Onboarding Survey Questions for New Hires
Fourteen questions written for somebody about a month into a new job. Take the survey below, copy any block of it, or open the whole template in the editor.
- 14questions
- About 4 minto complete
- Day 30the one send this page recommends
- New starterswho answers it
Pick one moment: why this survey goes out on day 30
Almost every page on this subject hands you a menu of send moments and picks none of them, so a reader who arrived asking what to send leaves asking which of five. Here is the choice, and then ours.
| When you send it | What this moment can tell you | What it cannot |
|---|---|---|
| Before the first day | Whether the paperwork, the equipment order and the joining instructions arrived in time. | Anything about the job or the team. Neither has been met. |
| The end of the first day | Whether the day was organised, and whether anybody was expecting them. | Whether the training was enough. No work has been done yet. |
| Day 30 | All three things a new starter can report on: what the job is, whether they can do it, whether the team has taken them in.[1] | How the job turns out. Most of the first year is ahead. |
| Day 90 | A settled account, with a probation conversation in view. | The first week, except as a memory of it. |
Somebody is a new starter for about a year, not a fortnight. The meta-analysis behind this page defined newcomers as people in a new job at a new employer for thirteen months or less, and the studies in it collect data at entry, at three months and at six months.[1] A survey sent in week one is not measuring the year. It is measuring a Tuesday. If a single day is what you want judged, the orientation survey questions do that job instead.
What a new starter can report on changes as the month goes on. On day two they can tell you whether their login worked and whether anybody was expecting them. They cannot tell you whether the training was enough, because they have not done the job yet. By thirty days all three of the measured things are answerable at once, and of the three it was social acceptance, the slowest to arrive, that was related to every outcome the meta-analysis looked at.[1]
The common failure is asking every moment in one late send. The US government instrument does exactly that: it has sections headed the first day, the first week and the first month, all inside a questionnaire that lands after ninety days.[2] An item about whether a desk was ready is a different item at ninety days than it was on day two. That is not a criticism of the wording. It is what happens to any question asked long after the thing it asks about.
Practice has settled near the same place. Cornell runs its new hire survey six times a year and invites people sixty to ninety days after they start.[3] The swap sets below turn this template into a day-1 or a day-90 send, so picking a different moment does not mean finding a different page.
If your organisation calls this an induction rather than onboarding, the induction feedback survey template is the same job in British and Australian vocabulary, and it is built around a repeating cadence rather than one send.
The onboarding survey questions, block by block
Fourteen questions in five blocks, about 4 minutes to answer. Two swap sets sit underneath them for the other two moments, marked and not counted in the fourteen, so copying the lot gives you twenty-one.
What the job is
Three statements about the job itself. A newcomer can answer all three at a month, which is the first point at which the answers mean anything.[1]
- I could describe what my job is for by the end of my first week.Tied to the first week, so a low answer points at the handover, not at the person.
- Somebody has told me what a good first three months looks like.A disagree here is the cheapest thing on the form to fix: somebody has to say it out loud.
- The job so far matches how it was described to me when I was hired.The only item about the hiring, which is why it asks about the description and not the recruiter.
Whether you can do it
Confidence in the work, asked only about work that has actually happened. Somebody three days in cannot answer any of these, which is the argument against a week-one send.
- I am confident about the parts of the job I have done so far.Bounded to work already done, so nobody rates a job they have not started.
- The training I have had prepared me for the work I am actually doing.The comparison a new starter can make and a trainer cannot.
- I know who to ask when I get stuck.Disagree is a map problem, not a training one. It is fixed with a name, not a course.
Whether the team has taken you in
The slowest of the three blocks to become answerable, and the one that was related to all five outcomes in a meta-analysis of 70 samples of new starters.[1]
- My team has made room for me.Deliberately vague about how, because teams do it differently and nobody is grading a ritual.
- I have had a conversation here that was not about a task.An event rather than a feeling, so the answer is a fact the respondent can check.
- I know what the people around me are working on.Separates somebody who has been welcomed from somebody who has been seated.
The first days, while they are still recent
Two facts rather than two judgements. Both are about a window the respondent has just left, which is the reason this page sends at a month rather than at ninety days.
- When did your accounts, logins and equipment first all work?Asks when it worked, not whether. A yes hides the four days somebody spent waiting.
- Who did you turn to most in your first two weeks?A routing fact, not a rating. The last option is the one to count.
In their own words, then one rating
Both text boxes are optional and both sit before the end, so the survey does not finish on an empty box and people still have a reason to type.
- What would have made your first week easier?Asks for a fix rather than a complaint, which also makes it forwardable.
- What did you have to work out for yourself that somebody could have simply told you?The most useful prompt on the form for whoever writes the onboarding checklist: every answer is a line missing from it.
- So far, joining here has gone the way I expected it to.Kept last so it is answered after the detail rather than instead of it.
Swap set: sending on day 1 instead
Four items that replace the first three blocks if you send on the first day. Nothing about the job survives the swap, because none of it has happened. Two of the four cover the gap between accepting the offer and starting.
- Everything I had to do before my first day arrived in time to do it.A pre-boarding item, about your paperwork, so a low answer has a named owner already.
- I knew where to be, and who to find, before my first day.The second pre-boarding item, and the one that separates a warm start from a confusing one.
- By the end of today I had met the people I will work with most.Answerable on day one, and one of very few things that is.
- I know what I am doing tomorrow.A first day compressed. If this is low, nothing else about the day mattered.
Swap set: sending at day 90 instead
Three items that replace the first-days block at ninety days. The first-days questions come out because by then the answers are being remembered rather than reported, which is the failure mode of a single late send.[2]
- The work I am doing now is the work I was hired to do.At ninety days there is enough of the job to compare against the description.
- Someone has talked with me about how my first three months went.Whether the conversation happened, kept separate from any judgement of who should have held it.
- I could explain to a new starter how decisions get made here.The one item worth waiting ninety days for: a month is not long enough to answer it.
Every line here was written for this page. The fourteen in the template are already built, so you can open them in an online survey tool and change any wording before it goes out. There is no licensed instrument for onboarding to reproduce or to avoid, which is unusual for an employee survey and worth knowing before you start rewriting from a blog post.
Two items are worded the way they are because of what they are not. The question about what the job is asks about the end of the first week rather than making a standing statement about work in general, and the equipment question asks when things started working rather than whether they do. Both are moment questions on purpose. The item about the job matching its description is the only one here about hiring, and it is asked of an employee about a job; the process that got them there is the interview feedback survey.
Your first cohort is six people, and that changes both promises you can make
This is the part every ranking page on this subject leaves out, and it is the part that decides whether the survey is worth sending.
A new-hire cohort is defined by a start date, so it is small by construction. If six people joined in March, the March cohort holds at most six answers. Nobody can promise anonymity to a group of six, and the enterprise tools do not try: Microsoft documents its lifecycle surveys as identifiable and sets the response threshold below three for them, while its confidential surveys default to hiding any score behind fewer than five responses.[4] That default of five is a useful number to borrow even though nothing on this page enforces it.
It gets smaller the moment you add a filter. Start date, team, site and role are four fields that feel like reporting and behave like a name. A study of US census data found that a five-digit postal code, a sex and a date of birth were enough to make 87 per cent of the population unique.[5] That is about a country rather than your survey, but the arithmetic underneath is the same, and it is why this template collects no demographics.
So there are two honest options and no third. Either say plainly who reads the answers and ask nothing you would not ask to somebody's face, which is what the block below does, or pool three or four cohorts and report them together. Pooling costs you the chance to act while the person who answered is still new, so this page recommends the first. Either way, choosing a sample size is a different exercise from setting a reporting threshold.
What to put at the top of the survey
Who reads your answers [NAME], [ROLE], reads these. They are not handed to your manager one answer at a time. Only [N] people started here this month, so this is confidential rather than anonymous. Please answer as though the person named above can tell which answers are yours, because on a group this size they can. If something needs dealing with today rather than counting later, email [NAME] at [EMAIL] instead. This form is not the place for it.
| What to do with the answers | What it stops |
|---|---|
| Read the written answers the day each one arrives, one person at a time. | Holding the thing you could have fixed this week until a quarterly report. |
| Report no block, and no split by team or site, on fewer than five answers. | A chart of four answers, which is four answers. |
| Wait for four cohorts in a row before you call anything a pattern. | Rebuilding a joining process because two people in March agreed. |
| Freeze the wording once it goes out. | A trend that is a record of you editing the questions. |
When do they stop being new? Somewhere in the first year, on the definition used above.[1] After that a person has a baseline and a trend, which is what the employee pulse survey questions are for. If somebody goes in month five instead, the exit interview questions and the answers they gave at day 30 are worth reading side by side: the two forms ask different things about the same few months.
Running the day 30 send
Five decisions this survey needs that a general survey does not.
Fire it from the start date, never from the calendar
The date belongs to the person, not to the month. Whoever keeps the hire list sets one reminder per starter at thirty days, so somebody who joined on the 28th is not asked on day three along with everybody else.
Put a name at the top before you put a link in an inbox
On a group this size the respondent will work out who reads it. Being told is better than deducing, and the block above is there to be pasted straight in.
Keep the recall window shorter than the survey is old
Ask about the first week while the first month is still running. Anything further back is being remembered, and a remembered answer is not worth the question.
Send the written answers where they can be acted on, as themes
One person naming a missing account is an errand. Three cohorts naming it is a broken step in the joining process, and only the second belongs in a report.
Give it a year before you compare one cohort with another
Four cohorts is the earliest point at which a gap between two of them is worth a meeting. Until then you have a set of accounts, and accounts are read rather than counted.
Questions people ask about onboarding surveys
How many onboarding survey questions should you ask?
Fourteen, for one moment. The long lists elsewhere are menus covering three or four sends at once, so counting them against one survey compares a catalogue with an instrument. Pick the moment first and the length follows.
Is there a free onboarding survey you can reuse without permission?
Yes, and almost nobody mentions it. The US Office of Personnel Management publishes a 42-item New Employee Survey, sectioned by moment from before the first day through the first ninety days, as a work of the federal government.[2] It is a retrospective PDF rather than something sendable, so read it for the wording and send something shorter.
Who else should you ask, besides the new hire?
Their manager sees the other half of the same month, and a manager form is worth running. Keep it separate: the two are answered by people with different information and different stakes, and merging them produces a number that describes neither.
What changes for a remote or hybrid new hire?
Two items, and no extra block. The equipment question matters more, because a late laptop cannot be solved by borrowing one from the next desk, and the item about a conversation that was not about a task stops being incidental.
What is the difference between an onboarding survey and an orientation survey?
The unit being judged. An orientation survey rates one induction event: how it was run, what was covered, whether the timing worked for the people sitting through it. An onboarding survey rates the first weeks of the job itself, and it is answered long after any event has finished.
Methods and sources
What this template is based on
The send moment and the three things a new starter can report on come from a meta-analysis of newcomer adjustment.[1] The comparison moments come from a published US federal instrument[2] and from one large employer that publishes its own cadence.[3] The cohort-size point comes from enterprise survey documentation[4] and from work on how few fields identify somebody.[5]
Question licensing
No licensed instrument governs onboarding questions. The academic socialization scales are journal-published, so they are named here rather than reproduced; the US federal instrument is public domain. Copy, edit and translate every line on this page, then freeze the wording: a changed question is a new trend.
Limits and disclosure
The meta-analysis is correlational, mostly self-report, and its samples were mostly White. It tells you which measures travel together, not that moving one moves another. The federal instrument is undated, US-only, and published with no response data. The employer publishing findings is one university. The identifiability study used 1990 census data and describes a population, not a survey. Five per group and four cohorts are working rules, not validated thresholds.
SuperSurvey makes the editor this template opens in. Everything here is free to use with or without it.
References
- Bauer, T. N., Bodner, T., Erdogan, B., Truxillo, D. M. and Tucker, J. S. "Newcomer adjustment during organizational socialization: A meta-analytic review of antecedents, outcomes, and methods." Journal of Applied Psychology 92, no. 3 (2007): 707-721. Read from the open-access copy at PDXScholar.
- U.S. Office of Personnel Management and the CHCOC Subcommittee for Hiring and Succession Planning. New Employee Survey (sample instrument, 42 items). opm.gov (PDF)
- Cornell University Division of Human Resources. "New Hire Survey." Working at Cornell. hr.cornell.edu
- Microsoft. "Manage Viva Glint confidentiality thresholds." Microsoft Learn. learn.microsoft.com
- Sweeney, L. Simple Demographics Often Identify People Uniquely. Carnegie Mellon University, Data Privacy Working Paper 3, Pittsburgh 2000. dataprivacylab.org (PDF)
Send it thirty days after they start
Fourteen questions, about 4 minutes, no demographics asked. Open it in the editor, change any wording, and set the reminder from the start date.
Use this template