shipping, quality and commercial. Each section has one named owner. Each section has its own comment thread. A decision about packing is now written in the packing section — on the screen the person is already looking at, in the place the next person will look.
| Measured | Before | After |
|---|---|---|
| Decisions recorded in the record | 39.2% | 81% |
| Time to record one decision | 4 minutes | 25 seconds |
The compliance number moved because the cost moved. Nobody became more disciplined. The four minutes became twenty-five seconds, and at twenty-five seconds a merchandiser writes it down without stopping to weigh whether it is worth it. Recording a whole quarter now takes about fifty minutes.
That is the move to take away from this lesson. When a discipline is not being followed, the first question is not how do we enforce it. It is how many seconds does it take. The second question is who gets the benefit. You can nearly always change one of those. You can nearly never change willingness.
Sections, not locks; owners, not gatekeepers
Splitting the record into sections does three things at once. Only the first is about tidiness.
- It gives comments somewhere to live, right next to what they are about. That is the cost reduction above.
- It makes the ownership of lesson 3 work in practice. A record has one accountable owner per state. A section has a standing owner across all states. Those are different jobs and you need both. The record owner answers for the order. The section owner answers for the packing method.
- It reduces collisions without a lock. Six people working in six different sections of one record are not competing for it — as long as the system writes fields rather than whole records. That is lesson 6's fix, and this is the working habit that makes it pay.
A section owner is not a gatekeeper. Anybody may write in a section. The owner is the person who is told when it changes, and who answers for what is in it. A gate slows work down and people route around it. A name that gets notified slows nothing, and it is the reason a wrong value is caught in a day. Lesson 4 found the same thing: a notified name beat an expensive split of permissions, at no cost at all.
The handover, which is where six people become a problem
One more thing about several people on one record. At any moment, most of them are not looking at it. So the expensive event is not simultaneous work — that is lesson 6. It is the moment responsibility moves and the receiving person does not know what they have received.
The cheap discipline is one line, written by the person handing over, in the section: what state this is in, what is outstanding, and what I was about to do next. Tanvara's version is three bullets and takes under a minute. It exists because of the shade standard in lesson 3 — an item that fell into the gap between two people. Neither of them had ever written down what they believed the other was doing.
Check yourselfYour team is small and everybody knows everything. Do you need this?Show the answer
Try the test rather than the argument. Pick a decision made four months ago on a closed order — why that trim supplier, why that packing method — and ask somebody to find the reasoning in ten minutes. A small team stores its decisions in shared memory. That works beautifully, and it has two failure modes: somebody leaves, and time passes. Both are certain. Writing it down costs seconds. Not writing it down costs a re-decision that contradicts something a buyer already approved.