Staying and leaving
Exit survey
Send this during someone's last two weeks, once the handover is under way and they have stopped worrying about the reference. Eight questions, and we'd read number 2 before the others.
Opens this survey in Gists as a draft you can edit. Sign in with Slack, Google or your work email. 30 days free. No card needed.
Before you run this
Written answers can carry a report of harassment, a safety problem or something worse. If one reaches you, a duty to act may override what people expected about confidentiality, depending on where you are and on your own policy, so agree with your legal adviser how you will handle it before you send this. In Gists, written answers appear as themes by default, so a single report may never reach you. Set where leavers raise anything serious in Reporting and support (Surveys › Planning), which shows with every survey, and don't let an exit survey be the only route.
The decision it helps with
Which reason for leaving do you try to fix first, and is it something you can fix at all?
Running it
Send it in the last two weeks, once the handover has started and any reference has been agreed. People who are worried about a reference soften what they write, and no wording removes that entirely, so read this alongside the conversations you have anyway. Hold the results until you have at least five answers.
What the results might look like
These are made-up examples to show the patterns, not real data.
Most who raised concerns say nothing changed
of 12 leavers over a year, 7 answered “Yes, but nothing changed” on question 2.
- What it means
- People say raising it internally didn't help. Check how concerns get handled.
- What it doesn't mean
- It doesn't tell you the concerns were reasonable, or that they would have stayed had someone listened.
- What to do
- Trace three or four recent concerns to see where they stopped, and name who now answers them.
“Relationship with my manager” piles up in one area
5 of 9 leavers from one department tick it, against 1 of 20 elsewhere.
- What it means
- There is something to look into in that department.
- What it doesn't mean
- It isn't proof about a particular manager. Five people can share a department and have different managers, and leavers from a difficult team talk to each other.
- What to do
- Have someone outside the line read the department's stay survey and talk to its current staff before anything is decided.
Low scores on role match, high on the people
average 2.4 on question 3 and “The people I worked with” picked by most.
- What it means
- People liked it here and the job wasn't what they were sold.
- What it doesn't mean
- It doesn't say the job is bad. Plenty of people stay in jobs that differ from the advert.
- What to do
- Give the three lowest-scoring roles to whoever writes the adverts and runs interviews, and compare with day-30 answers.
What to tell people afterwards
Once a year, tell the company the top reasons people gave and the one thing you are changing. Keep departments out of it unless the groups are large.
What to measure next, and when
Read the stay survey in the same quarter, and check day-30 role-match scores for the roles you changed about two hiring cycles later.
The eight questions
Which of these played a part in your decision to leave?
Pick any
- Better opportunity (career growth)
- Better opportunity (pay/benefits)
- Relationship with my manager
- Work-life balance
- Dissatisfied with the work itself
- Company culture or direction
- Personal reasons
- Something else
Did you raise your concerns with anyone before deciding to leave?
Pick one
- Yes, and it helped
- Yes, but nothing changed
- No
- I did not have concerns
How well did the role match what you expected when you joined?
1 = very different, 5 = very close
How supported did you feel by your manager?
1 = not at all supported, 5 = fully supported
What did you value about working here?
Pick any, optional
- The people I worked with
- The work itself
- Flexibility
- Learning opportunities
- Pay and benefits
- The mission
- Not sure
Overall, how would you rate your time working here?
1 = very poor, 5 = excellent
Would you recommend this organization to a friend as a place to work?
Pick one
- No
- Probably not
- Unsure
- Probably
- Yes
Is there anything you would like leadership to hear?
Written answer, optional
We wrote these questions for this survey and they are not a validated scale. You can edit them once the survey opens in Gists. If you copy them into another tool, check what it records about who answered.
Why the reason field is the weak one
Question 1 lets people tick every reason that played a part, because records of a single reason have a poor track record. Campion gathered accounts from 325 former employees, their supervisors and their replacements, then compared them with a university's personnel files. The files came off badly, especially because they recorded a single reason for each departure (Campion, 1991).
Lee and Mitchell's model of quitting has four separate decision paths, each involving different events and thought processes (Lee & Mitchell, 1994). A person who got an unsolicited offer and a person who has been miserable for a year can both tick “Better opportunity (career growth)”.
Question 2 asks whether they raised it with anyone first. When people pick “Yes, but nothing changed”, somebody tried to sort the problem out from inside, and we would take ten of those in a year more seriously than whatever the reason chart shows.
Next to the stay survey
Pair this with the stay survey, which goes to people who are still with you, and compare what each group says about managers and growth.
Who sees the answers
Results show only when 5 or more people answer. A team below 5 rolls up into the team above it. A group is also held back when comparing it with another group could reveal a small team. HR and named leaders see every team with 5 or more answers. Each manager sees only their own team. Employees see no team scores. Themes only by default. Word-for-word answers are a company setting, off unless turned on, and employees see a notice before they write.
Related
Sources
- Campion, M. A. (1991). Meaning and measurement of turnover: Comparison of alternative measures and recommendations for research. Journal of Applied Psychology, 76, 199-212. doi:10.1037/0021-9010.76.2.199
- Lee, T. W., Mitchell, T. R. (1994). An Alternative Approach: The Unfolding Model of Voluntary Employee Turnover. The Academy of Management Review, 19, 51. doi:10.5465/amr.1994.9410122008