Messaging → Email Escalations: What Qvasa Measures and Where to See It
When a messaging conversation can't be finished in chat, the ticket gets moved to email. Qvasa now records that moment directly — when it happened, who escalated, why, and what happened next — and turns it into dashboard metrics: how many tickets escalate, at what rate, which agents and groups are escalating, how fast the email side picks them up, and whether they get solved.
What Counts as an Escalation
An escalation is the moment a ticket leaves a messaging conversation for email (or another support channel) for the first time. Qvasa watches the ticket's channel history and records the switch from the messaging run to the support channel. A ticket escalates at most once: bouncing back to messaging and then to email again doesn't create a second escalation.
Each escalation is classified by how it happened:
Routing trigger — a Zendesk trigger transformed the ticket to email. Qvasa also keeps the trigger's routing reason.
Organic email — the conversation simply continued over email.
Undetermined — the switch was seen in real time before Zendesk's audit explained it. If the audit later supplies the reason, the record is upgraded in place.
What Qvasa Records for Every Escalation
When it escalated, how long the ticket had been in messaging, and the time from the first agent chat to the escalation.
Who escalated — the last messaging agent before the switch: the agent whose chat started last before it, or, if nobody chatted, the agent assigned during the messaging conversation. Qvasa notes which of those two it used. Bot-only conversations with no agent are still counted as escalations, but credited to nobody.
The situation at the switch — the assignee, the Zendesk group, the agent who performed the change (when it wasn't a trigger), and how many agent chats happened before it.
The first email reply — the first public agent reply sent over email (or another non-messaging channel) at or after the switch, who sent it, and how long it took. Replies typed into the messaging composer, trigger comments and system comments don't count.
The first solve after the escalation — when it was solved, by which assignee, and whether that was the same agent who escalated it. A reopen and re-solve never overwrite the first solve.
All of this is stored once per ticket and is what every widget below reads from, so the numbers agree with each other and with the table.
Before You Start
Escalation tracking is switched on by Qvasa for your account. It needs ticket channel history, and the widgets don't appear until it's on — if you can't find them, reach out and we'll enable it.
Tracking is forward-only. Escalations are recorded from the moment tracking is enabled; there is no historical backfill. Pick date ranges that start after that day.
Sunshine Conversations message tracking is what lets Qvasa credit the escalating agent from their chats and compute the chat-based rates. Most messaging accounts already have it.
Anyone who can edit a dashboard can add these widgets. Search the widget library for "escalation" — the quick stats, charts and by-agent breakdowns are all there.
Step 1: The Headline Numbers
Add Quick stat widgets for the numbers you want to watch. Every count has a ticket drill-in, so you can click through to the tickets behind it.
Messaging → Email Escalations — how many tickets escalated in the date range.
First Agent Chats Escalated to Email and its Rate — the share of messaging tickets whose first agent chat ended that later escalated. Same denominator as First Agent Chat Solve Rate, so the two read as a pair. Lower is better.
Avg Time from First Agent Chat to Email Escalation — how long agents worked the chat before handing it to email.
Avg Time to First Email Reply After Escalation — how long the customer waited for the first email after the hand-off.
Escalations Without an Email Reply — escalated tickets nobody has replied to over email yet.
Escalated Tickets Solved and Escalated Tickets Solved by the Escalating Agent — whether escalated tickets get closed, and whether the escalator closed them. A percentage quick stat of the second over the first gives the escalator's share.
Escalation counts are dated by when the ticket escalated. The two solved counts are dated by when the ticket was first solved after escalating, so a ticket that escalated on Monday and was solved on Tuesday counts under Tuesday there.
Step 2: Who Is Escalating
Add a Donut chart (shown as a horizontal bar, donut or simple table) to break escalations down by people:
Messaging → Email Escalations by Agent — credited to the escalating agent (the last messaging agent).
Chats Ended in Email Escalation by Agent and its Rate — of each agent's chats, how many (and what share) were the last chat before an escalation. This is the fairest per-agent view because it's normalised by how many chats the agent handled. Lower is better.
Messaging → Email Escalations by Group — the Zendesk group the ticket was in at the moment of the switch.
First Email Replies After Escalation by Agent and Avg Time to First Email Reply by Agent — the email side of the hand-off: who picks escalated tickets up and how fast.
Escalated Tickets Solved by Agent — who closes them.
Agent filters follow the credit. When you filter a dashboard or widget by agent, group, skill or agent tag, each escalation metric narrows on the agent that metric credits: the escalator for escalation counts, the replier for first-reply metrics, the solver for solve metrics.
Step 3: Trends Over Time
Add a Basic chart for Messaging → Email Escalations Over Time, First Agent Chats Escalated to Email Rate Over Time or Avg Time to First Email Reply After Escalation Over Time. The columns of the escalations chart add up to the quick stat for the same date range. For a per-agent trend, add a Multi data chart — Messaging → Email Escalations by Agent Over Time or Chats Ended in Email Escalation Rate by Agent Over Time — and stack it.
Step 4: The Escalations Table
For the full record, add a Configurable table and pick Messaging → Email Escalations under Ticket Management. One row per escalated ticket, newest first, with every column Qvasa recorded: escalated at, ticket, escalated by (and how that was determined), escalation kind and routing reason, from/to channel, time in messaging, time from first agent chat, agent chats before escalation, assignee and group at escalation, who changed it, first email reply at/by and the time to it, solved after escalation at/by, solved by escalator, and links to the ticket in Qvasa and Zendesk. Choose the columns you want, filter it like any ticket table, and export it to CSV.
Step 5: Escalation Columns on the Agents Table
Every by-agent metric above is also an Agents Table column, so you can put escalations next to the rest of an agent's numbers — chats handled, replies, solves — in one row per agent and sort by any of them. Click a cell to inspect the tickets behind it.
First Chat Solve No Longer Credits Escalated Tickets
Previously a single messaging chat that was escalated to email and then solved over email could count as a First Chat Solve, because the ticket was solved after one chat. With escalation tracking on, an escalated ticket stays in the denominator but is never a solve — across the First Chat Solve quick stats, donuts, charts, the agents table and scorecards. Expect First Chat Solve Rate to drop slightly the day tracking is enabled; that's the corrected number.
FAQ
Why don't my escalation numbers go back further? Tracking records escalations from the day it was enabled on your account. There is no backfill, so earlier date ranges show nothing.
A ticket escalated but shows no escalating agent. Why? Nobody chatted and nobody was assigned during the messaging conversation — typically a bot-only conversation. It counts in the totals but not in any by-agent breakdown.
Why is the escalation kind "Undetermined"? The switch was seen live before Zendesk's audit explained it. If the audit later shows a routing trigger at the same moment, the kind is upgraded automatically.
An agent replied over email but the escalation still shows no first reply. Only public agent replies sent over a non-messaging channel count. Replies typed into the messaging composer, trigger comments and system comments don't. A reply from an agent Qvasa didn't know at the time can't be credited either.
The escalations count and the table row count don't match my by-agent donut. The donut only includes escalations with an attributable agent; the count and the table include bot-only escalations too.
Which widgets need Sunshine Conversations message tracking? The ones built on chats — First Agent Chats Escalated to Email (count and rate), Avg Time from First Agent Chat, and Chats Ended in Email Escalation by Agent (count and rate). Everything else needs only escalation tracking.