Metrics And Their Definitions

How to use this reference

This glossary covers the metrics shown across the ten tabs in Reporting. A report’s date range and timezone determine which dates its records belong to. The available filters differ by report, and a filter only changes a metric when that metric has the matching data dimension. For filter behavior and saved views, see How To Use Report Filters And Saved Views.

Cards summarize the selected scope; charts and tables show how those values vary by date, queue, team, agent, status, or category. Some period cards use pooled counts or totals instead of averaging daily percentages. The definitions below call out those differences where they matter.

Ticket Operations: Volumes

• Created Tickets: Tickets created during the selected period. The report uses the queue membership currently recorded for each ticket, so team and agent filters do not narrow this measure.
• Solved Tickets: Tickets solved during the period and attributed to staffable human agents. AI-only resolutions are excluded.
• Unsolved Tickets: Tickets still in New, Open, Pending, or Hold in the latest available snapshot for the selected dates.
• Incoming Volume Over Time by Queue: Created tickets grouped by date and current queue. Use this to see when incoming demand rose and which queues received it.
• Incoming Volume by Day of Week & Hour: Created tickets grouped by the weekday and local hour of their creation, displayed as a heatmap.

For steps to compare incoming demand and work waiting, see How To Review Volume And Backlog.

Ticket Operations: Backlog

• Total Backlog: All unsolved tickets in New, Open, Pending, and Hold at the end of the selected period, based on the latest available backlog snapshot. It is a point-in-time count, not the number created that day.
• New, Open, Pending, and Hold: The unsolved tickets in each status at the end of the period, regardless of when they were created. Pending tickets are typically waiting for the requester; Hold tickets are typically waiting on an internal party.
• Ticket Age: Median time since creation for New and Open tickets in the latest backlog snapshot. Pending and Hold are excluded. Historical queue membership is not available, so this measure is not shown when a queue filter is selected.
• Historical Backlog Over Time: Daily end-of-day counts split by New, Open, Pending, and Hold. The total is the sum of those status counts; Median Age is the median age of New and Open tickets for that day.
• Created vs Solved Over Time: Daily created tickets appear above zero and solved tickets below zero. Net is created minus solved for the day; a negative net means more tickets were solved than created that day.

Ticket Operations: Service

• Incoming Tickets: Tickets created each day, using the same queue scope as Volumes. Team and agent filters do not apply because tickets are often unassigned when created.
• Solved Tickets: Tickets solved by staffable human agents. AI-only resolutions are excluded. Queue filtering uses tickets’ current queue and team and agent filters apply to the assigned agent scope.
• AHT (Average Handle Time): Time logged by agents through Zendesk Time Tracking divided by the number of distinct tickets touched. Time is attributed to the day it was logged, not the day the ticket was solved. The period card counts each ticket once across the whole period, so it is not the average of daily AHT values. It includes only time captured after Time Tracking was configured; earlier activity is not backfilled.
• Worked Hours: Verified time in productive agent statuses, calculated from deduplicated adherence intervals. This measure has no queue dimension, so a queue filter does not narrow it.
• Agent Productivity: Solved tickets divided by worked hours. The period card divides period totals; each daily row is that day’s ratio. With a queue filter, solved tickets narrow to those queues while worked hours remain across all queues.
• Backlog: The closing unsolved-ticket count from the last available snapshot in the selected period. It includes New, Open, Pending, and Hold.
• SLA: Fulfilled SLA targets divided by fulfilled plus breached targets. Only tickets with a fulfilled or breached result enter the denominator.
• CSAT: Good ratings divided by all counted good and bad ratings. The period value pools rating counts across days rather than averaging daily percentages. Excluded ratings do not count.

The Service Metrics over Time table shows these measures by day, with a period summary. For AHT setup and capture requirements, see How To Configure Zendesk Time Tracking For AHT.

Workforce Time: Adherence

• Adherence: Adherent scheduled minutes divided by verified scheduled minutes. It measures whether the recorded agent state matched the scheduled activity at the right time. Minutes without verified status are excluded from this score.
• Conformance: Actual productive minutes divided by scheduled productive minutes. It measures whether the agent worked the scheduled total, regardless of when during the day the work occurred.
• Status coverage: Verified status minutes divided by elapsed scheduled minutes in the selected period. Read it with adherence: low coverage means less of the scheduled time could be evaluated.
• Scheduled Hours: Total productive hours in the published schedule for the selected scope.
• Agents: Agents with report rows in the selected scope. The table also shows Days, the number of dates represented for each agent.
• Adherence & Conformance Trends: Daily values for the two percentages above. The By agent table also includes scheduled time and status coverage; interval detail compares scheduled activity with the observed state.

Adherence requires a published schedule and tracked agent states. See How To Review Adherence for guidance on interpreting coverage.

Workforce Time: Absence

• Requests: Time-off requests that overlap the selected dates and match the team and agent scope, regardless of status.
• Approved and Pending: Counts of requests in those statuses. The table can also show declined and cancelled requests.
• Approved Hours (Requested): The sum of requested hours on approved requests, as recorded on the request.
• Approved Hours (Effective): Approved requested hours remaining in schedules after reviewed overlap resolutions. If another approved request takes precedence over an overlapping interval, the affected request can have fewer effective hours, including zero.
• Request type totals: The Vacations, Sick, PTO, or other type chips show the number of matching requests of each type.

See How To Review Absence And Overtime for the source records behind these figures.

Workforce Time: Overtime

• Entries: Overtime requests plus overtime activities recorded directly in the published schedule, with duplicate request-backed schedule activities counted once.
• Approved and Pending: Counts of entries with those statuses. A schedule-direct overtime activity is treated as approved.
• Approved Hours: Sum of hours for approved entries. Request hours use the recorded request value when present; schedule activities use their start and end times.
• Activity, Type, and Status: The detail table identifies the overtime activity, whether it came from an agent request, agent timeline, or schedule, and its decision status. Start, end, and hours describe the entry’s time window.

Performance: KPI Scorecard

The metrics in each scorecard row depend on the KPI model that applies to the agent. Administrators can choose metrics, weights, thresholds, scoring values, and comparison types. See How To Configure KPI Targets for the model settings.

• Agents: Staffable agents with scorecard rows in the selected scope.
• Average Score: Mean Total Score across the displayed agents. Its displayed maximum is the mean of the maximum scores in their applicable models.
• Top Score: Highest Total Score among the displayed agents, shown with that agent’s model maximum.
• Average [metric] and Team CSAT: Team-average raw values for enabled model metrics; CSAT is shown as Team CSAT. CSAT uses pooled good and bad ratings.
• Solved: Number of tickets solved by the agent in the period.
• FRT (First Reply Time): Average time from ticket creation to the first agent reply, in minutes. Lower is better.
• Resolution (Full Resolution Time): Average time from ticket creation until the ticket is solved, in minutes. Lower is better.
• AHT (Average Handling Time): Average logged handling time per ticket touched, in minutes. It depends on Zendesk Time Tracking data. Lower is better.
• CSAT: Good ratings divided by counted good and bad ratings, expressed as a percentage. Higher is better.
• QA: Optional quality metric: the agent’s average effective QA grade for the selected dates, as a percentage. Manager reviews are applied and test runs are excluded. No scored grade appears as a dash, not zero.
• Team/Goal Comparison: Percentage difference between the raw metric value and the comparison value configured in the applicable KPI model, such as a team average or fixed goal. Interpret the direction using the metric’s higher-is-better or lower-is-better setting.
• Metric Scores: Each enabled raw metric mapped to its configured threshold and score bands.
• Total Score: The weighted combination of the metric scores in the applicable model. The score scale is model-specific.
• Rating and Salary Bonus: Optional bands that map Total Score to a rating and bonus percentage. A band includes its lower bound and excludes its upper bound; the last band has no upper limit. Salary Bonus appears only when configured and the viewer has permission.
• Agent leaderboards: Up to three boards list the top five agents for selected metrics, ordered by the effective KPI weight. Only agents whose applicable model includes that metric can appear.
• Productivity metrics: When enabled, Work Days counts days with observed productive work, Scheduled Days counts days with planned productive activity, Work Hours is observed productive time, and Overtime Hours is productive time scheduled as overtime. Productivity (Hourly) is solved tickets per Work Hour; Productivity (Daily) is solved tickets per Work Day.

Performance: My Performance

My Performance shows agent results against that agent’s targets and over the selected date range. See How To Review An Agent’s Performance and How To Configure 'My Performance' Targets.

• Solved Tickets: Tickets solved by the selected agent during the period.
• CSAT: The share of counted ratings marked good. Excluded ratings remain visible in Rated Tickets but do not count in CSAT.
• Worked Hours: Verified time in productive statuses from deduplicated adherence intervals, compared with the published schedule. A missing published schedule or status tracking can leave this measure unavailable.
• Adherence and Conformance: Use the definitions in Workforce Time: Adherence above. The dashboard can show both as summary cards and daily trends.
• Salary Bonus: The bonus percentage from the KPI model applied to the selected agent for the selected dates, when configured and enabled for display.
• Solved Tickets Trends: Solved-ticket counts grouped by day, or by hour for a single-day view. The target line, when present, is the agent’s configured daily solved target.
• Public Comments: Count of public staff replies in each day or hour bucket.
• Handle Time: Total handle time logged in each bucket, not an average. It comes from captured Zendesk Time Tracking entries.
• CSAT Trends: Good and bad rating counts and the CSAT percentage over time. Excluded ratings are omitted from the score.
• Agent Activity Trends and interval detail: Daily adherence and conformance values; for a single agent, selecting a day opens the scheduled and observed interval details.
• Rated Tickets: The ticket-level list of ratings, including excluded tickets and the exclusion status. The rating filter changes the list and download, not the CSAT card or trend.

In All agents (average) mode, solved tickets and worked hours are averaged per agent. CSAT, adherence, and conformance are pooled ratios, not averages of the agents’ percentages.

Performance: Engaging

• Eligible agents: Agents marked eligible for the selected rewards period and filters.
• Participation rate: Eligible agents with at least one award divided by eligible agents. The hint gives the participant count.
• Total points awarded: Sum of award points for the report scope.
• Average points per eligible agent: Total points divided by the eligible-agent count.
• Rewards unlocked: Agents whose period points reached an active reward threshold. For an open period, the card is labeled Projected.
• Claims fulfilled: Count of report rows with a fulfilled claim status. For an open period, the card is labeled Projected; the value hint shows the estimated cost of fulfilled claims.
• Points by category and Points awarded over period: Award-point totals grouped by activity category and award date.
• Claims by status: Count of claims in each status.
• Overall Points and category leaderboards: Top five eligible agents by total points and, when there are awards, by up to two leading categories.
• Engaging detail: Award count, Total points, Leading category, Unlocked threshold, Selected reward, Claim status, and Estimated reward cost for each agent. Cost is shown in the configured reward currency.

Planning: Forecasting

Forecasting reports completed reforecast runs using the accuracy each run stored when it finished. See How To Review (Re)Forecast Accuracy for how to compare runs.

• Average reforecast accuracy: Arithmetic mean of scored completed runs in the selected completion-date period. Every run counts once, regardless of its duration or volume.
• Change over period: Accuracy of the last scored run minus the first scored run, in percentage points. It compares two runs, not two period averages.
• Initial forecast accuracy: Average backtest accuracy of the selected models for the distinct parent forecasts represented. It measures held-out historical data when each forecast was built, so it is context rather than a like-for-like comparison with reforecast accuracy.
• Scored runs: Completed runs with a stored accuracy score, out of completed runs in the selected period. Unscored runs are excluded from the average, change, and chart; N/A is not zero.
• Reforecast accuracy per completed run: The stored run score: 100% minus incoming weighted absolute percentage error over elapsed intervals. A miss larger than actual volume is floored at 0%, not shown as a negative score.
• Change vs previous: Percentage-point difference between a scored run and the previous scored run. The first scored run has no comparison.
• Intervals evaluated: Number of elapsed intervals used to evaluate a run. Actual volume is the volume recorded across that run’s window.
• Forecast, Run type, Queue scope, and Completed: Identify the parent forecast, whether the run was scheduled or manual, which queues it covered, and when it finished in the report timezone. A queue filter selects runs to display; the score remains for every queue covered by each run.