Published August 16, 2026

Zendesk SLA Policies, Set Up Properly

An SLA is a promise with a clock attached. Most SLA problems are really clock problems.

The targets that matter

First reply time. The one customers actually feel. Get this right before you configure anything else.

Next reply time. Stops a ticket going quiet after a promising start, which is the complaint pattern nobody reports until it's a churn conversation.

Periodic update. Useful for long investigations where silence is the real failure.

Resolution time. Set it if you can really control it. If resolution depends on engineering, an aggressive target just produces breaches nobody can act on.

Business hours and pauses

Decide whether the clock runs on calendar hours or business hours, and be consistent. A four-hour first reply target means something different at 11pm on a Saturday depending on that choice.

Use pending status deliberately. When you are waiting on the customer, the clock should pause, otherwise you breach on their delay and your report describes their behaviour rather than yours.

How duplicates corrupt SLA reporting

This is the part rarely discussed and worth understanding, because it makes your numbers lie in both directions.

When one problem arrives as two tickets, one gets answered promptly and the other sits. The customer is content, since they got their reply, so nobody escalates. The forgotten twin ages until it breaches.

Your report now shows a breach nobody complained about, which teaches everyone to distrust the report. Meanwhile the quickly-closed duplicate counts as a compliant ticket and inflates your attainment.

Merging duplicates on arrival removes both distortions. Then exclude tickets tagged `closed_by_merge` from SLA reporting and the numbers start describing reality. There is an SLA attainment calculator if you want to see the effect on your own figures.