Zendesk Schedule Adherence Report

Most support teams that miss service levels are not understaffed on paper. They are staffed for the day but not for the interval — agents online at the wrong hour, clocked into the wrong workstream, or absorbed by work that was never on the schedule.

Schedule adherence is the metric that exposes that gap. It compares what an agent was scheduled to do with what they actually did, minute by minute. This guide covers how to build the report in Zendesk Workforce Management, which supporting metrics keep it honest, and how to connect adherence to the queue metrics leadership actually asks about.

What this report should answer

What percentage of scheduled time was spent on the scheduled activity?
Where does out-of-adherence time go — untracked time, unscheduled workstreams, or general tasks?
Which intervals lose the most coverage, and do they overlap with peak demand?
Are agents present but misallocated, or genuinely absent?
Does adherence explain the intervals where first reply time and SLA risk got worse?

Use the glossary for definitions rather than restating them: schedule adherence, shrinkage, agent utilization, agent concurrency, and forecast accuracy. Keep the report beside the support metrics dashboard.

How adherence is calculated in Zendesk WFM

The formula is simple:

Adherence rate = (Time in adherence / Total scheduled time) × 100

The subtlety is in what counts as “in adherence.” An agent is in adherence when they are clocked into a workstream, general task, or unified agent status that they are actually scheduled for at that moment. Working hard on the wrong thing counts as out of adherence, and that is the point — the metric measures coverage of planned demand, not effort.

Two structural facts shape every report you build on it:

WFM activity tracking is linear and sequential. An agent can only be in one activity at any moment. There is no partial credit for multitasking, which is why adherence and agent concurrency must be read together on live channels.
Only published schedules count. Scheduled time is derived from published schedules. A shift that was planned but never published contributes nothing to the denominator, which quietly inflates adherence for the days you forgot to publish.

How to build the report

Zendesk WFM offers two reporting surfaces, and you want different things from each.

1. Start with system reports for the attendance picture

System reports carry the metrics that answer “were they here at all”:

Metric	What it tells you
Scheduled time	Total time the agent was scheduled to work (published schedules only).
Total time	Total time the agent was logged in, including all activities, general tasks, and untracked time.
Attendance (rate)	Hours actually worked against hours scheduled, regardless of adherence.
Late (occurrences / duration)	How often and how long agents clocked in after shift start.
Left early (occurrences / duration)	The mirror image at the end of the shift.
Overtime (occurrences / duration)	Unplanned time beyond the schedule.
Occupancy rate	How much of tracked time was spent on productive work.

Attendance and adherence are different questions. An agent can score 100% attendance and 60% adherence by being present for every scheduled hour while working on the wrong queue.

2. Move to custom reports for the adherence breakdown

Custom report metrics give you the operating detail:

Adherence rate — the headline.
In adherence time and Out of adherence time — the raw numerator and its complement, which matter because a 90% adherence rate on a 4-hour shift and on a 9-hour shift are very different amounts of lost coverage.
Idle time and Productive time — where the out-of-adherence time actually went.
Paid time and Paid general task or status time — the cost view.
Solved points per paid hour and solved points per productive hour — output normalised against time, which stops adherence from becoming a pure compliance metric.
First response time after assignment — the bridge between adherence and customer-visible speed.

3. Build the interval view, not the daily view

This is the step most teams skip, and it is the one that makes the report useful. A daily adherence average of 88% can hide an interval at 55% that happens to be your busiest hour.

Report adherence per 15- or 30-minute interval, then overlay:

forecast or actual contact volume for the same interval
queue wait time or answer time for the same interval
SLA status for tickets created in that interval

Coverage problems are almost always interval problems. See Zendesk Peak Hours Report for the demand side of that overlay.

4. Use the agent activity view for diagnosis, not scoring

The Agent activity page visualises each agent’s day as a sequence of workstreams, general tasks, unified agent statuses, and untracked time, with their schedule alongside. It is the right tool for answering “what actually happened in that hour” and the wrong tool for performance management. Hovering an entry shows the workstream or general task, ticket ID, start and end time, and duration — which is usually enough to distinguish a mis-clocked agent from a genuine escalation that ran long.

How to interpret the patterns

Adherence is low, attendance is high

People are at work but on the wrong activity. The usual causes are unclear workstream mapping, agents forgetting to switch status, or ad-hoc escalation work that never appears on a schedule. This is a process fix, not a discipline problem.

Adherence is high, service levels still miss

The schedule itself is wrong. Everyone did what they were told; the plan under-covered demand. Move to Zendesk Forecast Accuracy Report — perfect adherence to a bad forecast produces exactly this signature.

Out-of-adherence time is mostly untracked time

Agents are doing work the system cannot see: side conversations, internal collaboration, tooling gaps, or handoffs happening outside Zendesk. See Zendesk Side Conversations and Internal Collaboration Metrics. Untracked time is a measurement problem before it is a behaviour problem.

Adherence dips at shift boundaries

Late clock-ins and early departures cluster at the edges. The Late and Left early metrics quantify this precisely. Small shifts at the boundary matter more than the raw minutes suggest, because handover intervals are often when queues are deepest.

One workstream has structurally poor adherence

The mapping between agent status and that workstream is probably wrong, or the workstream’s conditions do not match the work agents actually receive. Fix the mapping before drawing conclusions about the team.

Common mistakes

Targeting 100% adherence. Perfect adherence means nobody handled anything unexpected, which is not a functioning support team. Most teams find 85–90% is the honest ceiling once breaks, coaching, and escalations are accounted for.
Reporting daily averages only. Interval coverage is the thing that actually breaks; daily numbers smooth it away.
Forgetting unpublished schedules. Unpublished shifts contribute no scheduled time, so adherence looks better than reality.
Comparing adherence across teams with different task mixes. A team with heavy project work will always look worse than a pure queue team.
Treating adherence as a performance score. It is a coverage diagnostic. Using it for rankings drives status gaming and destroys the data.
Ignoring the update delay. Adjusting past schedules or adding time off can take up to an hour to appear across the platform, so same-hour reporting will disagree with itself.
Reading adherence without shrinkage. If your staffing model assumed no shrinkage, adherence will absorb blame that belongs to the plan.

What to do when adherence drops

Split the drop into attendance versus allocation. Late, Left early, and Attendance rate answer the first; In/Out of adherence time answers the second.
Identify the intervals, not the days. Overlay demand for those intervals.
Categorise the out-of-adherence time: untracked, unscheduled workstream, or general task.
Check whether unscheduled work is recurring. Recurring unscheduled work belongs on the schedule, not in the exception column.
Verify workstream and status mappings before coaching anyone.
Re-measure at interval level after two weeks, and confirm the change in first response time or SLA compliance for those same intervals.

Where this report fits in your dashboard

Keep it beside:

Together they show whether the plan was right, whether people followed it, and whether the queue felt the difference.

FAQ

What is a realistic adherence target? Most support teams land between 85% and 90% once breaks, coaching, and unplanned escalations are included. Set the target from your own out-of-adherence breakdown rather than a benchmark.

Does adherence apply to asynchronous work, or only live channels? Both, but it means different things. On voice and messaging, an interval out of adherence is an immediate coverage hole. On email, it shifts work later rather than losing it, so pair adherence with backlog rather than answer time.

Why does adherence look different an hour after I edit a schedule? Past schedule adjustments and added time off can take up to an hour to propagate. Avoid pulling reports immediately after editing.

Should agents see their own adherence? Yes, as a self-service diagnostic. Publishing rankings turns adherence into a status-management game and the data stops describing reality.

Is high occupancy with low adherence a problem? It usually means people are busy on unplanned work. That is a scheduling and intake problem — the work is real, it just was not planned for.

Connect coverage gaps to first reply time and SLA risk - start free