Research question and scope
This protocol measures the share of backup assignments that have not been tested before a planned leave. The unit is one lane backup assignment for one leave. Tested means the backup has handled at least one real case in the lane with the primary available. The measure describes coverage readiness, not the backup's general ability.
Set the scope to one team, one leave calendar, and a fixed period. Record the lane list, the backup names, and the leave dates. If a lane has more than one backup, count the primary backup and note the second separately. A list with several untested names is not stronger than one tested name.
Keep the review separate from a performance judgment. The purpose is to find a gap before the leave, not to grade the backup. A neutral framing makes it easier for the team to report an untested assignment. State the leave and the lane in the report.
Record whether the leave was planned more than two weeks in advance. A short-notice leave may not allow a test, which changes the interpretation. Note the lead time. The same untested assignment is more concerning when there was time to prepare.
Record whether the lane changed since the last test. A lane that added a tool or a rule may need a new test even if the backup passed before. The change date matters. A recent test on an old process is weak evidence.
Population and unit of analysis
Define the population as every planned leave in the period that has a named backup for at least one lane. The unit is the lane assignment, not the person, because one person may back up several lanes. A backup tested in one lane is not automatically tested in another. Count each lane separately.
Exclude unplanned absences from the main measure and report them separately. An unplanned absence rarely allows a test, but the coverage gap is still relevant. Include planned leaves with no named backup as a separate failure. The two failures need different fixes.
Record the practice date, the case type, and the help the backup received. A practice that was only a shadowing session is different from a case handled independently. Define the minimum test before collecting the data. A clear minimum prevents a generous interpretation.
Include lanes with no named backup as a separate failure. A lane with no name is worse than a lane with an untested name, but both need attention. Report the two counts. The owner can then choose the first fix.
Include the backup's other commitments during the leave. A person who is backing up three lanes at once may not be available. The load is part of coverage. Report the number of lanes per backup.
Definitions that make the measure reproducible
Define a tested backup as a person who handled at least one real case in the lane in the previous ninety days with the primary reachable. Define a shadowed backup as a person who observed but did not handle a case. Define an untested backup as a name with neither. Publish one example of each.
Decide whether a case handled during a previous leave counts as a test. It should count if the person completed the case and the quality check passed. A case handled under pressure without support may be even better evidence. Record the context so the reader can judge.
Define the quality check used for the test. Use the same check as the normal sample. A test that is not checked is only an activity. Record the result and any correction. The quality check turns the test into evidence.
Define a recent test as one within the last ninety days. A test from a year ago may not reflect the current process. Use the same window for every lane. A recent test is stronger evidence.
Check whether the test case was a typical case or an easy one. An easy case may not show the gaps. Use a case with one ordinary complication. Record the case type so the reader can judge.
Sampling and data collection
Collect the leave calendar and the lane list before the leave begins. Match each lane to a named backup and to the practice record. If the practice record is missing, mark the assignment untested rather than assuming it happened. A missing record is a finding.
Interview the primary and the backup separately for a small sample. Ask what the backup handled and what they had to ask. The two answers may differ. Record the difference rather than resolving it. A gap in perception is useful information.
Capture the lane, the leave dates, the backup name, the practice date, the case type, and the quality result. Store the data with the leave record. A count without the dates cannot show whether the test was recent. Recency matters for a fast-changing lane.
Ask the backup whether they felt ready, not only whether they passed the test. A person who passed but felt unsure may need another practice. The self-report is context, not proof. Record it beside the quality result.
Record the time the backup would have to respond to an urgent item during the leave. A backup who is only available in the evening may not cover a morning queue. The hours are part of the plan. Report them with the assignment.
Measuring untested coverage
Report the share of lane assignments with a tested backup, a shadowed backup, or no test. Also report the share of leaves with no named backup at all. The two measures show different risks. A leave can have a name and still have no coverage.
Break the result down by lane and by backup. A lane with a high untested share may need a scheduled practice. A backup with several untested lanes may be over-assigned. Compare only lanes with similar complexity and volume.
Report the age of the last test. A test from a year ago may not reflect the current process. Show the median age and the oldest test. A recent test is stronger evidence than an old one, even if both are recorded as tested.
Report the share of lanes with a tested backup and the share of leaves fully covered. A leave may have one tested lane and one untested lane. The second measure shows the full risk. Show both. Report the count of records behind each share so the reader can judge the base.
Limitations and next action
The measure cannot show how the backup will perform under a real deadline or an unusual case. It also cannot show the primary's willingness to let go. Note any observed handoff friction. A tested backup can still fail if the primary does not share the state.
Give the result to the team lead and the lane owners before the next leave. Schedule one practice for the highest-risk untested lane. Record the date and repeat the review after the leave. A short practice is cheaper than a failed handoff.
Keep the backup list current and remove names that no longer work in the lane. A stale name creates false confidence. Review the list at a fixed interval and after any role change. A small, tested list is better than a long, untested one.
Schedule the practice for the highest-risk untested lane before the next leave. Record the date and the result. A practice that is scheduled is more likely to happen than a general reminder. Review the list after the leave.
Add the backup test to the leave request checklist so it is scheduled at the same time as the time off. A separate reminder is easy to postpone. The checklist should name the lane, the backup, and the test date. A small process change keeps the list from becoming a document nobody uses.
Ask the primary to confirm that the backup has access on the first morning. A test that happened weeks earlier may not cover a changed password or a new queue. A quick access check on the day is a small but useful safeguard. Add the access check to the leave checklist.
