The lag
The five that surfaced
The five removals whose absence became a problem, in the order they surfaced, with what each one cost and how the organisation found out.
The standing slot for work with no ownerThe fastest of the five problems
week 5 · surfaced after 14 weeks
Reading supplier contracts before renewal
week 2 · surfaced after 22 weeks
The quarterly reconciliation of the project ledger
week 3 · surfaced after 31 weeks
Shadowing new starters through their first week
week 6 · surfaced after 47 weeks
Logging why a job was turned downThe slowest on the record
week 9 · surfaced after 64 weeks
Surfaced 5
Five removals were written down with the week each of them stopped. Every one of them produced an event inside the 78 weeks watched. Of these, five surfaced as a problem. All five removals whose absence surfaced as a problem, in the order they surfaced.
The standing slot for work with no ownerSmall jobs belonging to nobody stopped happening at all
week 5 · surfaced after 14 weeks
Reading supplier contracts before renewalTwo renewals went through on terms nobody had read
week 2 · surfaced after 22 weeks
The quarterly reconciliation of the project ledgerAn audit asked for a year of figures and nobody could assemble them inside a fortnight
week 3 · surfaced after 31 weeks
Shadowing new starters through their first weekThree of the four people who joined that year left inside eight months
week 6 · surfaced after 47 weeks
Logging why a job was turned downThe same reason for refusing was worked out from scratch four times
week 9 · surfaced after 64 weeks
Surfaced 5
Five removals were written down with the week each of them stopped. Every one of them produced an event inside the 78 weeks watched. Of these, five surfaced as a problem. Five of the thirty-eight. Lags of 14, 22, 31, 47 and 64 weeks.
Five removals out of thirty-eight produced a problem rather than a request. They are set out above with their lags. What follows is each one in turn, because the mechanisms are different and the differences are the useful part.
Week fourteen: the standing slot
An hour a week with no fixed agenda, which appeared on a diary as an empty meeting and was removed by somebody outside the team on exactly that basis.
What it had been doing was absorbing work that belonged to nobody. Small jobs with no owner arrived constantly and got picked up in that hour. When it stopped, they did not get reassigned; they stopped happening. Fourteen weeks later a backlog of them had grown large enough that somebody counted it.
The lag here is the shortest of the five and the mechanism is the simplest: the work accumulated visibly. Nothing had to be inferred. Somebody simply had to notice that a pile was growing, and piles announce themselves eventually.
Week twenty-two: the contracts
Reading supplier contracts before they renewed. Removed as duplication — the terms were on file and could be looked up.
They could. Nobody did. Two renewals went through on terms nobody had read, one of which had changed. The organisation found out when the invoice came.
This is a precaution whose failure is invisible until the specific occasion arrives. The occasion was a renewal date, which is why twenty-two weeks: that is simply when the next renewal fell.
Week thirty-one: the reconciliation
Quarterly reconciliation of the project ledger, removed on the grounds that the figures existed elsewhere.
They did exist elsewhere. What the reconciliation had been doing was not producing figures but assembling them, and nobody outside the team knew that the assembling was where the time went. When an auditor asked for a year's figures, the assembling had to be done from scratch for four quarters at once, under time pressure.
The output of a procedure is visible from outside. The work of producing it is not, and that asymmetry is where several of these removals came from.
Week forty-seven: the shadowing
New starters had spent their first week sitting alongside somebody doing the job. It was not a programme and it had no name; it was what happened. It stopped because the people who would have been sat alongside no longer had a spare day.
Three of the four people who joined that year left inside eight months. Nobody attributed this to the shadowing at the time, and it is not provable now. What is on the record is that the pattern was noticed at week forty-seven, that the organisation went looking for a cause, and that the shadowing was the change it found.
This is the weakest of the five as evidence and the most expensive if true. It is included exactly as it stands, with its uncertainty, for both reasons.
Week sixty-four: the log
A log of why jobs were turned down. Removed without discussion — it lived in a Friday-afternoon habit, and Friday afternoon was the day that went.
Fourteen months later somebody noticed that the same unusual request had been assessed and refused four separate times by four people, each of whom had worked out the reasoning from first principles.
The lag is sixty-four weeks because that is how long it took for a rare event to recur often enough for the repetition itself to become visible. There was no earlier moment at which anybody could have found out, short of being asked.
What the five have in common
Not much, and that is the point worth taking away.
They came from three different functions. Two were error-catching, three were knowledge-carrying. Three were removed deliberately and two stopped without anybody proposing it. Their lags span fifty weeks.
What they share is one thing: in every case the absence was visible only to somebody who was not in the room, and only at a moment that had not yet arrived. An auditor. A renewal. A new starter's second month. A recurrence of a rare request.
That is the whole category. It is not a category of important things or of big things. It is the category of things whose observers are occasional.
What each one cost
The record does not carry money figures for any of the five, and inventing them would be worse than leaving the gap. What it carries is a description, and the five are very unequal.
| removal | what it cost |
|---|---|
| the standing slot | a backlog of about thirty small jobs, cleared over a fortnight |
| the contracts | one renewal at changed terms, absorbed for a year |
| the reconciliation | roughly a fortnight of unplanned work under audit pressure |
| the shadowing | probably contributed to three departures |
| the refusal log | the same analysis done four times |
Two of those are inconveniences, one is a real but bounded cost, one is unquantified and probably large, and one is small and annoying.
The lags do not sort with the costs. The fastest of the five produced the second-smallest consequence and the slowest produced the smallest of all. The two expensive ones sit at twenty-two and forty-seven weeks, in the middle of the range.