What a handover note must say so the next shift does not start over

A handover note that records what was finished is useless. What the next person needs is the state of everything open. Finished work needs no reader.

The release that stopped at six

Priya ran the release from Bengaluru and stopped at half past six. One migration had been applied and a second was queued behind it. The queued one had not been run.

Tom in Toronto picked it up nine and a half hours later. He opened the ticket and could not tell whether the second migration had been attempted. The note did not say.

He spent forty minutes reading deployment logs to answer one question. A single line in the note could have settled it. He logged the gap afterwards.

What her first note contained

Her template had three headings: done, doing, blocked. Under done she listed six tickets, which took most of the available space. That section was also the least useful one.

Under doing she wrote only the migration number. Under blocked she wrote nothing, because nothing had formally failed yet. Both fields were technically correct.

The note was accurate and complete by its own rules. It still left Tom unable to choose a first action. Accuracy and usefulness had come apart.

A task list is not a handover

A task list answers what happened, which the reader can also reconstruct from the ticket history. It does not answer what is now true. Those are different questions.

Only the second question decides what a person does in their first hour. The first one mostly satisfies the writer. The writer is not the reader.

Priya’s list was a summary of her afternoon. Tom needed a description of the system as it stood at midnight. Neither document existed yet.

Describing the state of the work

The useful line was the one she added on the second night. Migration two is written, not run, and depends on the index rebuild. It carries three facts at once.

Each of those facts changes what the reader does next, and in what order. A list of tickets carries none of them. It only carries sequence.

State means what exists now, what has been executed, and what is waiting. Effort belongs nowhere in that sentence. Neither does optimism.

Naming the next action and its owner

Every handover should end with one action and one name. The name can belong to the person reading it in the morning. It must still be written.

Her second note said Tom should rebuild the index before running migration two. It also said he owned the decision to continue. That second half is unusual.

Naming an owner also names who may stop. That permission keeps a hesitant reader from improvising alone at seven in the morning. It also limits the damage.

The thing that breaks first

Each handover should carry a prediction. It names the part most likely to fail in the next twelve hours, and the signal that shows it failing. Hers named one.

She wrote that the rebuild would probably time out near four hundred thousand rows. The job log would show a lock wait, she said. That is checkable in one query.

A prediction is falsifiable, which is what makes it better than a warning. A warning says be careful. A prediction says look here at this number.

Where the uncertainty sits

Writers leave doubt out because it looks like weakness in a status update. In a handover, doubt is the most valuable content available. It is also the rarest.

Priya was not sure the staging copy had the same row count as production. She had assumed it did, and said so in one line. The assumption was wrong.

Tom verified it in four minutes and found a gap of nine thousand rows. He adjusted the batch size before starting anything. The line saved him a failed run.

Context the reader cannot reconstruct

Some knowledge lives only in the writer’s head. Which colleague answers questions at midnight, which dashboard lies, which test is flaky. None of it is written down.

Every item costs the reader an hour to rediscover alone. Some of it cannot be rediscovered without waking someone. That is the expensive category.

She added a short section called known traps. It became the first thing Tom read each morning. It was four lines long.

Links that survive the night

A link is useful only if it opens for the reader. That makes permissions part of the handover rather than an afterthought. Most writers never check.

In the first week two of her three links failed for Tom. Both pointed at a private dashboard he had never been added to. He worked around them with logs.

She started pasting the query text beside each link. The reader could then run it without waiting for access. The link became a convenience, not a dependency.

The note Tom wrote back

Tom’s reply took six lines and went at the top of the next note. It said what he had run, what he had skipped, and why. It took him four minutes.

That reply became the handover for the following night. The format had survived a full round trip without a meeting. That is the only proof that matters.

The round trip is the real test of any template. A handover that cannot be answered in the same shape will not be reused. It will be replaced by a call.

Rewriting the template after two weeks

After ten handovers they cut the done list to one line. The space went to state, next action, and known traps. The note got shorter and the mornings got faster.

That is the trade the original template had exactly backwards. It rewarded completeness of history and punished clarity of state. Nobody had noticed for a year.

Nothing was added to the process. Three headings were removed and one sentence of prediction became compulsory. The change took twenty minutes to agree.

What the morning stopped repeating

The forty minutes of log reading disappeared first. The duplicate migration attempt went next, three weeks later, in a different timezone. Nobody celebrated either.

Priya still writes the note at half past six. Most evenings it now takes eleven minutes instead of twenty-five. The quality went up as the length came down.

The real saving is not the writing time. It is that neither of them holds the whole release in memory overnight. That is what makes the second day work.

If a handover is a description of state rather than a report of effort, what would you cut from the note you wrote last?