Philippines staffing blog ·

Set pause rules for SLA clocks that agents can defend

A clock that stops for every waiting reason measures nothing, so pause rules need named conditions, evidence, and a review the client can follow.

Set pause rules for SLA clocks that agents can defend help desk operations illustration

Direct answer

Operational guidance

Decide what the clock is measuring

A first-response clock measures the time until the customer hears from the desk; a resolution clock measures the time until the request is closed. The two need different pause rules because they fail in different ways. Write the definition at the top of the policy and repeat it in the customer-facing description. An agent should be able to say what the clock counts without opening a manual. Write both definitions in the same place so an agent never has to choose which one applies.

Choose the clock that matches the promise the client actually made. A promise of a reply within four hours cannot be measured by a resolution timer that runs for days. If the client made both promises, report both and do not blend them into one number. Blended measures hide the failure that matters most to the customer. If the client's promise is vague, ask for the sentence the customer will see and measure against that.

Name the conditions that stop the clock

List the states that legitimately stop the clock: waiting on the customer, waiting on a third-party vendor, and waiting on a scheduled change window. Each state needs a definition an agent can apply and a record that proves it. Waiting on the customer is not a pause reason until the desk has asked a specific question and recorded the time. Otherwise the clock stops whenever work becomes inconvenient. Give each state a code so the report can separate a vendor wait from a customer wait without reading notes.

Keep the list short. Every additional pause state creates a way to look fast while the customer waits. If a state cannot be tied to an event with a timestamp, it does not belong on the list. Publish the list with examples so agents and reviewers apply the same test. Review the list once a quarter and remove any state that no longer matches how the desk works.

Require evidence at the moment of the pause

A pause should create a record: who set it, why, and when it will be reviewed. The record is what lets a reviewer confirm the state was real rather than assumed. Make the reason a required field with a fixed set of options instead of free text. Fixed options can be counted and compared, while free text cannot. Make the pause visible on the ticket header so the next agent sees the state before opening the thread.

Ask the agent to write the customer-facing sentence at the same time. A pause that the customer does not understand looks like neglect, even when the desk is waiting correctly. The sentence should say what the desk needs and when the customer will hear next. That single habit prevents most duplicate tickets during a wait. Keep the approved sentence in the knowledge base so every agent uses the same wording for the same wait.

Cap the pause and force a review

Give every pause a maximum duration after which it expires and the clock resumes. A pause without a limit becomes a hidden backlog that no report shows. The cap should match the request type: a vendor answer may take days, while a customer reply should be reviewed within one. When the cap expires, the agent must either resume or record a new, justified pause. Show the expiry time in the queue view so an agent can act before the clock resumes on its own.

Review expired pauses weekly and read a sample. Look for pauses set at the end of a shift, pauses set without a customer message, and pauses that repeat on the same ticket. Each pattern points to a different fix: a handoff problem, a missing question, or a policy that nobody will decide. Fix the pattern rather than the individual record. Assign the review to a named person and record the date, so the check does not depend on spare time.

Report the clock honestly

Report the percentage of time a ticket was running, not only the final elapsed time. A ticket open for five days with four days paused is a different service than one that ran for five days. Show the running time, the paused time, and the number of pause events beside each other. The three numbers together tell a story that any one of them hides. Define the calculation in the report itself so two readers cannot compute it differently.

Do not compare a period with new pause rules to an earlier period without labelling the change. A rule change can improve the reported number without improving the service. Put the effective date on the chart and keep the old definition available for reference. A reader should never have to guess which rule produced a figure. Keep a short note of rule changes beside the report, with the date each took effect.

Agree the rules with the client in writing

Put the pause definitions, the caps, and the evidence requirement in the service agreement or an attached schedule. The client should know which waits are excluded from the promise before a dispute, not after. Ask the client to name the person who can approve a change to the rules. That approval keeps a temporary exception from becoming a permanent one. Include an example of a correct pause and a rejected one so the client can see how the rule is applied.

Test the rules with one real ticket per pause state before the first reporting period. Confirm that the timer stops, the record is created, and the report shows the state correctly. Fix the tool before agents build habits around a broken clock, because a rule that depends on a workaround will not survive a busy week. Keep the test record beside the rule so the next reviewer can see what was checked. Repeat the test after every platform upgrade, since timers and fields often change without notice.

Authoritative sources

Related planning pages