What stopped
Who decides what goes
On this record the removals proposed by the people who did the work were safer than the ones proposed from above, and the reason is not that they were wiser.
The monthly slide deck nobody presented
week 1 · still not missed
Printed copies of the weekly rota
week 0 · asked back after 1 week
The daily stand-up
week 0 · crept back after 3 weeks
Reading supplier contracts before renewalTwo renewals went through on terms nobody had read
week 2 · surfaced after 22 weeks
Not missed 1Crept back 1Asked back 1Surfaced 1
Four removals were written down with the week each of them stopped. One of them has not been asked for in the 78 weeks since. Of the rest, one was asked back, one crept back and one surfaced as a problem. One removal of each of the four kinds, from the thirty-eight counted here.
Thirty-one of the thirty-eight removals were proposed by somebody. The list records who, in the loose sense of whether the proposal came from the team that did the thing or from somewhere else.
It is a crude distinction and it is the only one available. Twenty-two came from inside the team that performed the procedure. Nine came from outside it — from a manager, from another team, or from the person coordinating the transition.
The result
| proposed by | n | surfaced as a problem |
|---|---|---|
| the team that did it | 22 | 1 |
| somebody else | 9 | 2 |
| nobody — it just stopped | 7 | 2 |
Again: small numbers, and the difference between one in twenty-two and two in nine would not survive any statistical test worth the name. What makes it worth printing is that the direction matches a mechanism that is independently visible in the record.
Why inside proposals are safer
Not because the people doing a job know best whether it matters. That is the flattering version and this record does not support it — two of the fourteen never-missed entries were defended at length by the people who performed them.
The mechanism is narrower. Somebody who proposes removing their own procedure continues to be in the place where the consequences land. They are the first to see the gap, they have standing to say so, and raising it costs them nothing because they proposed it. Being right about a removal and being wrong about it both get reported, and quickly.
Somebody who proposes removing another team's procedure is not in the place where the consequences land. If the removal turns out badly, the people who notice are the people who lost the argument. Reporting it means going back to someone who overruled them and saying that the thing they said would happen has happened — which is a sentence people find unexpectedly difficult to say, and which, when said, is heard as a continuation of the original argument rather than as information.
The question is not who knows most about the procedure. It is who will be in the room when it fails, and whether saying so will be heard as news or as point-scoring.
The two that came from outside
Both are instructive and neither involves anybody behaving badly.
The first was the quarterly reconciliation, stopped on the reasonable grounds that the figures it produced were available elsewhere. They were. What was not available elsewhere was the act of assembling them, and nobody outside the team knew that the assembling was the part that took the time. This is an ordinary information problem: the person proposing could see the output and not the process.
The second was the standing slot for work with no owner, stopped because it appeared on the diary as an hour with no agenda. From outside it looked like a meeting that had lost its purpose. From inside it was the only mechanism by which small jobs belonging to nobody got picked up. When it went, those jobs stopped happening, and it took fourteen weeks for anybody to connect the two.
In both cases the proposal was sensible on the evidence available to the person making it, and in both cases the evidence available to that person was systematically missing the thing that mattered.
What this is not an argument for
It is not an argument for letting each team decide what it removes. Run that way, a transition produces removals in exactly the places where the organisation has slack and none where it has waste, because nobody volunteers to remove the thing they believe in.
Nine removals came from outside and seven of the nine were fine. Two of the fourteen things nobody has missed were removals the performing team argued hard against. Outside proposals are doing necessary work.
What it is an argument for
One sentence added to any removal proposed from outside the team that performs it:
Who will notice if this was a mistake, and will they tell us?
If the answer to the second half is "the people who objected," then the removal needs a check that does not depend on them volunteering it. Not because anybody is being obstructive — on this record nobody was — but because the channel runs uphill through an argument that has already been lost, and information does not travel well up that particular slope.
On this record that check would have cost one conversation, twice a year, with two teams. It would have caught the standing slot by about week six instead of week nineteen, and the reconciliation by the first quarter end instead of by an audit.
The twenty-two, and what they chose
Worth describing, because the pattern of what a team volunteers is not random either.
Of the twenty-two removals proposed from inside the team that performed them, fourteen were things that the proposer personally found tedious, five were things that had visibly stopped being used, and three were proposed by somebody other than the person who did the work but within the same team.
Nobody volunteered a procedure they believed in. That is not a criticism; it is what volunteering means. But it describes the limit of the inside channel exactly: it produces the removals that are cheap for the organisation and welcome to the person making them, and it will not produce a removal that somebody is attached to, however little it is doing.
Of the fourteen entries nobody has missed, eight came from inside and six from outside or from nowhere — which, against a split of twenty-two to sixteen, means outside proposals produced slightly more of the clean savings per proposal than inside ones did.