IT survey questions, and which twelve to send
Two surveys, not one. Twelve questions for everyone who uses IT, once or twice a year, ticket or no ticket. Four more for the end of a closed ticket. Answer the twelve below, copy any block, or open it in the editor.
- 12questions
- About 3 minto complete
- Plus 4for a closed ticket
- Two numbersout of the twelve
Two IT surveys, and which one you actually need
Two instruments, two names, and mixing them is why so many IT question lists read as one undifferentiated pile.
The industry names them transactional and relationship. An independent analyst brief on service desk feedback puts it in one sentence: implement two complementary satisfaction surveys, a transactional one capturing satisfaction with individual ticket experiences, and a relationship one capturing broader satisfaction among the entire user base.[3]
This page ships the relationship survey. Twelve questions, sent to everybody who uses IT, once or twice a year, ticket or no ticket. It says what IT is like to live with, and nearly nobody runs it, because a ticket form is easier to set up and arrives on its own.
Running only the ticket survey has a cost that is easy to miss. The same brief is blunt about the failure state: it can be worse to ask people for feedback and do nothing with it than not to ask at all.[3] A ticket form gives you ratings on individual jobs, and nothing anybody can act on at a budget meeting.
| The transactional survey | The relationship survey | |
|---|---|---|
| Who is asked | Whoever raised the ticket that just closed. | Everybody who uses IT, ticket or no ticket. |
| When | On the close of one ticket. | On a fixed cycle, once or twice a year. |
| What it answers | Did that job get done, and what did it take. | Is IT holding up its end for the people who depend on it. |
| What it cannot see | Anybody who never raised a ticket. | The detail of any single ticket. |
| On this page | The four at the end of the bank. | The twelve in the survey above. |
The 12 IT support survey questions for employees
Every question, its scale, and what an answer is telling you. Copy one block, copy all sixteen, or open the survey and rewrite the wording to match your own organization.
Who is answering, in three taps
Three cheap questions first, because they are the splits that make every rating below readable. A warehouse supervisor and a developer are not rating the same service.
- Which part of the organization do you work in?The split to apply before you average anything. Use your own department names, and merge any group small enough that a comment identifies its author.
- Which of these have you needed IT for in the last three months?What this person has leaned on IT for. The last option is the one that matters: it marks the people the ticket system never saw.
- Where do you go first when something is not working?The item that turns the results into a map of where support actually happens. Everyone who does not answer the help desk is doing work the desk never records.
What the problems cost
One question that would be absurd on a customer survey and is the point of an employee one. A customer can leave; a member of staff absorbs the delay and keeps working, so time is the currency.
- In the last month, roughly how much working time did technology problems cost you?A band, not a number, because nobody recalls this to the minute. Report the share choosing half a day or more.
What IT gives you to work with
Two statements on one 5-point agree scale about what a whole user base can judge without knowing anything about IT: whether they can get in, and whether the machine keeps up. The published instrument here splits IT into three commitments and only one is the desk.[4]
- Getting into the systems I need to use is straightforward.Access, asked once and plainly. Low scores from people who also picked access on question 2 point at joiners and movers.
- The machine I work on is quick enough for the jobs I do on it.The device, framed as speed rather than specification: a respondent knows they are waiting, not what is in the box.
How support works when you need it
Three statements on the same scale about the desk itself: the knowledge of the people who answer, the speed of a fix, and whether a reported problem can be followed without chasing.
- The people who help me know enough about the tools I use.Knowledge, asked about the tools this person uses rather than IT in general, so a low score names a gap.
- Problems I report get fixed soon enough for me to carry on.Speed, judged against getting the work done rather than a service level you set. This is the one your staff act on.
- I can see what is happening with a problem I reported without having to chase it.Visibility. A queue that is fast and silent scores badly here, and the fix is a status message, not another engineer.
The two numbers, and one open box
The only two items that produce a figure anybody outside IT will ask for. Both carry their own not-asked option rather than leaning on question 2, so nobody is stranded.
- The last time you asked IT for help, was it sorted out on that first contact?First contact resolution, asked of the customer: one of the two accepted ways to measure it, and the other one needs a monthly audit.[1]
- Thinking of the quality of support you get from IT, which of these is closest?Satisfaction, worded to match the published definition: the score is the percent choosing satisfied or very satisfied, never an average.[2]
- What is the one technology thing that most gets in your way?The only place somebody can name a problem the other eleven missed. Read these first.
The four to attach when a ticket closes
The transactional block, and a different survey with a different job. These four are NOT in the twelve above: publish them as a second copy and paste its link into the message that closes a ticket.
- Is the problem you raised in this ticket now fixed?Ask the person, not the ticket status: marked resolved and actually gone are two different facts.
- How many separate contacts did it take before this was fixed?First contact resolution at ticket level, counted rather than judged. Anything past one contact names a queue to look at.
- Getting this fixed took more of my time than I expected.Effort, set against the respondent's own expectation rather than a target you chose. The validated instrument here asks for a minimum, a desired and a perceived level on every item;[4] this is that idea for one tap.
- Is there anything about how this was handled that we should know?The box the agent who closed the ticket reads the same day. Optional, and last.
All twelve are built and running in the preview above, so an online questionnaire maker opens them in a couple of minutes. The four ticket-close items are deliberately not in that survey: make them a second copy.
Four subjects that are not in here, on purpose. Security awareness, onboarding, training needs and remote-working kit get folded into the long IT menus, and each is a survey with its own respondents and reporting. For the first, the employee cyber security survey asks about behaviour rather than opinion.
The four questions to attach when a ticket closes
The transactional block, where its link goes, and what a link can honestly do at that moment.
What this is not. It does not read your ticket system, cannot fire when a status changes to resolved, cannot attach the agent who closed the job to the answer, and does not send text messages. Triggering on ticket close, scoring by agent and drilling into a queue are what an ITSM or help desk platform sells, and ServiceNow, Freshservice, Jira Service Management, Zendesk and HappyFox all do it. This is a survey behind a link.
The method that gets the split back. A link cannot read a ticket, but it can belong to a queue. Publish the survey more than once, one copy per queue or tier, and paste each copy's link into the close message that queue already sends. The answers arrive separated, so you can compare hardware against access requests, or first line against second, without the survey ever seeing a ticket.
| Publish one copy per | Paste its link into | What you can then compare |
|---|---|---|
| Queue | The close message that queue already sends. | Hardware against access requests against software faults, on one wording. |
| Tier | The close message used at that level. | First line against second, so an escalation shows up in the answers. |
| Site or region | The close message that location sends. | Whether a low score is a desk problem or a building problem. |
Send it once. One link in the message that closes the job, no reminder, no second ask on the same ticket.
Two cases where a different template is the right one. If the people contacting your desk are paying customers rather than colleagues, the customer support survey keeps the loyalty and channel questions this page drops. If the desk is a phone line and you want the answer while the caller is on it, the IVR survey questions template carries a script written to be read aloud.
Two numbers a service desk is judged on, and how to work them out
Satisfaction and first contact resolution. Both have published definitions, both are routinely computed the wrong way, and the twelve above give one of each.
Satisfaction is a percentage of people, not an average of ratings. The published definition for a support organization is the percentage of customers either satisfied or very satisfied with the quality of support they receive,[2] which is why question 11 offers five worded levels rather than a bare one to five. Count the top two boxes, divide by everyone who answered, report that.
It is also why one desk's score cannot be read against another's. Surveys here run from one to forty questions, in multiple-choice, fill-in-the-blank and interview form, with scoring systems offering from two choices to twelve, and satisfaction ends up carrying the greatest variability of any service desk metric.[2] Keep your own wording still instead of chasing somebody else's number.
First contact resolution has two formulas and only one is worth reporting. Gross divides the contacts you resolved on the first contact by every contact that came in. Net divides by everything that came in minus the contacts nobody at first line could have resolved, and it is by far the more relevant of the two; gross produces a distorted picture and is often misinterpreted.[1] Hardware break-fix and physical moves are the standard carve-outs, because no desk resolves them remotely.[1]
Question 10 is a real measurement, not a proxy. There are two accepted ways to measure it: an agent ticks a box on the ticket, or customers are asked in a follow-up survey whether their issue was concluded on the initial contact.[1] The first needs a monthly audit of sampled tickets to stay honest; the second is the survey you are reading. If a large share of your contacts arrive by email, the emerging standard is that resolution within one business hour counts as first contact,[1] a rule to settle before you publish a number.
The people who never filed a ticket
Every instrument on this subject is addressed to somebody who already contacted support. The most useful answers here come from the people who did not.
A ticket survey can only reach a ticket. Somebody who gave up, asked the person at the next desk, found a workaround or decided it was not worth the wait has left no record in your system, and no ticket-close form can reach them. They are the people whose problems were just small enough, or just annoying enough, not to report.
That is what questions 2 and 3 are for. Question 2 ends with an option for having needed IT for nothing, and question 3 offers five destinations that are not the help desk. Filter to everybody who chose one of those five and read them on their own: the cost question, the access statement and the open box, from people the desk never met.
What to do with the result. A department that routes around the desk is not automatically asking for more staff on the queue, so check the three cheaper explanations first: a self-service answer that works, a permission that should not need a ticket, or software that fails often enough to have taught people not to bother. If it is the software, the software evaluation scorecard asks what this survey cannot. If it is the self-service layer, the chatbot feedback questions measure whether it resolves anything. If people route around one screen rather than a service, that is a usability survey template job.
What it still cannot tell you. It collects what people say about the service, not what happened on any one job, and it will not reconcile with your ticket queue. Somebody rating the desk poorly may be remembering one bad week in March, which is why you run it on the same wording every year, and alongside the four ticket questions rather than instead of them.
IT satisfaction survey FAQ
What should an IT satisfaction survey ask employees?
Six things: which part of the organization they are in, what they have needed IT for, where they go first when something breaks, what the problems cost them in working time, how the service felt on access, equipment, knowledge, speed and visibility, and one written answer.
How many questions should an IT help desk survey have?
Fewer than the lists suggest. Surveys here run from one to forty questions, with scoring systems offering from two choices to twelve, which is why one desk's score cannot be read against another's.[2] The validated academic instrument here is thirteen items and takes about twenty minutes, because it asks three ratings on every one.[4] Twelve is the version people finish; four is what a closed ticket carries.
Can this fire automatically when a ticket is closed?
No. It is a survey at its own link, pasted into a close message or embedded in a page. It does not read your ticket system, does not know when a status changed, and does not send text messages. Firing on closure, scoring by agent and drilling into a queue are what an ITSM platform sells. One copy per queue gives that split without the integration.
Should an IT survey be anonymous?
The twelve-question one, yes, and say so before people answer: it asks for no name and no email address, and the answers arrive in the account it was made in. The department question is the only thing that could narrow anybody down, so merge small departments before you report. The ticket version arrives attached to a job somebody just raised, so do not promise anonymity there.
What is a good satisfaction score for an IT help desk?
A published average exists and is worth reading with its year attached, which is how the row on our technology survey templates hub carries it. The more useful answer is that the number cannot be read alone: a desk under severe budget restrictions doing well at 75 per cent, and a generously funded one doing badly at 85, is the published illustration.[2]
This asks about the desk and the tickets it closes. When the question is the whole estate rather than one support queue, the technology survey asks what people sign into, what gets in the way, and whether the last rollout was ever adopted.
Methods and sources
What this template is based on
The two-instrument structure and the names transactional and relationship follow an independent analyst brief on service desk feedback.[3] The satisfaction item matches the published top-two-box definition, and the point that a score cannot be read without knowing what the desk is funded to do comes from the same source.[2] The first contact resolution item, the two formulas, the carve-outs and the email rule come from the companion article on that metric.[1] The ticket block's effort item sets a rating against the respondent's own expectation, the one-tap version of the minimum, desired and perceived design used by the validated higher education instrument for internal IT service quality.[4]
Question licensing
Nothing here reproduces a licensed, paid or standardised instrument. The TechQual+ core instrument is named, described and cited only: its thirteen items are not reprinted, because techqual.org no longer resolves and no operative licence can be read. SERVQUAL items are not reproduced either. Every question here was written for this page and is free to copy, edit and send anywhere.
Limits and disclosure
The benchmark figures quoted here are from 2011 and 2012 and are stated with those years attached, not as what normal looks like today. No response-rate figure is offered: the ones circulating for IT surveys are vendor self-reports with no method or sample size. The survey measures what people say about a service, not what your ticket system recorded, and the two will not reconcile.
SuperSurvey makes the editor this template opens in. Everything here works with or without it.
References
- Rumburg, J. Metric of the Month: First Contact Resolution. HDI Industry Insider, MetricNet, 2011. thinkhdi.com
- Rumburg, J. Metric of the Month: Customer Satisfaction. HDI Industry Insider, MetricNet, 2012. thinkhdi.com
- Info-Tech Research Group. Take Action on Service Desk Customer Feedback. infotech.com
- Chester, T.M. (principal investigator). A Higher Education TechQual+ Study: 2015 TechQual IT Survey Project, Washington State University Tri-Cities, 2015. tricities.wsu.edu (PDF)
Send the twelve, then attach the four
Twelve questions, about three minutes, no name asked. Open it in the editor, rewrite the department list to match yours, and pick the month.
Use this template