Zendesk Round Robin

Everyone asks for Zendesk round robin and most teams don't want it. What they want is for nobody to cherry-pick and for nobody to be buried.

There's no simple Zendesk round robin toggle

Worth saying plainly, because a lot of search results imply otherwise. Zendesk has never shipped a "distribute tickets one each in rotation" switch in the way a phone system distributes calls.

What exists instead is a set of mechanisms that get you to even distribution by different routes, each with its own trade-offs.

Guided mode and Play. Agents take the next ticket from a view rather than browsing it. Pull rather than push, which distributes evenly by construction because the fastest agent takes the most work.
Omnichannel routing. Zendesk assigns work from queues based on agent availability and capacity across channels. Not round robin in the strict sense, and closer to what most teams want.
Marketplace round robin apps. Third-party apps that hold a rotation counter and assign in turn. This is the only route that is genuinely round robin, and it is somebody else code.
Triggers plus a rotation of your own. A webhook to your own service, which decides the assignee and writes it back. Total control, and now you own an assignment engine.

Pull beats push, most of the time

Guided mode is underrated. You define a view, ordered how you want work taken, and agents click to take the next one. They cannot skip, can't browse, cannot leave the awkward ticket for someone else.

The distribution that results is not equal, and that is the point. It's proportional to how much each agent can actually handle right now, which is more useful than equality. Somebody in a long phone call takes fewer tickets. Somebody having a productive morning takes more. Nobody is holding twelve unread assignments while a colleague sits idle.

The cost is agent autonomy, and some teams hate it. It works best on high-volume, relatively uniform queues, and badly on specialist work where the right person matters more than the next person.

Why routing to a group beats routing to a person

This is the argument worth having before you build anything.

A ticket assigned to a group is available to everyone who can help. A ticket assigned to a person is unavailable to everyone else, whatever that person is currently doing.

Consider what breaks when you assign directly.

Absence. The assignee is on holiday, off sick, or in an all-day workshop. The ticket sits there. Your rota was correct on Monday and is wrong on Wednesday, and no round robin counter knows that.
Load is invisible. Round robin counts tickets, not effort. Five password resets and five enterprise escalations are the same number and nowhere near the same afternoon.
Ownership goes early. Once a name is on a ticket, colleagues stop reading it. The queue is no longer a shared responsibility, it is ten private inboxes, and shared responsibility is what makes a support team resilient.
Duplicates get split. Two tickets about one problem, round-robinned to two agents. Both investigate. Both reply. The customer gets different answers and now they don't trust either.

The pattern that survives contact with reality: route to a group with triggers based on the actual content of the ticket, then let capacity-aware routing or guided mode hand out individual work inside that group. Assign directly only where continuity genuinely matters, such as a named account or an escalation somebody already owns.

If you do want true round robin

Sometimes you have a real reason. Sales-adjacent queues where every ticket is roughly the same size and fairness is contractual. Regulated work where "who was given this" needs an auditable rule.

In that case, build it around three things.

First, availability. A rotation that assigns to unavailable agents is worse than no rotation, because the ticket looks handled. Whichever mechanism you use has to check agent status, and ideally your business hours and group setup too, so overnight tickets aren't dealt to people who are asleep.

Second, a cap. Nobody should hold more than a set number of open assignments, whatever the rotation says. When they are at the cap, they're skipped.

Third, a reassignment sweep. An automation that unassigns anything untouched for a few hours and returns it to the group. Otherwise the tickets dealt to the agent who called in sick are invisible until the SLA breaches.

Measure the right thing

Teams introduce round robin to fix a fairness complaint and then measure tickets assigned per agent, which was never the problem.

Measure time to first response by hour and by group, unassigned time, the spread of open tickets held per agent, and reassignment rate. If reassignment is high, your routing is guessing rather than deciding, and the fix is better trigger conditions rather than a better rotation.

And if the complaint was that one agent takes all the easy tickets, the answer is guided mode or a queue they cannot browse, not a counter. A rotation still lets someone solve the easy one and leave the hard one open, which is the behaviour you're actually trying to stop.

FAQ

Frequently asked questions

How do you distribute tickets evenly in Zendesk?

There's no native Zendesk round robin assignment. To auto assign tickets Zendesk gives you group routing plus omnichannel routing, and true Zendesk ticket assignment that distributes tickets evenly comes from a marketplace app or the API.

Does Zendesk have built-in round robin?

Not as a simple toggle. You get guided mode, omnichannel routing based on capacity, marketplace apps that implement a true rotation, or your own service driven by a trigger.

What is guided mode?

A setting that makes agents take the next ticket from a view rather than browsing it. Distribution ends up proportional to capacity, and cherry-picking becomes impossible.

Should tickets be assigned to a group or an agent?

A group, in almost all cases. Assigning to a person makes the ticket invisible to everyone else the moment that person is unavailable.

How do I stop tickets sitting with an absent agent?

An automation that unassigns anything untouched for a few hours and returns it to the group, plus a cap on how many open tickets one agent can hold.

Does round robin improve response times?

Only if unassigned time was your bottleneck. If the problem is cherry-picking or uneven effort per ticket, a rotation counts tickets and misses both.

One problem, two agents, two answers

Even distribution spreads duplicates across your team instead of surfacing them. Ticket Merger merges the pair before either agent starts work.

Start free trial

14-day free trial. No credit card required.