Write the list

The five columns

1. What stopped. One line, in the words people actually use.
2. The week it stopped. A number, counted from the start of the transition.
3. Whether anybody said anything. Three values: argued, asked, nothing.
4. What became of it. Filled in later, not now.
5. A sentence, where there is one worth writing.

Columns one to three are filled in at the moment the removal is agreed, in the same meeting, taking a minute or two each. Across eleven weeks and thirty-eight removals that is about an hour.

Column four is filled in on three later occasions. Each pass takes under an hour because most rows do not change.

Why column three is not optional

It is the cheapest column to maintain and it is the only one on this record that predicted anything.

Of the nine removals somebody argued about, none later surfaced as a problem. Of the seventeen nobody remarked on, four did. The median wait before anything happened was four and a half weeks for the argued, six and a half for the questioned, and eighteen for the unremarked.

Two conditions on it:

• Fill it in at the time. Reconstructed afterwards it becomes a judgement about whether the removal was controversial, which is a different thing and a much less useful one.
• "Nothing" is a value, not a blank. A blank reads as "not yet recorded". The whole finding rests on nothing-being-said as a positive observation.

Why column two is a number and not a date

Because every comparison on this site is between lags, and lags only compare if they are measured from a common origin. Week seven and week nine are instantly comparable; the fourteenth of a month and the third of another are not.

It also makes the arithmetic that matters trivial. The absolute week of an event is column two plus the lag, and that is the number that tells you whether a six-month review would have seen it.

The three failure modes

Not cost. Three hours against a transition that occupied a quarter of an organisation's attention is nothing. The list fails for structural reasons and each has a cheap answer.

It has no owner once the transition ends. Whoever coordinates a change is reassigned when the change is done, and the list belongs to a project that no longer exists. Answer: name a person, not a role, before the project closes.

Column four is not anybody's job. Nothing goes wrong if it is not filled in. It is the purest example of work whose absence is invisible — which is, with some irony, exactly the category this record identifies as dangerous. Answer: three calendar entries, made before the transition ends.

It is most interesting when nobody cares. At six months the list says very little. At eighteen months it says most of what is on this site. That interval is precisely the one over which organisational attention decays. Answer: nothing clever. Accept that the read-back will feel pointless at six months and do it anyway.

The mild disincentive

Worth naming because it is real and nobody mentions it.

A list of removals, read back, produces a short column of things that turned out to be mistakes. Nobody is punished for it. Nobody is rewarded either, and the document's most visible output is a record of one's own errors. A transition with no list has no such column and reports a cleaner story.

On this record that column has five entries. The organisation absorbed all five, and the list is the reason it knows there were five rather than two.

The smallest version that still works

One hour, not three: columns one to three filled in during the transition, and a single read-back at twelve months.

That loses the difference between the twelve and eighteen-month pictures, which on this record is two entries — one of them the sixty-four-week failure.

What does not work is any version without column three, which is a list with no signal in it, and any version without a read-back, which is a document nobody opens.

What a row looks like

Concrete, because the description above is more formal than the thing it describes.

Three rows. Nothing in them required anybody to think hard, and the whole of this site is in the third and fourth columns of thirty-eight of them.

The note column was left empty for twenty-two of the thirty-eight, and that is fine. It is the only optional field. It exists so that when somebody reads the row eighteen months later the entry means something — "the standing slot" is not self-explanatory to a person who was not there — and a note written at the time is worth considerably more than a reconstruction.