Helpdesk performance analytics track how quickly and effectively your team handles customer conversations, how much work is coming in, how satisfied your customers are, and how customers use your Help Center. When reviewing your team's performance, you can use this guide to understand each metric and how it's calculated.
Requirements
- Available on all Helpdesk plans
- Auto-QA metrics require the AI Agent add-on
- The Voice and SMS tabs require a Voice or SMS subscription
- All roles can review Helpdesk performance reports
Helpdesk performance analytics
Your helpdesk performance metrics are divided into three reports in the Helpdesk performance section of your Gorgias analytics: Overview, Channels, and Help Center. To open a report, follow the steps below:
- From your helpdesk, click the dropdown menu in the top-left corner.
- Select Analytics.
- In the menu, locate Helpdesk performance, then select Overview, Channels, or Help Center.
- Along the top of the Overview and Channels reports, click the tabs to switch between them.
The Overview report contains the Overview, CSAT, Ticket attributes, and Busiest times tabs. The Channels report contains a tab for each channel: Email, Chat, Social, Voice, and SMS. The Help Center report is a single page that shows how customers use your Help Center articles and search.
Available metrics
To find a metric's definition, select it from the lists below, or scroll through this article to browse every definition. Metrics are grouped by the report they appear in. The Overview report has four tabs (Overview, CSAT, Ticket attributes, and Busiest times), the Channels report has a tab for each channel, and the Help Center report is a single page.
Overview report metrics
Channels report metrics
Email tab
Chat tab
Social tab
Voice tab
SMS tab
Help Center report metrics
• Searches
Overview tab
The Overview tab surfaces high-level metrics across your whole helpdesk, letting you see how your team and AI Agent are handling your support workload, and how well you're meeting your SLAs.
| Achievement rate | |
| Metric type | Percentage |
| Definition | Percentage of tickets where your team met the SLA during the selected period. |
| How it's calculated | Calculated by dividing the number of tickets that achieved their SLA by the number of tickets that either achieved or breached their SLA, then multiplying by 100. Tickets are counted in the period where their SLA deadline fell, even if they were created earlier. |
| Notes Tickets with a pending SLA (one that hasn't been achieved or breached yet) are excluded. Each ticket is counted once, even if more than one SLA policy applies to it. Spam, trashed, and deleted tickets are excluded from the calculation. A period with no qualifying tickets shows an achievement rate of 0%. Voice call SLAs are reported separately in the Voice tab of the Channels report. | |
| Average CSAT | |
| Metric type | Average |
| Definition | Average satisfaction (CSAT) score from answered surveys sent during the selected period. |
| How it's calculated | Calculated by dividing the sum of all survey scores (1-5) by the number of answered surveys. Only surveys sent during the selected period are included. |
| Notes Surveys are counted in the period they were sent, not when the ticket was closed or when the customer responded. Surveys that were sent but never answered are excluded. A period with no answered surveys shows no value. | |
| Closed tickets | |
| Metric type | Volume |
| Definition | Tickets closed during the selected period that are still closed. |
| How it's calculated | Calculated by counting the number of unique tickets where the most recent close falls within the selected period and the ticket hasn't been reopened since. Tickets that were closed, but reopened within the selected period are excluded. Tickets that reopened and were closed again count once, based on their latest close. |
| Notes Spam, trashed, and deleted tickets are excluded from the calculation. Snoozed tickets count as closed until they reopen. | |
| Created tickets | |
| Metric type | Volume |
| Definition | Inbound and outbound tickets created during the selected period. |
| How it's calculated | Calculated by counting the number of unique tickets with a creation date within the selected period. |
| Notes Spam, trashed, and deleted tickets are excluded from the calculation. When you filter by agent, only outbound tickets the agent started are counted. Some channels, like social media, can have a creation date earlier than the customer's first interaction, for example when a customer comments on an older post. | |
| First response time | |
| Metric type | Median |
| Definition | Median time customers wait for the first reply from a human agent. |
| How it's calculated | Calculated by measuring the time in seconds from the customer's first message to the first human agent response, then taking the median across all qualifying tickets. A ticket is included in the selected period if the agent's first response fell within the period, even if the customer's first message did not. |
| Notes Only inbound tickets with at least one customer message and one agent message are included. Messages from AI Agent, Rules, and other automations don't count as a first response. Phone tickets are excluded. For merged tickets, the response time is taken from the ticket they were merged into. Time outside business hours is included. We use the median rather than the average so a few unusually slow tickets don't skew the number. | |
| Human response time after AI handoff | |
| Metric type | Median |
| Definition | Median time customers wait for a human agent to reply after AI Agent hands over the conversation. |
| How it's calculated | Calculated by measuring the time in seconds from AI Agent's last message to the first human agent response, then taking the median across all qualifying tickets. A ticket is included in the selected period if the human response fell within the period, even if AI Agent's last message did not. |
| Notes Only public messages are considered, so internal notes are excluded. If a human replies to a ticket AI Agent is handling without a formal handover, that message still counts as the first human response. For merged tickets, the response time is taken from the ticket they were merged into. | |
| Median response time | |
| Metric type | Median |
| Definition | Median time agents take to reply to customer messages across an entire conversation, not just the first reply. |
| How it's calculated | Calculated by measuring the time in seconds between each customer message and the agent reply that follows it, then taking the median across all exchanges in the selected period. When a customer sends several messages in a row, the time is measured from the first message in that series. |
| Notes Both open and closed tickets are included. Agent messages that don't follow a customer message, such as follow-ups, aren't counted. We use the median rather than the average so a few unusually slow tickets don't skew the number. | |
| Messages per ticket | |
| Metric type | Average |
| Definition | Average number of public messages exchanged in tickets closed during the selected period. |
| How it's calculated | Calculated by dividing the total number of public messages in qualifying closed tickets by the number of qualifying closed tickets. Customer messages, agent replies, Rule and auto-reply messages, and public AI Agent messages are all counted. |
| Notes A ticket qualifies if it has at least one customer message and at least one reply from an agent or a Rule. AI Agent messages don't qualify a ticket on their own, so a conversation between only a customer and AI Agent may be excluded. Internal notes are excluded. Messages are counted across the ticket's full history, up to 6 months before the start of the period. | |
| Messages received | |
| Metric type | Volume |
| Definition | Public messages your customers sent during the selected period. |
| How it's calculated | Calculated by counting the number of public customer messages sent within the selected period. |
| Notes A message is counted based on when it was sent. Messages on spam, trashed, and deleted tickets are excluded. | |
| Messages sent | |
| Metric type | Volume |
| Definition | Public messages sent from your helpdesk during the selected period. |
| How it's calculated | Calculated by counting every public message sent by agents (through the helpdesk, the API, or a phone integration) and by Rules within the selected period. Each message is counted individually, so multiple messages on the same ticket each add to the total. |
| Notes Internal notes are excluded. Messages on spam, trashed, and deleted tickets are excluded. Call logs from phone integrations, such as Aircall, are counted as messages. | |
| One touch tickets | |
| Metric type | Volume |
| Definition | Closed tickets resolved with a single reply. |
| How it's calculated | Calculated by counting the number of tickets closed during the selected period that have at least one customer message and exactly one reply from an agent or a Rule. |
| Notes Replies are counted across the ticket's full history, up to 6 months before the start of the period; the selected period only determines when the ticket was closed. Tickets created more than 6 months before the start of the period are excluded. Spam, trashed, and deleted tickets are excluded from the calculation. Call logs from phone integrations, such as Aircall, count as a reply. | |
| Open tickets | |
| Metric type | Volume |
| Definition | Tickets still open at the end of the selected period. |
| How it's calculated | Calculated by counting the number of tickets with an open status at the end of the selected period, among tickets created in the 180 days before the period end. |
| Notes Tickets created more than 180 days before the end of the period aren't counted. A ticket closed after the period ends still counts, since it was open at the end of the period. Reopened tickets count if they're open at the end of the period. Spam, trashed, and deleted tickets are excluded from the calculation. Because the count is a snapshot of the end of the period, changing only the start date doesn't change it. | |
| Resolution time | |
| Metric type | Median |
| Definition | Median time it takes to resolve a ticket, from the customer's first message to the ticket's latest close. |
| How it's calculated | Calculated by measuring the time in seconds from the customer's first message to the most recent time the ticket was closed, then taking the median across all qualifying tickets. A ticket is included in the selected period if it was closed (or closed again) during the period, regardless of when it was opened. |
| Notes Only tickets with at least one customer message and one reply from an agent or a Rule are included. If a ticket was closed more than once, only the most recent close is used. Time outside business hours is included. We use the median rather than the average so a few unusually slow tickets don't skew the number. | |
| Ticket handle time | |
| Metric type | Average |
| Definition | Average time agents spend actively working on a ticket, for tickets closed during the selected period. |
| How it's calculated | Calculated by dividing the total drafting time, working time, and phone call time on closed tickets by the number of closed tickets a human agent handled. A ticket counts as handled by a human if it has a public agent reply, a phone call, a drafted reply, or an internal note written by an agent. |
| Notes Tickets closed without any human involvement, such as bulk-closed tickets or tickets resolved by AI Agent, are excluded. Messages and internal notes from AI Agent or other automations don't count as human handling. If a work session is still open at the end of the period, its time stops counting at the end of the period. | |
| Tickets replied | |
| Metric type | Volume |
| Definition | Tickets that received at least one public reply from your helpdesk during the selected period. |
| How it's calculated | Calculated by counting the number of unique tickets with at least one public message sent by an agent (through the helpdesk, the API, or a phone integration) or by a Rule within the selected period. A ticket with multiple replies is counted once. |
| Notes Outbound tickets count, even if the customer never replied. Internal notes are excluded. Spam, trashed, and deleted tickets are excluded from the calculation. Call logs from phone integrations, such as Aircall, count as a reply. | |
| Tickets with achieved SLAs | |
| Metric type | Volume |
| Definition | Tickets that met their SLA during the selected period. |
| How it's calculated | Calculated by counting the number of unique tickets whose SLA deadline fell within the selected period and that met their SLA. |
| Notes Tickets with a pending SLA (one that hasn't been achieved or breached yet) are excluded. Each ticket is counted once, even if more than one SLA policy applies to it. Spam, trashed, and deleted tickets are excluded from the calculation. | |
| Tickets with breached SLAs | |
| Metric type | Volume |
| Definition | Tickets that missed their SLA during the selected period. |
| How it's calculated | Calculated by counting the number of unique tickets whose SLA was breached within the selected period. |
| Notes Tickets with a pending SLA (one that hasn't been achieved or breached yet) are excluded. Each ticket is counted once, even if more than one SLA policy applies to it. Spam, trashed, and deleted tickets are excluded from the calculation. | |
| Zero touch tickets | |
| Metric type | Volume |
| Definition | Closed tickets resolved without any reply. |
| How it's calculated | Calculated by counting the number of tickets closed during the selected period that have at least one customer message and no replies from an agent or a Rule. |
| Notes Replies are counted across the ticket's full history, up to 6 months before the start of the period; the selected period only determines when the ticket was closed. Tickets created more than 6 months before the start of the period are excluded. Spam, trashed, and deleted tickets are excluded from the calculation. Tickets with a call log from a phone integration, such as Aircall, don't count as zero touch. | |
Agent availability metrics
The Agent table at the bottom of the Overview tab includes metrics that show how long each agent was online, and how that time splits across availability statuses and phone calls. Custom availability statuses you've set up also appear as their own columns.
| Available time | |
| Metric type | Duration |
| Definition | Time an agent was online and set to Available. |
| How it's calculated | Calculated by adding up the time an agent had at least one Gorgias tab open while their availability status was Available, within the selected period. |
| Notes Time spent on a call or in call wrap-up is counted in On a call time and Call wrap-up time instead. Custom availability statuses have their own columns and aren't included. Only the Date and Agent filters affect this metric. | |
| Call wrap-up time | |
| Metric type | Duration |
| Definition | Time an agent spent in post-call wrap-up while online. |
| How it's calculated | Calculated by adding up the time between the start and end of each call wrap-up while the agent had at least one Gorgias tab open, within the selected period. |
| Notes Call wrap-up takes priority over the agent's availability status, so this time isn't also counted as Available or Unavailable time. Wrap-ups in progress at the start or end of the period are cut off at the period boundary. Only the Date and Agent filters affect this metric. | |
| On a call time | |
| Metric type | Duration |
| Definition | Time an agent spent on phone calls while online. |
| How it's calculated | Calculated by adding up the time from when an agent joined each call to when they left it, while they had at least one Gorgias tab open, within the selected period. |
| Notes Calls take priority over the agent's availability status, so this time isn't also counted as Available or Unavailable time. Calls in progress at the start of the period are counted from the period start if they began within the previous 7 days. Only the Date and Agent filters affect this metric. | |
| Active time | |
| Metric type | Duration |
| Definition | Estimated time the agent was actively using Gorgias during the selected period. Only time when Gorgias is the focused tab counts; time in another tab or app is not included. |
| How it's calculated | Estimated from time when the Gorgias tab is in focus during the selected period. |
| Logged-on time | |
| Metric type | Duration |
| Definition | Time the agent had at least one Gorgias tab open during the selected period, including Available and Unavailable time. |
| How it's calculated | Calculated by adding up the agent's online time across every status: Available, Unavailable, any custom availability statuses, On a call, and Call wrap-up, within the selected period. |
| Notes If the selected period ends in the future, time is counted up to the current time. Only the Date and Agent filters affect this metric. | |
| Closed tickets per active hour | |
| Metric type | Rate |
| Definition | Number of closed tickets assigned to the agent per hour of estimated active time during the selected period. |
| How it's calculated | Divide the number of closed tickets assigned to the agent by Active time, in hours, during the same period. |
| Closed tickets per logged-on hour | |
| Metric type | Rate |
| Definition | Number of closed tickets assigned to the agent per hour the agent had a Gorgias tab open during the selected period, including Available and Unavailable time. |
| How it's calculated | Divide the number of closed tickets assigned to the agent by Logged-on time, in hours, during the same period. |
| Messages sent per active hour | |
| Metric type | Rate |
| Definition | Number of messages sent by the agent per hour of estimated active time during the selected period. |
| How it's calculated | Divide the number of messages sent by the agent by Active time, in hours, during the same period. |
| Messages sent per logged-on hour | |
| Metric type | Rate |
| Definition | Number of messages sent by the agent per hour the agent had a Gorgias tab open during the selected period, including Available and Unavailable time. |
| How it's calculated | Divide the number of messages sent by the agent by Logged-on time, in hours, during the same period. |
| Tickets replied per active hour | |
| Metric type | Rate |
| Definition | Number of tickets replied to by the agent per hour of estimated active time during the selected period. |
| How it's calculated | Divide the number of tickets replied to by the agent by Active time, in hours, during the same period. |
| Tickets replied per logged-on hour | |
| Metric type | Rate |
| Definition | Number of tickets replied to by the agent per hour the agent had a Gorgias tab open during the selected period, including Available and Unavailable time. |
| How it's calculated | Divide the number of tickets replied to by the agent by Logged-on time, in hours, during the same period. |
| Unavailable time | |
| Metric type | Duration |
| Definition | Time an agent was online and set to Unavailable. |
| How it's calculated | Calculated by adding up the time an agent had at least one Gorgias tab open while their availability status was Unavailable, within the selected period. |
| Notes Custom unavailable statuses, such as Lunch, have their own columns and aren't included. Time spent on a call or in call wrap-up is counted in On a call time and Call wrap-up time instead. Only the Date and Agent filters affect this metric. | |
Ticket status chart
The Ticket status Sankey chart shows how the tickets created during the selected period split between inbound and outbound, and what status each ticket had at the end of the period. Every ticket is counted in exactly one status. When a ticket matches more than one, it's counted in the first that applies, in this order: Spam, Deleted & trashed, Closed, Reopened, Open.
The Email, Chat, Social, Voice, and SMS tabs of the Channels report each have their own ticket status chart, which uses the same definitions, filtered to that channel.
| All created tickets | |
| Metric type | Volume |
| Definition | Every ticket created in the selected period (inbound and outbound), including spam, deleted, and trashed tickets. |
| How it's calculated | Calculated by counting the number of unique tickets with a creation date within the selected period. |
| Notes Unlike the Created tickets metric, this total includes spam, trashed, and deleted tickets, so the two numbers can differ. | |
| Inbound | |
| Metric type | Volume |
| Definition | Created tickets started by a customer. |
| How it's calculated | Calculated by counting the number of created tickets where the first message came from the customer, plus all spam tickets. |
| Notes Spam tickets always appear under Inbound, even if your team started the ticket. | |
| Outbound | |
| Metric type | Volume |
| Definition | Created tickets started by your team. |
| How it's calculated | Calculated by counting the number of created tickets where the first message came from your company, including messages sent by AI Agent or other automations. Spam tickets are excluded. |
| Notes Spam tickets your team started are counted under Inbound instead. | |
| Open | |
| Metric type | Volume |
| Definition | Created tickets that are still open at the end of the selected period. |
| How it's calculated | Calculated by counting the number of created tickets with an open status at the end of the selected period that were not reopened during the period. |
| Notes Open and Reopened together make up every created ticket still open at the end of the period. This differs from the Open tickets metric, which counts open tickets created in the 180 days before the end of the period. | |
| Reopened | |
| Metric type | Volume |
| Definition | Created tickets that were closed, then reopened, and are still open at the end of the selected period. |
| How it's calculated | Calculated by counting the number of created tickets that were closed and reopened during the selected period and have an open status at the end of it. |
| Notes A ticket that was reopened and then closed again counts as Closed. | |
| Closed | |
| Metric type | Volume |
| Definition | Created tickets that are closed at the end of the selected period. |
| How it's calculated | Calculated by counting the number of created tickets with a closed status at the end of the selected period. |
| Notes Includes tickets that were reopened earlier in the period and then closed again. | |
| Deleted & trashed | |
| Metric type | Volume |
| Definition | Created tickets that were deleted or moved to trash by the end of the selected period. |
| How it's calculated | Calculated by counting the number of created tickets that are deleted or in the trash at the end of the selected period. |
| Notes A ticket restored from the trash before the end of the period is counted by its status instead. | |
| Spam | |
| Metric type | Volume |
| Definition | Created tickets marked as spam at the end of the selected period. |
| How it's calculated | Calculated by counting the number of created tickets with a spam status at the end of the selected period. |
| Notes Spam takes priority over every other status, so a spam ticket is never counted as open, closed, deleted, or trashed. | |
AI Agent funnel chart
The AI Agent funnel Sankey chart shows how the tickets created during the selected period moved through AI Agent: whether they came from a channel AI Agent supports, whether AI Agent worked on them, and how they ended. Every ticket is counted once at each step of the funnel.
| All created tickets | |
| Metric type | Volume |
| Definition | Every ticket created in the selected period (inbound and outbound), including spam, deleted, and trashed tickets. |
| How it's calculated | Calculated by counting the number of unique tickets with a creation date within the selected period. |
| Notes This total matches All created tickets in the Ticket status chart. | |
| AI Agent-eligible | |
| Metric type | Volume |
| Definition | Tickets from a channel AI Agent can support. |
| How it's calculated | Calculated by counting the number of created tickets from a channel AI Agent supports, plus any tickets AI Agent worked on in another channel. |
| Notes Spam, deleted, and trashed tickets follow the same rules as other tickets. | |
| AI Agent-ineligible | |
| Metric type | Volume |
| Definition | Tickets from channels AI Agent doesn't support. |
| How it's calculated | Calculated by counting the number of created tickets from a channel AI Agent doesn't support, where AI Agent didn't work on the ticket. |
| Notes If AI Agent worked on a ticket from an unsupported channel, the ticket counts as AI Agent-eligible instead. | |
| AI-ignore tag | |
| Metric type | Volume |
| Definition | Tickets AI Agent skipped because they carry your AI-ignore tag. |
| How it's calculated | Calculated by counting the number of AI Agent-eligible tickets AI Agent didn't work on that have the AI-ignore tag. |
| Notes If AI Agent worked on a ticket before the tag was added, the ticket counts as Covered by AI Agent instead. | |
| Covered by AI Agent | |
| Metric type | Volume |
| Definition | Tickets AI Agent worked on, including tickets it answered and tickets it handed over to your team. |
| How it's calculated | Calculated by counting the number of AI Agent-eligible tickets where AI Agent sent at least one message, or a billable automated interaction was recorded, during the selected period. |
| Notes Tickets handled by an AI Journey aren't counted as covered, since Journeys are flows you set up rather than AI Agent conversations. | |
| Not covered by AI Agent | |
| Metric type | Volume |
| Definition | Tickets from an AI Agent-eligible channel that AI Agent didn't work on. |
| How it's calculated | Calculated by counting the number of AI Agent-eligible tickets AI Agent didn't work on that don't have the AI-ignore tag. |
| Notes Tickets with the AI-ignore tag are counted separately. | |
| Resolved by AI | |
| Metric type | Volume |
| Definition | Tickets AI Agent closed without handing over. These are billed as automated interactions. |
| How it's calculated | Calculated by counting the number of covered tickets with a billable automated interaction during the selected period. |
| Notes If AI Agent hands over a ticket after an automated interaction was billed, the ticket still counts as Resolved by AI. | |
| Handover by AI Agent | |
| Metric type | Volume |
| Definition | Tickets AI Agent passed to a human agent. |
| How it's calculated | Calculated by counting the number of covered tickets that weren't billed as an automated interaction, where AI Agent's latest outcome was a handover. |
| Notes Each ticket is counted once, even if AI Agent handed it over more than once. | |
| Reopened by your team | |
| Metric type | Volume |
| Definition | Tickets AI Agent worked on where a human agent stepped in after AI Agent's last message. |
| How it's calculated | Calculated by counting the number of covered tickets that weren't billed or handed over, where a human agent replied after AI Agent's latest message. |
| Notes These tickets aren't billed as automated interactions. | |
| No reply from shopper | |
| Metric type | Volume |
| Definition | Tickets where AI Agent reached out with a Shopping Assistant engagement feature and the shopper never responded. |
| How it's calculated | Calculated by counting the number of covered tickets started by AI Agent outreach that didn't lead to a billable conversation with the shopper. |
| Notes These tickets aren't billed as automated interactions. | |
| Open and in progress | |
| Metric type | Volume |
| Definition | Open tickets AI Agent is working on that are still in progress. |
| How it's calculated | Calculated by counting the number of covered tickets that don't match any other outcome: they haven't been resolved by AI, handed over, picked up by your team, or closed with no reply from the shopper. |
| Notes These tickets move to another outcome once AI Agent finishes working on them. | |
CSAT tab
The CSAT tab of the Overview report shows how customers rate your support in satisfaction surveys, and how your agents score in Auto-QA reviews. Every survey metric is counted in the period the survey was sent, not when the ticket was closed or when the customer responded.
The Average CSAT metric also appears on this tab. To learn how it's calculated, see Average CSAT under Overview tab.
| Auto-QA Accuracy | |
| Metric type | Average |
| Definition | Average score for the factual accuracy of agent responses, across manually reviewed tickets. |
| How it's calculated | Calculated by dividing the sum of accuracy scores by the number of tickets with an accuracy score in the selected period. |
| Notes Accuracy is scored manually; AI doesn't evaluate it. If a ticket is re-scored, only its latest score is used. A period with no reviewed tickets shows no value. | |
| Auto-QA Brand voice | |
| Metric type | Average |
| Definition | Average score for how well agent responses match your brand vocabulary, greetings, sign-offs, and tone of voice, across manually reviewed tickets. |
| How it's calculated | Calculated by dividing the sum of brand voice scores by the number of tickets with a brand voice score in the selected period. |
| Notes Brand voice is scored manually; AI doesn't evaluate it. If a ticket is re-scored, only its latest score is used. A period with no reviewed tickets shows no value. | |
| Auto-QA Communication | |
| Metric type | Average |
| Definition | Average score for agents' empathy, clarity, patience, positivity, and adaptability, across reviewed tickets. |
| How it's calculated | Calculated by dividing the sum of communication scores by the number of tickets with a communication score in the selected period. |
| Notes Communication is scored automatically by AI. Only closed tickets with at least one customer message and one agent message are evaluated automatically. If a ticket is re-scored, only its latest score is used. A period with no reviewed tickets shows no value. | |
| Auto-QA Efficiency | |
| Metric type | Average |
| Definition | Average score for how quickly agents handled tickets and how few touches they needed, across manually reviewed tickets. |
| How it's calculated | Calculated by dividing the sum of efficiency scores by the number of tickets with an efficiency score in the selected period. |
| Notes Efficiency is scored manually; AI doesn't evaluate it. If a ticket is re-scored, only its latest score is used. A period with no reviewed tickets shows no value. | |
| Auto-QA Internal compliance | |
| Metric type | Average |
| Definition | Average score for whether agents followed your internal processes, such as using the correct macros, tags, escalation, and merging procedures, across manually reviewed tickets. |
| How it's calculated | Calculated by dividing the sum of internal compliance scores by the number of tickets with an internal compliance score in the selected period. |
| Notes Internal compliance is scored manually; AI doesn't evaluate it. If a ticket is re-scored, only its latest score is used. A period with no reviewed tickets shows no value. | |
| Auto-QA Language proficiency | |
| Metric type | Average |
| Definition | Average AI-generated score (1-5) for agents' spelling, grammar, and syntax in the language of the conversation, across reviewed tickets. |
| How it's calculated | Calculated by dividing the sum of language proficiency scores by the number of tickets with a language proficiency score in the selected period. |
| Notes Language proficiency is scored automatically by AI. Only closed tickets with at least one customer message and one agent message are evaluated automatically. If a ticket is re-scored, only its latest score is used. A period with no reviewed tickets shows no value. | |
| Auto-QA Resolution completeness rate | |
| Metric type | Percentage |
| Definition | Percentage of reviewed tickets where the agent addressed all of the customer's questions. |
| How it's calculated | Calculated by dividing the number of tickets marked as complete by the number of tickets with a resolution completeness result in the selected period, then multiplying by 100. |
| Notes Only closed tickets with at least one customer message and one agent message are evaluated automatically. Manually recorded results are also included. If a ticket is re-scored, only its latest score is used. A period with no reviewed tickets shows no value. | |
| Auto-QA Reviewed tickets | |
| Metric type | Volume |
| Definition | Closed tickets that were reviewed, automatically or manually, during the selected period. |
| How it's calculated | Calculated by counting the number of closed tickets that received an automatic or manual review within the selected period. |
| Notes Only closed tickets with at least one customer message and one agent message are evaluated automatically. | |
| Response rate | |
| Metric type | Percentage |
| Definition | Percentage of surveys sent during the selected period that received a rating. |
| How it's calculated | Calculated by dividing the number of unique tickets with an answered survey by the number of unique tickets with a survey sent within the selected period, then multiplying by 100. |
| Notes Surveys are counted in the period they were sent, not when the customer responded. Surveys that were sent but never answered only count toward the total sent. A period with no surveys sent shows no value. | |
| Surveys answered | |
| Metric type | Volume |
| Definition | Surveys sent during the selected period that received a rating. |
| How it's calculated | Calculated by counting the number of unique tickets with a survey sent within the selected period that received a customer rating (1-5). |
| Notes Surveys are counted in the period they were sent, not when the customer responded. Each ticket is counted once, even if a survey was sent to it more than once. | |
| Surveys sent | |
| Metric type | Volume |
| Definition | Tickets that were sent a satisfaction survey during the selected period. |
| How it's calculated | Calculated by counting the number of unique tickets where a satisfaction survey was sent within the selected period. |
| Notes Each ticket is counted once, even if a survey was sent to it more than once. Surveys are sent automatically after a ticket is resolved. | |
Ticket attributes tab
The Ticket attributes tab of the Overview report measures ticket volume by intent, tag, or ticket field value. Its breakdown tables use the following metrics for each row. Each attribute is counted against its own date: intents use the date the ticket was created, tags use the date the tag was added, and ticket field values use the latest value set for the field during the period.
| Tickets | |
| Metric type | Volume |
| Definition | Tickets associated with each intent, tag, or ticket field value. |
| How it's calculated | Calculated by counting the number of unique tickets in each row: Intents: tickets created during the selected period, grouped by the intent assigned when the ticket was created. Tags: tickets where the tag was added during the selected period and is still applied. Ticket field values: tickets grouped by the latest value set for the selected ticket field during the period. |
| Notes Each ticket counts once per row. A ticket with several tags appears in the row of each tag. Tickets with no detected intent appear in the No Assigned Intent row. Parent ticket field values add up the tickets of the values below them. | |
| Share of total | |
| Metric type | Percentage |
| Definition | Each row's share of the tickets in the table. |
| How it's calculated | Calculated by dividing the number of tickets in a row by the total number of tickets in the table, then multiplying by 100. For intents, the total is every ticket created during the period, including tickets with no intent. For tags, the total is tickets with at least one tag. For ticket field values, the total is tickets with a value for the selected field. |
| Notes A ticket can carry more than one tag, so shares in the Tags table can add up to more than 100%. If the total is zero, no percentage is shown. | |
| Trend | |
| Metric type | Percentage |
| Definition | Percentage change in tickets for each row compared to the previous period of the same length. |
| How it's calculated | Calculated by subtracting the previous period's tickets from the current period's tickets, dividing the result by the previous period's tickets, then multiplying by 100. |
| Notes An increase shows as a positive trend and a decrease as a negative trend. If a row had no tickets in the previous period, no trend is shown, unless it also has no tickets in the current period, in which case the trend is 0%. | |
Busiest times tab
The Busiest times tab of the Overview report maps one metric at a time on a heatmap of hours and weekdays. Its metrics use the same definitions as Created tickets, Tickets replied, Closed tickets, Messages received, and Messages sent on the Overview tab. Each cell adds up every occurrence of that hour and weekday in the selected period, shown in your own timezone.
Channels report
The Channels report has a tab for each channel: Email, Chat, Social, Voice, and SMS. The ticket metrics on each tab use the same definitions as the Overview tab, filtered to that channel's tickets. Each tab's tickets created metric, such as Email tickets created, follows the Created tickets definition, and each tab's ticket status chart follows the Ticket status chart definitions.
The Email tab includes both email and contact form tickets, and the Social tab combines every social channel you've connected.
Voice tab
Along with Voice tickets created and the Voice ticket status chart, the Voice tab reports on the phone calls themselves. The Call outcome chart splits Total calls into the inbound and outbound call metrics below. Voice data is available from January 1, 2024.
| Abandoned calls | |
| Metric type | Volume |
| Definition | Inbound calls where the caller waited at least 10 seconds and hung up before any agent's phone rang. |
| How it's calculated | Calculated by counting the number of unique inbound calls where no agent's phone rang, the caller didn't request a callback, the system didn't end the call, and the caller hung up after waiting at least 10 seconds. |
| Notes Calls where the caller hung up in under 10 seconds count as Cancelled calls. Calls where an agent's phone rang count as Missed calls. This metric is hidden when you filter by agent, since no agent is linked to these calls. | |
| % Abandoned calls | |
| Metric type | Percentage |
| Definition | Percentage of inbound calls that were abandoned. |
| How it's calculated | Calculated by dividing the number of abandoned calls by the number of inbound calls, then multiplying by 100. |
| Notes The total is all inbound calls, not just unanswered ones. This metric is hidden when you filter by agent, since no agent is linked to these calls. | |
| Achievement rate | |
| Metric type | Percentage |
| Definition | Percentage of call queue exposures where your team met the SLA. |
| How it's calculated | Calculated by dividing the number of queue exposures that met the SLA by the number of queue exposures with an SLA result, then multiplying by 100. A queue exposure is one call passing through one queue, so a call routed through several queues is evaluated once for each queue. |
| Notes Queue exposures are excluded when the queue has no SLA set, the call went unanswered after less than 10 seconds of waiting, the call came in outside business hours and the queue doesn't track SLA outside business hours, or the caller requested a callback within the SLA threshold. A call that overflows to another queue counts as a breach for the queue it left. A period with no qualifying queue exposures shows no value. | |
| Average talk time | |
| Metric type | Average |
| Definition | Average time agents spent talking to customers on answered calls. |
| How it's calculated | Calculated by dividing the total talk time across answered inbound and outbound calls by the number of answered calls. |
| Notes Unanswered calls are excluded. Hold, mute, and transfer ringing time all count as talk time. For a transferred call, the full talk time is credited to the assigned agent. | |
| Average wait time | |
| Metric type | Average |
| Definition | Average time inbound callers waited before an agent answered or the call ended. |
| How it's calculated | Calculated by dividing the total wait time across inbound calls by the number of inbound calls. For answered calls, wait time runs from when the call arrived to when an agent picked up. For unanswered calls, it runs until the call ended, for example when the caller hung up, reached voicemail, or requested a callback. |
| Notes Wait time includes time spent in IVR menus and recorded messages before the call reaches a queue. Calls still in progress are excluded. | |
| Callback requests | |
| Metric type | Volume |
| Definition | Inbound calls where the caller requested a callback. |
| How it's calculated | Calculated by counting the number of unique inbound calls where the caller chose the callback option. |
| Notes A call counts as a callback request even if an agent's phone already rang. Callback requests are included in Unanswered calls, so don't add the two together. | |
| % Callback requests | |
| Metric type | Percentage |
| Definition | Percentage of inbound calls where the caller requested a callback. |
| How it's calculated | Calculated by dividing the number of callback requests by the number of inbound calls, then multiplying by 100. |
| Notes The total is all inbound calls, not just unanswered ones. | |
| Calls with achieved SLAs | |
| Metric type | Volume |
| Definition | Call queue exposures where your team met the SLA. |
| How it's calculated | Calculated by counting the number of unique queue exposures that met the queue's SLA threshold. A queue exposure is one call passing through one queue. |
| Notes A call routed through several queues can add more than one queue exposure. A queue exposure that overflows to another queue counts as breached. Queue exposures without an SLA result are excluded. | |
| Calls with breached SLAs | |
| Metric type | Volume |
| Definition | Call queue exposures where your team missed the SLA. |
| How it's calculated | Calculated by counting the number of unique queue exposures that breached the queue's SLA, for example because the wait went over the SLA threshold or the call overflowed to another queue. A queue exposure is one call passing through one queue. |
| Notes A call that moves between queues can count as breached in one queue and achieved in another. Queue exposures without an SLA result are excluded. | |
| Cancelled calls | |
| Metric type | Volume |
| Definition | Inbound calls that ended in under 10 seconds, before any agent's phone rang. |
| How it's calculated | Calculated by counting the number of unique inbound calls that ended in less than 10 seconds before any agent's phone rang. Calls with no recorded duration also count as cancelled. |
| Notes Calls where the caller waited 10 seconds or more count as Abandoned calls. This metric is hidden when you filter by agent, since no agent is linked to these calls. | |
| % Cancelled calls | |
| Metric type | Percentage |
| Definition | Percentage of inbound calls that were cancelled. |
| How it's calculated | Calculated by dividing the number of cancelled calls by the number of inbound calls, then multiplying by 100. |
| Notes The total is all inbound calls, not just unanswered ones. This metric is hidden when you filter by agent, since no agent is linked to these calls. | |
| Inbound answered | |
| Metric type | Volume |
| Definition | Inbound calls answered by an agent. |
| How it's calculated | Calculated by counting the number of unique inbound calls that were answered. |
| Notes Each call counts once, even if it rang several agents. If an agent answers a call and then transfers it to someone who doesn't pick up, the call still counts as answered. In the Agent table, only the agent who picked up is credited. Calls still in progress are excluded. | |
| Inbound calls | |
| Metric type | Volume |
| Definition | Inbound calls received during the selected period. |
| How it's calculated | Calculated by counting the number of unique inbound calls, however they ended. |
| Notes Includes answered, missed, abandoned, cancelled, and callback-requested calls. Calls still in progress are excluded. | |
| Inbound declined | |
| Metric type | Volume |
| Definition | Inbound calls an agent actively declined. |
| How it's calculated | Calculated by counting the number of unique inbound calls each agent declined. Each decline is credited to the agent who declined the call. |
| Notes If several agents decline the same call, each of them is credited. A declined call can still be answered by another agent. Calls still in progress are excluded. | |
| Inbound transfers | |
| Metric type | Volume |
| Definition | Inbound calls an agent transferred to another agent, a queue, or an external number. |
| How it's calculated | Calculated by counting the number of unique inbound calls where a transfer started by the agent was accepted. Each transfer is credited to the agent who started it. |
| Notes If a call is transferred more than once, each agent who transferred it is credited once. Outbound call transfers are excluded. Calls still in progress are excluded. | |
| Missed calls | |
| Metric type | Volume |
| Definition | Inbound calls that weren't answered after an agent's phone rang or after the system ended the call. |
| How it's calculated | Calculated by counting the number of unique unanswered inbound calls where an agent's phone rang or the system ended the call, including calls routed to voicemail. |
| Notes Calls an agent declined that nobody else answered, and calls forwarded to an external number that wasn't answered, also count as missed. Callback requests are excluded. In the Agent table, Inbound missed counts every call that rang an agent and that they didn't pick up, even if a teammate answered it, so it can be higher than this metric. | |
| % Missed calls | |
| Metric type | Percentage |
| Definition | Percentage of inbound calls that were missed. |
| How it's calculated | Calculated by dividing the number of missed calls by the number of inbound calls, then multiplying by 100. |
| Notes The total is all inbound calls, not just unanswered ones. | |
| Outbound calls | |
| Metric type | Volume |
| Definition | Outbound calls placed during the selected period. |
| How it's calculated | Calculated by counting the number of unique outbound calls. |
| Notes Calls still in progress are excluded. | |
| Total calls | |
| Metric type | Volume |
| Definition | Inbound and outbound calls during the selected period. |
| How it's calculated | Calculated by counting the number of unique inbound and outbound calls. Each call is counted once, no matter how many agents it rang. |
| Notes Calls still in progress are excluded. | |
| Unanswered calls | |
| Metric type | Volume |
| Definition | Inbound calls that ended without being answered. |
| How it's calculated | Calculated by counting the number of unique inbound calls that were missed, abandoned, cancelled, or ended with a callback request. |
| Notes Equals the sum of Missed calls, Abandoned calls, Cancelled calls, and Callback requests, so don't add those on top of it. Outbound calls are excluded. | |
Help Center report
The Help Center report shows how customers use your Help Center articles and search. Each metric counts every tracked event, so repeat views, searches, and clicks by the same visitor all count separately.
| Article views | |
| Metric type | Volume |
| Definition | Times your Help Center articles were viewed. |
| How it's calculated | Calculated by counting every tracked article view within the selected period. Repeat views by the same visitor count separately. |
| Notes This isn't a count of unique visitors. Traffic identified as bot activity is excluded. The Views column in the Articles table shows this metric for each article. | |
| Articles clicked | |
| Metric type | Volume |
| Definition | Clicks on articles from Help Center search results. |
| How it's calculated | Calculated by counting every tracked click on an article in search results within the selected period. Repeat clicks count separately. |
| Notes In the Search terms table, this metric is shown for each search term, and terms that returned results but got no clicks show 0. Traffic identified as bot activity is excluded. | |
| Click-through rate | |
| Metric type | Percentage |
| Definition | Clicks on search results as a percentage of searches that returned results. |
| How it's calculated | Calculated by dividing Articles clicked by Searches with results, then multiplying by 100. |
| Notes One search can lead to several clicks, so the rate can be higher than 100%. In the Search terms table, the rate is calculated for each search term. Traffic identified as bot activity is excluded. | |
| No-result searches | |
| Metric type | Volume |
| Definition | Help Center searches that returned no articles. |
| How it's calculated | Calculated by counting every search within the selected period that returned zero articles. Repeat searches for the same term count separately. |
| Notes Searches with no recorded search term or result count are excluded. The result count reflects the articles available at the time of the search, so publishing new content doesn't change past results. Traffic identified as bot activity is excluded. | |
| Searches | |
| Metric type | Volume |
| Definition | Searches submitted in your Help Center. |
| How it's calculated | Calculated by counting every tracked search submitted within the selected period, including repeat searches by the same visitor. |
| Notes Searches count whether or not they returned results, and even if the search term wasn't recorded. Traffic identified as bot activity is excluded. | |
| Searches with results | |
| Metric type | Volume |
| Definition | Help Center searches that returned at least one article. |
| How it's calculated | Calculated by counting every search within the selected period that returned at least one article. Repeat searches for the same term count separately. |
| Notes In the Search terms table, this metric is shown as Search count for each search term. Traffic identified as bot activity is excluded. | |