Examine a handover Use this when work waits for information, clarification, or approval between people. Begin with a few recent instances that the people involved can describe accurately. The six-step method and short worksheet come first. A filled fictional example and a reusable observation-to-decision record follow farther down. 1. Record what happened. Choose three recent handovers of the same kind. Record when the work was ready, when it was passed on, when the information was complete, when review actually started, and when it was completed. Include returns for clarification. 2. Ask both people. Ask the sender what they believed was complete. Ask the recipient what they needed to proceed. Record the missing information or decision in their own terms. Check whether the same issue occurred in each instance. 3. Locate responsibility. Identify who supplies the information, who makes the decision, and who can approve a change. Account for privacy, safety, contractual, and operational requirements. 4. Choose the next step. If the records do not yet support a change, agree a bounded observation first: what to record, who will record it, how many instances, when to review, and when to stop. Otherwise, agree one small reversible change, such as adding a required reference to a handover note. Record the expected effect, trial period, who will maintain it, and the conditions for stopping. Participants can question the explanation, correct the records, or decline extra recording or a trial. 5. Carry out the agreed next step. For an observation, record the agreed events without changing the workflow. If a change has been agreed, try it on a small agreed set of handovers. Record waiting, clarification requests, completion, and the extra effort required from each person. Keep sensitive content in its authorized location. 6. Decide from the observations. Review the observation or trial with both people. Compare similar work and account for changes in volume or difficulty. Decide whether more observation is needed or whether to choose a trial. For a trial, keep, revise, or remove the change. Record the reason and who owns the next step. WORKSHEET Work and trigger: Recent instances and dates: Sender and recipient: Ready / sent / received / actionable / review starts / completed times: Missing information or decision: Decision owner and permission: Next observation or agreed change and expected effect: Observation or trial size, review date, and stop conditions: Observed result and additional effort: Decision, owner, and next review: Check whether the intended recipient could act, whether work was actually completed, and whether effort moved to another person. Pause if the change creates errors, exposes information, or exceeds the agreed scope. A short trial may require repetition before it supports a lasting decision. REASONING FILLED EXAMPLE — FICTIONAL TEACHING RECORDS All records and outcomes in this example, including the later case D, are invented. Question and competing explanations Where does elapsed time accumulate in an ordinary internal request queue, and what should we investigate before changing the workflow? Missing information may delay preparation; routing, scheduled reviews, or competing work may delay review starting; review itself may take time. Agree what the events mean Sent = request submitted. Actionable = required information complete, whether or not a reviewer is available. Review starts = recorded start of review. Completed = recorded finish. For actual work, check these definitions with the sender, coordinator, and reviewer. A logged status may not coincide with the actual event; mark uncertainty instead of inventing a time. Invented records — same day Request | Sent | Actionable | Review starts | Completed A | 09:00 | 09:00 | 09:18 | 09:22 B | 09:30 | 09:30 | 09:46 | 09:50 C | 10:00 | 10:08 | 10:24 | 10:28 A and B arrived complete. C needed a missing reference, supplied at 10:08. Calculate three separate intervals, in minutes Sent-to-actionable = actionable minus sent. Actionable-to-review = review starts minus actionable. Review-to-completion = completed minus review starts. Total elapsed = completed minus sent = the sum of the three intervals. Calculated from the invented records Request | Sent-to-actionable | Actionable-to-review | Review-to-completion | Total A | 0 | 18 | 4 | 22 B | 0 | 16 | 4 | 20 C | 8 | 16 | 4 | 28 For C: 10:08 - 10:00 = 8; 10:24 - 10:08 = 16; 10:28 - 10:24 = 4. 8 + 16 + 4 = 28 minutes. These are elapsed intervals, not hands-on effort; interruptions may be included. Observation, explanation, and limits The largest interval in each record is after information is complete and before review starts. How a request reaches an available reviewer is therefore worth investigating. These records do not establish that the reviewer was absent, explain the cause, or show whether the pattern is typical. They cannot establish staffing needs or justify staffing cuts. Supplying C's reference earlier might shorten the first interval while leaving the later wait. This is a prediction to test, not an observed improvement. A and B already had complete information; a missing-reference explanation does not account for their waits. Decision that follows Choose further observation in step 4 before selecting a workflow change. With the coordinator, sender, and reviewer, agree to follow the next five comparable requests of the same type and review requirements. Participants can question the definitions, correct the record, or decline the additional observation. Keep current service, responsibilities, and required checks intact. Alongside the four times, record missing information, routing, workload, competing commitments, exceptions, and the extra recording effort. Use request codes in the shared record and keep sensitive content in its authorized location. Review together when the fifth request is completed or after five working days, whichever comes first. Stop the extra recording if it disrupts service, creates an information risk, or agreement is withdrawn. Fewer records may leave the question open. Five requests bound effort; they do not establish a representative sample. If complete requests repeatedly wait before review, use verified context to choose one agreed, reversible workflow trial. If preparation or review duration instead accounts for the delay, revise the focus. Record the decision owner, expected effect, and review date. Consider quality, transferred burden, consent, and the purpose of the work alongside speed. HYPOTHETICAL LATER CASE D — INVENTED, NOT A TRIAL RESULT Request | Sent | Actionable | Review starts | Completed D | 11:00 | 11:00 | 11:02 | 11:22 Calculate D's three intervals and total. A and D both total 22 minutes. What would comparing only totals hide? What remains supported, and what would you observe next? What would you check before treating D as comparable to A, B, and C? Reasoning check: 0, 2, and 20 minutes; total 22. A, B, and C remain valid observations: their longest interval is before review starts. A and D both total 22 minutes, but A spends 18 minutes waiting to start review and four minutes from review to completion; D spends two and 20 minutes respectively. Comparing only totals would hide that difference. D broadens the inquiry without removing the reason to examine the earlier waiting. Check task difficulty, interruptions, and timestamp meanings. If D is comparable, investigate review duration too. If it is an exception, keep it with its context. Do not discard a record just because it challenges the explanation. One new case does not establish a universal cause. REUSABLE OBSERVATION-TO-DECISION RECORD Question and the decision it could change: Scope, comparable work, participants, decision owner, and agreement: Event definitions, timestamp source, and uncertain or missing values: Observed records (keep sensitive content in its authorized location): Request | Sent | Actionable | Review starts | Completed | Workload / exceptions | | | | | | | | | | | | | | | Calculated intervals, in minutes: Request | Sent-to-actionable | Actionable-to-review | Review-to-completion | Total | | | | | | | | | | | | What was observed, including context and other participants' corrections: Explanation being tested: Other possible explanation: What these records cannot establish: Next decision and why the observations support it: Observation that would change the decision: Next observation or agreed trial; predicted effect, owner, size, and review date: Consent, limits, added burden, stop conditions, and how to restore the prior workflow: What happened; keep, revise, stop, or gather more evidence; reason and next owner: Try the reasoning on another instance Ask a colleague to apply the worksheet to another safe instance and explain their decision. Record what they found unclear and when they would ask for help. They should be able to calculate the intervals and challenge the explanation. What another person needed to use this method: