Key takeaways
- Count a misroute only when a message is later moved because the first queue could not handle it.
- Separate a rule defect from a sender ambiguity so the assistant is not blamed for a vague request.
- Record who discovered the misroute and how long the client waited.
- Rewrite two label descriptions with real examples before adding a new queue.
Question and scope
This study asks how often a client message is placed in the wrong queue by the inbox rules a virtual assistant follows. It treats the routed message as the unit and the first queue assignment as the event. The scope is one shared inbox, one client group, and a fixed four-week window. The result describes that inbox and should not be read as a benchmark for every support team.
A message is eligible when it arrives from a client address and is assigned to a named queue within one business day. Internal notes, vendor messages, and spam are excluded. A misroute is counted when the message is later moved to a different queue because the first assignment could not handle it. A move made only to balance workload is recorded separately.
The study also records who corrected the route and how long the message waited before the correction. A misroute found by the assistant is different from one found by the client. The correction point is part of the result, not an anecdote. Report the counts by queue pair and by discoverer.
| Route state | Evidence | Follow-up |
|---|---|---|
| Intended queue | First assignment matches guide | Spot-check monthly |
| Rule defect | Guide clear, route different | Rewrite the label |
| Sender ambiguity | Message fits two queues | Add an example |
| Operational move | Queue closed or owner away | Record separately |
Definitions and evidence standard
Define the intended queue before reviewing the messages. Use the queue guide, the client contact map, and the task type as the reference. If the guide does not name a queue for a message type, mark the route as undefined rather than guessing. An undefined route is a finding about the guide, not a failure by the assistant.
Use the message thread, the queue history, and the correction note as the evidence. A message that was moved without a note can still be counted if the history shows the move. Do not rely on a memory of why it moved. The reason should be visible to a reader who was not in the inbox.
Separate a rule defect from a sender ambiguity. A rule defect exists when the guide gives a clear instruction and the message was routed against it. A sender ambiguity exists when the message could reasonably fit two queues. Code the two conditions separately so the result does not blame the assistant for a vague request.
Sampling and measurement
Collect every eligible client message in the window and freeze the list before coding. Include messages from more than one client and more than one task type. If the window contains a launch or a holiday, report it separately. Record the number of messages, the number of clients, and the number of queues.
Measure the number of messages assigned to the intended queue, the number moved once, and the number moved more than once. Add the time from arrival to the first assignment and the time from the first assignment to the correction. Report the median rather than only an average. A few slow corrections can distort an average.
If a message was moved for a reason unrelated to the rule, such as a queue being closed, record it as an operational move. A move caused by a person being on leave is not a rule defect. Report the operational moves beside the misroutes. The separation keeps the rate meaningful.
Findings and interpretation
Present the queue-to-queue pattern before suggesting a cause. A cluster from billing to scheduling may mean the two labels sound similar to a new assistant. A cluster from one client may mean the contact map is stale. Read the actual threads behind the largest pair before naming a reason.
Compare the misroute rate with the number of new message types introduced in the same week. A week with several new types may produce more errors because the assistant is learning. Report the new-type count with the result. The comparison keeps a learning curve from being read as carelessness.
Check whether the misrouted messages were corrected before the client noticed. A correction found internally is cheaper and may show that the review is working. Report the discoverer with the queue pair. The pair tells the team where to invest next.
Limitations
The study measures routes that were recorded, not every quiet reassignment. An assistant may move a message before anyone sees the first assignment. The result therefore describes visible misroutes, not total routing effort. It also cannot show whether the client was affected.
The queue guide may change during the window, which can make an early message look wrong by the later standard. Freeze the guide version and report any change. The sample may also be biased toward messages that were easy to retrieve. State both limits beside the counts.
Deciding whether a message could reasonably fit two queues requires judgment. Two reviewers may disagree at the margin. Record the disagreements and resolve them with examples before reporting. A calibration set of ten messages improves consistency.
Practical next step
Choose the queue pair with the largest correctable cluster and rewrite the two label descriptions. The descriptions should use the words clients use, not internal shorthand. Add one example of a message that belongs in each queue. Test the rewrite on the next twenty messages.
Add a field to the routing note that records the reason for a move. The field can be a short list: wrong topic, wrong client, missing owner, or queue closed. A small list is faster than free text and easier to count. The note should take less than a minute to complete.
Review the change after two weeks and compare the pair count with the baseline. Keep the rewrite if it reduces moves without increasing the time to first assignment. If the pattern does not move, the cause may be the intake form rather than the labels. The test should be read with the people who route the messages.
Comparing message types
Compare a new request with a follow-up on an existing thread. A follow-up may route correctly because the thread already names the owner, while a new request depends on the label guide. Report the type with the result. A difference may be a normal feature of the work rather than a defect.
Look at client-facing messages separately from internal messages. A client message may carry a higher cost when it is misrouted because the client can see the delay. Internal messages may be moved quietly. Pooling the two can hide where the client risk sits.
Compare messages that arrived through a form with messages that arrived through a personal email. A form may collect a topic field while an email may not. Report the channel with the route. A difference between channels points to a simple repair at intake.
What the evidence does not show
The study does not show that a wrong queue always causes a bad outcome. A capable queue may forward the message quickly and the client may never notice. It also does not measure the quality of the reply. The result describes the first assignment, not the full service.
It does not prove that the assistant ignored the guide. The guide may be hard to search, or the message may have arrived with a misleading subject line. The result describes the route, not the intention. The review should ask what made the wrong queue look plausible.
It cannot be compared across teams without the same queue definitions and message mix. A different client, tool, or review culture will produce a different pattern. The local result should guide a local experiment. A broader claim would need a broader study.
Using the result in a manager conversation
Open with the queue map and ask the manager which pair they expected to be the problem. The answer often reveals an assumption about the labels that the record does not support. The conversation can then focus on the pair with the largest gap. That focus is more useful than a general discussion about accuracy.
Ask the assistant which message was hardest to route and why. The answer may point to a missing example or a client name that appears in two places. Use that message to revise one label. A concrete case is easier to act on than a percentage.
Close with one change, one owner, and one review date. Record the baseline so the next review can compare the same pairs. If the change does not move the pattern, stop and investigate rather than adding another label. The goal is a visible shift, not a longer guide.
Sources and methodology
The study follows eligible client messages through a shared inbox for four weeks, coding the first queue, every later move, the discoverer, and the time to correction against a frozen queue guide.
- U.S. Small Business Administration - Hire and manage employees current.
- U.S. Office of Personnel Management - Training and development current.
- U.S. Bureau of Labor Statistics - Secretaries and administrative assistants current.
- U.S. Cybersecurity and Infrastructure Security Agency - Secure Our World current.
Source count: 4. Last verification date: November 10, 2026.
Questions about the findings
Is every moved message a misroute?
No. A move made to balance workload or cover an absence is operational. Count it separately from a routing defect.
Should the assistant be blamed for a vague client message?
No. Code it as a sender ambiguity and use it to improve the example set or the intake fields.
How many messages are enough for a local study?
Enough to see a stable pattern, often a few hundred. Report the count so a reader can judge whether the rate is meaningful.
Put the finding to work
Review the full research library, compare cluster coverage inside Onboarding And Retention Research, and pair these findings with our VA candidate screening support.
