Satisfaction surveys (CSAT)

When a ticket is closed, the requester receives an e-mail with five numbered lines, one per rating from one to five stars. Each line is a link of its own, and one click is the whole answer.

The mail does not go out immediately. The system waits an hour after closing, and from then on a background service sends the surveys that are due every ten minutes. The hour is deliberate: a ticket that is reopened straight away should not trigger a survey.

There is exactly one survey per ticket. Even if a ticket is reopened and closed again later, the system does not ask a second time.

The link needs no customer account and is valid for 30 days. Until then your customer can change the rating — a misclick on the wrong star is more common than abuse.

A comment is optional. Clicking a star is already a rating; anyone who wants to add something finds a field for it on the page and confirms with “Update rating”.

The page shows only the number and the title of the ticket. Description, comments and history are not on it: the link is a right to rate, not a right to read — it can be forwarded, or land in a shared mailbox.

The click from the mail only writes the rating once the page has loaded. That is why virus scanners and preview fetchers do not rate your tickets: they fetch the address, but they run no JavaScript. For a human being it still is one click.

The rating that comes back sits on the ticket, where agents and administrators of the responsible team can see it. The customer never sees it there, not even their own.

Not every closed ticket is asked. Without an address for the requester no mail goes out at all. Nor does a survey go out if the requester closed the ticket themselves or if their address is on the blocklist. Merged duplicate reports as well as the reports attached to a major incident stay out of it too — resolving an incident closes every attached report with a single click, and without that exception each reporter would be surveyed about the same piece of work. External people also only receive the survey after giving consent. The card “Who receives the survey” explains who counts as external.

Every survey e-mail ends with a link to unsubscribe. After a rating, the survey page also shows the link “Stop receiving surveys”. The card “Consent and unsubscribe” explains both.

The survey has exactly one place to set it up: under “Settings → General”, in the card “Customer Satisfaction Score (CSAT)”. This card explains the three controls for sending. The field “Own e-mail domains” on the same card is explained in the card “Who receives the survey”.

“Send satisfaction surveys” switches sending on; it is off as a factory setting. Only tickets closed after you switch it on are surveyed — otherwise your entire backlog would receive a mail in one go.

If you switch it off again, the ratings you already have stay visible. Only nothing new goes out any more.

Above the switches you see the address the links are built from. It is there to be checked, not to be edited: you change it in the one place where it is maintained, and the hint next to it takes you there.

The middle switch, “Per-agent evaluation”, belongs to the report. What it does there, and why it is off as a factory setting, is on the card about the report. It stays usable even while the survey is switched off.

“At most one survey per requester within” limits how often the same person is asked. The factory setting is 7 days: someone who reports several tickets within that window is still asked only once.

With 0 you ask on every closed ticket. For an internal helpdesk that is usually too much, because the same people report again and again; a customer desk with many different senders rarely reaches the limit at all.

The survey is deliberately plain. The scale is fixed at one to five stars, and so are the one-hour delay and the 30-day validity. Two different scales in the same database would mean the report averages things that cannot be compared.

The survey distinguishes two groups of recipients. Employees of your company receive it without being asked first. Everyone else counts as external. External people only receive the survey after they have given consent.

Employees can unsubscribe at any time. External people can withdraw their consent at any time. The card “Consent and unsubscribe” explains how.

You decide who counts as an employee in the field “Own e-mail domains”. It is under “Settings → General” in the card “Customer Satisfaction Score (CSAT)”. Only administrators can see and change the field. Enter the domains of your company there, separated by commas. Then click “Save domains”.

An address counts as an employee if its domain is in the list. An entry also covers its subdomains. So “example.com” also applies to “it.example.com”, but not to “myexample.com”.

When saving, the system tidies up the list. Capital letters become small letters, an “@” at the start is dropped, and duplicate entries disappear. If an entry is not a valid domain, the message “These entries are not valid domains:” appears with this entry. Nothing is saved in that case.

If the field is empty, every address counts as external. That is the factory setting.

For a single account you can override the domain. Open the account under “User management” with “Edit”. In the field “Satisfaction survey” you choose “Automatic by domain”, “Employee” or “External”. The factory setting is “Automatic by domain”. Next to it, “Result: …” shows what the domain list means for this account.

The choice on the account applies to the e-mail address of that account. It helps, for example, with an employee whose address has a different domain.

After the update to 1.11.0, the field contains the allowed domains of your single sign-on providers. If you do not use single sign-on, the field is empty.

As long as the survey is switched on and nobody has saved the field, the card shows a box: “Since version 1.11.0, external people receive the survey only after giving consent. Enter your company's e-mail domains so that your employees keep receiving the survey.” Only administrators see the box.

Administrators also see the same sentence as a banner at the top of every page. As long as the box is shown, the banner appears as soon as at least one survey was not sent in the last 30 days. “Open the survey settings” leads to the card. “Hide for 30 days” hides it for your account for 30 days.

There is one thing to do: enter the domains and click “Save domains”. This also applies if the list is already correct or is meant to stay empty. After that, the box and the banner disappear.

Below the field, a line counts the surveys that were not sent in the last 30 days, for example “Surveys not sent in the last 30 days: 2 without consent, 1 after unsubscribing.” Each ticket counts once. Other reasons, such as the blocklist, are not counted. The link “Show consents and unsubscribes” next to it opens the list from the card “Consent and unsubscribe”. Only administrators see the line and the link.

The rule only applies to the satisfaction survey. All other e-mails go out as before, for example the replies of agents.

An external person only receives the survey after giving consent. There are two ways for that. The person agrees in the customer portal, or an agent records the consent on the ticket.

In the customer portal, a checkbox with the consent text appears when a ticket is created: “Yes, I would like to receive a short satisfaction survey by e-mail after my tickets are closed. I can withdraw this at any time.” The box is not ticked in advance. It only appears for customers whose address counts as external and who have not agreed yet. The consent only applies to the customer’s own address.

The profile has the switch “Satisfaction surveys by e-mail”. For an external person, switching it on is the consent. The consent text stands below it. Switching it off is an unsubscribe. For employees the switch is on as long as they have not unsubscribed. Below it, the profile shows the state, for example “Consent given on …” or “Unsubscribed on …”.

On the ticket, agents and administrators see the line “Satisfaction survey” below the requester’s details. It shows “Employee”, “No consent”, “Consent given on …” or “Unsubscribed on …”. If the ticket has no address for the survey, the line is missing. If the survey is switched off, the line says so. Then only “Record unsubscribe” is offered. The unsubscribe applies as soon as the survey is switched on again.

“Record consent” opens a dialog with the consent text. Under “How did the person give their consent?” you choose “By phone”, “In person” or “In writing”. Then you tick “The person gave me their consent.” Only then can the consent be recorded. There is no free-text field. The button only appears for external people without a valid consent.

“Record unsubscribe” needs no tick. After it, the person receives no more surveys.

Only people who may edit the ticket see both buttons. Both also work for requesters without an account. Each entry writes a line into the ticket history, for example “Consent to the satisfaction survey recorded (by phone).” The line is internal. The customer does not see it.

What counts is the state at the moment the survey would be sent. A consent recorded up to 14 days after closing still leads to a survey. An unsubscribe takes effect at once, also for surveys that are already due.

Every survey e-mail ends with the sentence “You can stop receiving these surveys at any time:” and a link. It leads to a page with the question “Stop sending satisfaction surveys to …?” Only the first letter and the domain of the address are shown there. The unsubscribe only happens with the button “Unsubscribe”. Opening the link alone changes nothing.

After a rating, the survey page also shows the link “Stop receiving surveys”. E-mail programs that support it also show their own unsubscribe button on the survey e-mail. One click on it unsubscribes directly.

After unsubscribing, the page offers an employee “Subscribe again”. The same applies if they open the link again later. External people do not see the button. They only come back through a new consent.

Unsubscribing needs no account and no sign-in. It also works when the survey is switched off or the license has expired. It only stops the survey. All other e-mails still arrive.

If the link is no longer valid, the page says so: “This link is no longer valid. Reply to one of our e-mails if you no longer want to receive surveys.” If someone writes to you for that reason, you record the unsubscribe yourself.

All entries are on the page “Survey consents and unsubscribes”. You open it with the link “Show consents and unsubscribes” in the card “Customer Satisfaction Score (CSAT)”. Only administrators may open it. The page shows one row per address with the state, the channel and the date. Under “State” you filter by “Consented”, “Unsubscribed” or “Subscribed again”. The field “E-mail address” searches for an address.

“History” shows every entry of an address. That includes the date, the state and the channel. For a consent recorded by an agent, the way follows in brackets, for example “By phone”. “Done by” names the account that acted. “Ticket” links to the ticket where it happened. “Text version” and “Language” say which consent text the person saw and in which language. This history is your proof.

“Record unsubscribe” at the top of the page records an unsubscribe for any address. Every row that is not unsubscribed has the same button. Use it when someone asks you directly to stop the survey.

You cannot record a consent on this page. It has to come from the person or from an agent who spoke with them. Nothing can be deleted here, because the history has to stay as proof. The unsubscribe dialog says so: “This address will no longer receive satisfaction surveys. The entry stays as proof and cannot be removed.”

After an anonymization a row “anonymized” can appear. It stands for an unsubscribe whose address was removed. It has no history. The card “Anonymize people” explains it.

Under “Reports” satisfaction has a section of its own, “Customer satisfaction (CSAT)”. It appears in the same report as everything else, but it only follows two of the filters: the period and the team. Without a date it covers the last 30 days.

Five tiles sit at the top. “Average score” is the average of the stars, “Satisfaction rate (4-5 stars)” tells you what share was satisfied, “Response rate” is how many answered, and “Surveys sent” counts the surveys that went out. Below both rates you find, in small print, the fraction they were built from.

“Closed without survey” is the fifth tile. It counts the closed tickets that were never asked, with the total number of closed tickets underneath. Without that number you would take a rate as the picture of your customers, and it would rest on a subset you cannot see.

The number that matters most is not the average, it is the response rate. A good score built on few answers says little about your customers.

Below that comes the distribution: for every star count from five down to one a bar shows how often it was given, with the number next to it. Then comes “Trend”, one row for each day somebody answered, with the date, that day’s average as a bar and the number of answers. Last come “By agent” with one row per agent and “Latest comments” with what people actually wrote. A “By team” breakdown joins them as soon as more than one team has rated tickets.

You can switch the per-agent breakdown off. “Per-agent evaluation” is off as a factory setting, because ratings per person are performance data — in many companies the works council has a say in that, and with cloud providers this evaluation often cannot be switched off at all.

The switch takes effect on the server and not only on screen: with it off, the server does not deliver the breakdown at all.

The single rating on a ticket is untouched by this and stays visible to the team. The switch governs evaluation across people, not what is shown on one case. There is no export of the CSAT section of its own. Agents only see teams here whose “Agent Permissions” allow seeing reports.

The “Satisfaction” filter narrows the report down to ratings. “Rated only” shows rated tickets, “Not rated” shows the unrated ones, and with “Score from” and “Score to” you can look at every ticket with one or two stars. The filter applies to the table and to both exports.

A rating can trigger a rule. In the rule editor under “Settings → Automation” there is a condition for it, “Satisfaction rating (CSAT)”, and next to it you pick “is at most”, “is at least”, “is” or “is not”. The third field holds the stars, from one to five, with the number beside them.

The usual case is “is at most 2”. Above the rule you then read the sentence the editor writes along: “When a ticket was rated 2 stars or fewer, then send an e-mail to the assignee.”

This rule needs no time condition. A new rule comes with the condition “Time without any activity” “longer than” 3 “days” in the “WHEN” block; remove it so that the block is empty. That makes it the exception among the rules: all the others wait for nothing to have happened for a while, this one waits for an event.

As an action you have everything a rule can do anyway: send a mail, raise the priority, hand the ticket to another team, or set a follow-up.

One thing works differently here. Rules normally leave closed tickets alone, but a rating almost always arrives on a closed ticket — so a rule with this condition reaches closed tickets as well. Every other rule still does not.

The condition never applies to a ticket without a rating, and that includes “is not” — otherwise “not five stars” would hit your entire unrated backlog. If you want to know how many did not answer, that is the response rate in the report.

The rule acts once per rating. Below it, “Log” opens the table “What this rule did” with one row per ticket, so you can see when it ran and what it did.