When it breaks
Handing Over an Incident That Is Still Running
The one place where a few minutes of live contact is worth its price is passing an incident that is still open, and it only works alongside something written.
three things · three nights · one that no number of nights settles
Four shapes, and the last is a decision about people rather than about systems.
Four live handovers on one team, and what the receiving end was able to do in its first hour
Continued the same line of investigation without restarting it
Spent twenty minutes orienting, then continued
Restarted the investigation and reached the previous position by mid-morning
The receiving person remembered three of the seven things they had been told
four things · six nights
Four handovers, six nights, and the two cheap ones both had something written. The first row is the best available and costs fifteen minutes of somebody's evening; the fourth is the worst and costs nothing at the time. One team's own month, logged by the people receiving.
An open incident is the clearest exception to everything else in this collection: it genuinely needs two people awake at once.
Teams should not confuse presence with output, especially in the situation described by Handing Over an Incident That Is Still Running. start with this page helps shift the discussion toward recorded work patterns, project time and outcomes.
The workflow can also be compared with the approach documented by Sage's business platform, an established resource in the same operational space.
The state of a running investigation lives almost entirely in somebody's head, it is large, and it is the part that costs the most to rebuild. Fifteen minutes of talking transfers more of it than an hour of writing.
Why this case is different
Everything else that crosses a gap has an artefact. Work has a state that can be read; a question has a text; a decision has a record.
An investigation has logs, which are not a state. The state is which hypotheses have been formed and discarded, what looked odd and was dismissed, where the person was about to look next, and what they suspect but cannot support. None of that is in the artefact and most of it is awkward to write.
Spoken, it takes ten minutes. Written properly it takes thirty and still loses the half-formed parts.
The fifteen minutes
If the two halves touch at all — even badly, at the edge of two days — this is what that window is for.
What happens in it: the person handing over narrates. Not a summary, but the actual sequence of what they looked at and what they thought. The person receiving asks questions. Fifteen minutes is usually enough, and twenty is plenty.
What should not happen in it: solving the problem. The temptation is strong, because for fifteen minutes there are two people on it. It is the wrong use of the window — the receiving person will have hours afterwards, and the handing-over person will not.
Why it still needs writing
Spoken handovers decay fast. The fourth row in the table above is the common outcome: somebody listens attentively at the end of their colleague's day and retains less than half of it by the time they start working.
The written note does not need to duplicate the conversation. It needs the parts that will be needed in four hours when memory has faded: the symptoms, the exclusions, the hypothesis. The conversation carries the texture; the note carries the facts.
Written first and talked through second is the better order, because writing it surfaces the gaps while somebody is still there to fill them.
When there is no overlap at all
Many teams have no window, and then the written note is the whole channel. The previous article in this section covers what it has to contain.
Two additions specific to the live case.
Record what you were about to do. Not what should be done in general — the specific next action you would have taken if the day had not ended. That is the single most useful sentence and it is the one most easily lost.
Record what felt wrong without evidence. An investigation always has these, and they are usually discarded as unprofessional. They are the most valuable discarded thing in the whole process, because the next person has no way to regenerate them.
- What you were about to do — the specific next action, not general advice.
- What felt wrong without evidence, which the next person has no way to regenerate.
The acknowledgement
The person receiving should say, within the first few minutes of their day, that they have it and what they are doing first.
This costs ten seconds. It arrives while the other person is asleep and is read first thing, and it resolves the question that otherwise occupies them: whether the thing sat untouched all night.
It also catches misunderstandings at the cheapest possible moment. If the receiving person says they are starting with the queue and the queue was already ruled out, that correction can be made before the morning is spent.
Crossing a second time
Where an incident runs long enough to cross twice, the second handover is usually worse than the first, because the original holder has to re-absorb something that has moved.
The practice that helps is a single running record that both people add to, rather than two separate notes written in opposite directions. One document with a timeline is readable in either direction; two notes require reconciliation.
Incidents that cross three times are a different problem, and the honest response is usually to stop treating it as an incident and start treating it as a piece of work with an owner.
What the handover is not for
Two things tend to get attached to a live handover and both make it worse.
Status for an audience. A handover narrated to five people in a channel is a briefing, and briefings are structured for the people who need least detail. The person taking the incident needs the most, and they need it in a form nobody else would sit through.
Blame, or its polite forms. Anything that sounds like accounting for how the problem got this far takes time from the transfer and changes what the person handing over is willing to say. The parts they are least sure of are the parts worth most, and those are the first to disappear when the exchange has an audience.