Round robin sounds simple: rotate tickets one at a time across agents. Five different assignment modes hide under that label, each with a different fairness model. Pick wrong and you either burn out your top agent or starve your junior.
Variant 1: pure sequential
Tickets go to agent A, B, C, A, B, C. Simple and deterministic. Works if all your agents have identical skills and identical capacity. Breaks the moment one agent is on PTO (their slot still rotates and tickets pile up unrouted).
Avoid in production unless you have a small (3-5 agent) team with no specialization.
Variant 2: load-balanced
Tickets go to whichever agent has the fewest open tickets. Fair to capacity, agnostic to past assignments. Default in Freshdesk’s “Omniroute” feature.
Trap: an agent who closes tickets fast gets more work. The high performer becomes the dumping ground. Cap with “max tickets per agent” in the routing settings; once at cap, agent is skipped until they close some.
Variant 3: skill-based
Each ticket has skills required (parsed from category, source, language). Each agent has skills assigned. Routing matches. Configure under Admin, Skill-Based Routing.
Powerful for multilingual teams or specialized product lines. Cost: every new product feature needs a skill mapping update or new tickets get unassigned. Audit monthly.
Variant 4: time-zone-aware
For 24x7 teams across regions, do not round robin globally. Configure groups per time zone with their own round robin within. Tickets land in the group whose business hours are active.
Add a workflow: when a ticket comes in, check current time, route to the group that is in business hours. After-hours tickets queue for next-business-day or escalate to on-call.
Variant 5: hybrid (skill + load + zone)
Real teams need all three. Build skill groups within zone groups, with load-balancing inside each. The routing rule chain:
- Filter by zone (active business hours).
- Filter by skill required.
- Within the matched pool, load-balanced round robin.
This is configured in Freshdesk via Omniroute with multiple rule chains. Test in sandbox before flipping production.
Status-aware routing
All variants honor agent status. Online agents get assigned, Away/Offline are skipped. Train agents to set status before logging off; otherwise tickets queue against an absent agent and SLA breaches accumulate.
For a hard cap, use the “Working hours” feature: agents are auto-set offline outside their declared hours regardless of status setting.
Reassignment when an agent is overloaded
Round robin assigns at ticket creation, not on the agent’s queue depth changing later. If agent A gets six tickets in five minutes (round robin’s normal behavior), they are stuck unless you build a rebalance workflow.
The rebalance: a scheduled hourly check that finds agents at-cap with idle peers, reassigns the oldest unworked ticket to the idle agent. Run it during business hours only.
Reporting on fairness
Build a weekly report: tickets assigned per agent, average resolution time per agent, current open ticket count per agent. If one agent’s count is consistently 2x the median, the routing is misconfigured.
What to do this week
Identify which variant your highest-volume group uses. If it is sequential, switch to load-balanced. If you have multilingual customers, audit your skill mappings against the last 30 days of unassigned tickets.
