Research ·

Holding messages and duplicate contacts in outsourced support queues

Why a short holding message can reduce repeat contacts, and how to measure the effect without over-messaging customers.

Holding messages and duplicate contacts in outsourced support queues research hero

Key stats

3

timestamps compared per ticket

2

duplicate contact types reported

1

new fact required per holding message

Methodology and findings

Question and scope

This review asks whether a short holding message reduces the number of duplicate contacts a customer makes while a ticket is still open. The population is open tickets that cannot be answered in the first response because they depend on a vendor, a specialist, or a customer action. The unit is one ticket with a measurable gap between the first response and the next substantive update. The question matters because duplicate contacts add work to a queue that is already waiting.

The scope excludes automated acknowledgements that carry no case information. It also excludes tickets closed within the first response, because those customers never experienced a wait. The review does not measure whether the holding message improved satisfaction, since that would require a separate survey design. It looks only at the relationship between communication and repeat contact.

Defining a holding message

A holding message is a customer-facing update sent while the work is still in progress. It states what has been done, what is still needed, and when the customer will hear next. It is not a promise that the issue is nearly solved, and it is not a repeat of the original acknowledgement. The useful version contains at least one fact the customer did not have before.

Write the definition down before counting messages, because teams disagree about what qualifies. A message that only says the ticket is being reviewed adds little and may increase frustration. A message that names the vendor, the pending test, or the next checkpoint gives the customer a reason to wait. The difference is information, not length.

Why silence produces repeat contact

A customer who hears nothing cannot tell whether the request was received, whether anyone is working, or whether the answer was lost. The natural response is to ask again, often through a second channel. That second contact creates a new ticket, a duplicate investigation, and a longer queue. The original ticket may still be progressing while the team answers the copy.

Silence also changes the tone of later messages. A customer who has waited without information is more likely to escalate or to distrust the eventual answer. A holding message does not remove the wait, but it gives the customer a shared timeline. That timeline is often the difference between patience and a complaint. It also gives the agent a factual point to return to when the customer asks what happened.

Counting duplicate contacts

Define a duplicate contact before measuring it. A reply on the same thread asking for an update is not the same as a new ticket opened through another channel. Count both, but report them separately. The first is a sign of anxiety, while the second is a sign that the channel design allows the same request to enter twice.

Match duplicates by customer, request type, and a time window. A customer with two genuinely different problems should not be counted as a duplicate. Read a sample to check the matching rule, because names and email addresses vary across systems. The count is only useful if a reviewer can defend each match. Record the rule and the window beside the report so the next analyst applies the same standard.

The timing problem

The effect of a holding message depends on when it arrives. A message sent after the customer has already followed up does not prevent the follow-up, though it may prevent the next one. Set the first holding message before the expected wait becomes noticeable for that request type. The right moment differs for a password issue and a vendor dependency.

Record the time of the first response, the first holding message, and the first duplicate contact. That sequence shows whether the message came early enough. A team that sends good messages too late will conclude that messaging does not work, when the real problem is timing. The data should be read in order, not as monthly totals.

Channel differences

Email customers often wait longer before following up, while chat customers expect an almost immediate signal. A single cadence applied to both channels will be too slow for chat and too frequent for email. Set the interval by channel and by request type, and record which channel the customer used. The same ticket can move between channels, so note the change.

A phone callback is a different kind of holding message because it consumes agent time. Use it when the customer cannot easily read an update or when the issue needs a conversation. Do not replace every written update with a call, because calls are harder to schedule and harder to document. The channel should match the customer's situation, not the team's preference.

What a useful holding message contains

A useful message names the current state, the next action, the owner, and the next checkpoint. It avoids guessing at a cause and avoids promising a resolution time the team cannot control. If the customer must do something, the message says exactly what and why. If nothing is needed, it says that clearly so the customer does not invent a task.

Keep the message short enough to read on a phone and specific enough to be useful. One or two sentences of status plus one sentence about the next checkpoint is usually enough. Avoid restating the whole history, which buries the new fact. The customer should be able to tell what changed since the last message.

Risks of over-messaging

Frequent updates with no new information train customers to ignore them. A message sent every few hours to say the ticket is still open adds noise and may make the queue look busier than it is. Set a cadence the team can sustain with real facts, and skip a scheduled message when there is nothing new to say. Explain the cadence to the customer so a quiet period does not look like neglect.

Over-messaging also consumes the time needed to resolve the case. An agent who spends the afternoon writing status notes is not chasing the vendor or testing the fix. The cadence should be owned by the queue lead and reviewed against the actual wait. If the team cannot meet the cadence, shorten the promise rather than sending empty updates.

Measuring effect honestly

Compare duplicate contact rates for tickets that received a timely holding message with those that did not, within the same request type. Do not compare across unrelated periods without noting changes in volume or vendor behaviour. A simple before-and-after comparison can be misleading when the mix of work has changed. Keep the request type and the channel constant, and report the sample size beside the rate.

Report the median time to the first holding message, the share of tickets that received one, and the duplicate contact rate. Add the number of tickets where the message contained no new information. That last figure shows whether the cadence is producing substance or only activity. A team can improve the message before changing the schedule.

Limits of the review

The review cannot prove that a holding message caused a customer not to follow up. Customers have many reasons for waiting, and some would never chase a ticket. It also cannot capture contacts made through channels the desk does not see, such as a conversation with an account manager. The finding describes the pattern in the recorded queue.

Seasonality and staffing also affect the result. A holiday week may produce fewer duplicate contacts for reasons unrelated to messaging. Note those conditions beside the comparison and avoid a causal claim the design cannot support. The useful conclusion is usually about timing and content, not about a precise percentage. A direction that holds across two months is worth more than a single dramatic week.

A note for the next shift

The next shift should know which tickets received a holding message and when the next checkpoint falls. Put that information in the handoff note rather than leaving it in an agent's memory. A customer who was promised an update at four o'clock should not wait until the next morning because the promise was not carried over. The handoff is part of the message, even though the customer never sees it.

Applying the review

Take one month of open tickets that waited more than a day and extract the response times and duplicate contacts. Read the holding messages and mark whether each contained a new fact. Compare the duplicate rate for timely, substantive messages with the rate for silence or empty updates. The result should point to one change in cadence or wording.

Apply the change for a month and repeat the extraction. Keep the same request types and note any shift in channel mix. Review the oldest cases to see whether the message reached the customer who needed it most. A modest, informative cadence is better than an ambitious schedule that the team cannot keep. Write the cadence into the queue guide so it survives a shift change.

Sources

  1. ISO 18295-1:2017 customer contact centres - Contact-centre service measurement and customer communication concepts.
  2. Atlassian incident management handbook - Stakeholder update cadence during ongoing incidents.
  3. GSA customer experience guidance - Guidance on clear, timely customer communication.
  4. GAO Assessing Data Reliability - Methods for assessing the reliability of operational records.

Related research

Help desk staffing intake

Define the operating lane before staffing begins.

Share the queue, tools, schedule, approvals, and escalation limits. We will use them to prepare a focused help desk staffing brief.

Contact Us