Handoff Research
How caller repeat contact rates expose handoff gaps
A repeat contact review shows where a caller had to start over, and how to separate a genuine follow-up from a failed handoff.

Table of contents
Published November 4, 2026
A repeat contact is a caller who comes back, and the reason can be a normal follow-up or a failed handoff. The distinction matters because the two cases call for different fixes. A caller who returns to confirm an appointment is not the same as a caller who returns because the first agent promised a callback that never came. A repeat contact review separates the cases and points to the handoff that needs repair.
Question and scope
This review asks how often callers make a repeat contact within a defined window and what share of those contacts are linked to a failed handoff. The unit is one caller and one original contact. The observation window is seven days, with a longer window for complex cases. The review includes calls, messages, and written channels that can be matched to the same caller.
The scope excludes contacts that cannot be matched, such as a caller who uses a different number and does not identify themselves. Those contacts are a gap in the record and should be reported as such. The review also excludes contacts that are clearly unrelated, such as a new request months later. The matching rule should be written before the sample is drawn.
Define a repeat contact
A repeat contact is a contact from the same caller within the window that references the same issue or the same account. The reference can be explicit, such as a caller who says they are following up, or implicit, such as a caller who selects the same menu option and reaches the same queue. The definition should not rely on the agent's memory. Write the definition down before the sample is drawn.
Record the channel of the original contact and the channel of the repeat. A caller who called first and emailed second may have moved because the phone wait was long. A caller who called twice may have had no other option. The channel pattern is part of the finding.
Separate follow-up from failure
A genuine follow-up is a contact the team expected, such as a caller who was asked to confirm a time or provide a document. A failed handoff is a contact the team did not expect and the caller should not have needed. The two cases can look identical in the contact record. The review needs the original note to tell them apart.
Read the original note for a promised next step. If the note says the caller would receive a confirmation by a certain time and the caller returns after that time, the handoff may have failed. If the note says the caller would call back with a document, the return is expected. The note is the evidence, and a missing note is itself a finding.
| Repeat type | Original note | What it suggests |
|---|---|---|
| Expected follow-up | Caller asked to provide or confirm | The process worked as designed |
| Missed promise | Team promised a step by a time | The handoff did not complete |
| Unclear close | No next step recorded | The caller had no reason to wait |
| New issue | Different request | Not a handoff failure |
Match the contacts carefully
Use the strongest identifier available, such as a phone number or an account number, and a documented fallback. Do not merge records on a name alone when a unique identifier exists. A false match can turn two callers into one and distort the rate. Record the match method for each case so the review can be reproduced.
Review the matches with a second person for the first twenty cases. The review catches the ambiguous cases and tightens the rule. Freeze the rule once the coding begins and log exceptions. A rule that changes mid-review will produce a result that cannot be compared.
Read the pattern by queue and agent
Count the repeat contacts by queue, by original agent, and by the promised next step. A queue with a high share of missed promises may have a handoff problem rather than a knowledge problem. An agent with a high share may need coaching, and the pattern should be read with the sample size. A small number of calls can produce a misleading share.
Look at the time between the original contact and the repeat. A repeat within an hour may mean the caller did not understand the answer. A repeat the next day may mean the promised step did not happen. A repeat after a week may mean the issue returned. The timing narrows the cause.
Record the shift and the channel for each repeat. A repeat that arrives on a different channel may point to a wait problem rather than a handoff problem. The channel pattern is easy to collect and useful to review. It also shows whether the caller had another route.
Test a handoff fix
Choose one promised step that produces repeat contacts, such as a callback or a document request. Add a confirmation to the close and a record field for the promise. Measure the repeat rate for that step before and after. Keep the other steps unchanged so the result is interpretable.
A confirmation can be a written message, a read-back, or a scheduled reminder. The right form depends on the channel and the caller. The test should use one form and measure the result. A bundle of changes will not show which one helped.
Where the evidence stops
The review cannot show that the repeat contact was caused by the first call. The caller may have returned for a new reason that the record does not capture. The review can show that the same caller returned within the window and that the original note promised a step. That sequence is useful and it is not proof of cause.
The review also cannot show that a lower repeat rate means better service. A team that discourages callers from returning may lower the rate without solving the issue. Pair the repeat rate with a resolution review and a caller feedback sample. The combination is more trustworthy than the rate alone.
A protocol for one queue
Choose a month and define the repeat contact, the window, the matching rule, and the coding categories. Export the contacts and reconcile them with the original notes. Review the first twenty cases with a second reviewer and refine the examples. Freeze the definitions and log exceptions.
Report the repeat count, the share by type, the time to repeat, and the queue pattern. Add a short list of the handoffs the team intends to test. Repeat the review after the change with the same definitions. The comparison is more useful than a single month.
Limitations and next action
The sample is one queue and one month, and the results depend on the call mix and the record quality. A queue with incomplete notes will undercount missed promises, and the review should say so. The protocol should be repeated before it is used as a baseline. A single review is a description, not a benchmark.
The next action is to improve the promise record before expanding the review. A required next-step field and a short confirmation at the close are modest changes that make the repeat pattern visible. The team can then test one handoff and see whether the caller stops coming back. That is a more useful question than whether the repeat rate is high.