Short answer
A stay interview happens while change is still possible in the current employment relationship. An exit interview happens around departure and supports learning for future decisions. Use both with clear purpose, voluntary participation, honest confidentiality limits, human review, and named follow-up. Neither conversation predicts an individual's decision.
The practical difference is timing and decision
Stay and exit interviews can cover similar subjects, but they should produce different decisions.
| Dimension | Stay interview | Exit interview |
|---|---|---|
| Timing | During employment, before a resignation | After notice or around departure |
| Primary question | What helps this person do good work here, and what can we address now? | What should the organisation learn from this person's experience? |
| Useful output | An agreed work action, explanation, referral, or follow-up date | A routed individual matter or a theme for wider review |
| Typical owner | Manager, with an alternative route where needed | HR, trained interviewer, or another suitable person |
| Evidence limit | A current account does not reveal a future decision | A departing person's account does not prove the cause of departure |
| Main failure | Asking without the authority or intent to act | Collecting polite reasons without coverage analysis or follow-through |
NHS Employers' retention-discussion guidance presents stay conversations as discussions with individual employees about factors that may affect their decision to stay or leave, with questions adapted to local circumstances. Acas guidance describes an exit interview as a way to understand why someone is leaving and address outstanding matters. Together, the distinction is clear: one informs a live employment relationship, while the other learns at the point of departure.
Neither should replace a grievance, safety, discrimination, whistleblowing, or urgent wellbeing route. Explain that boundary before asking sensitive questions.
Choose the interview from the trigger
Use this decision guide:
| Trigger | Better starting point | Why |
|---|---|---|
| Employee is in role and a work change can still be discussed | Stay interview | There is time to agree an action and review it |
| Team has changed manager, schedule, location, or operating model | Stay interview or open team listening | Current employees can describe how the change works in practice |
| Employee has resigned | Exit interview | The departure decision has been made |
| Fixed-term or seasonal contract ends as planned | Exit interview with departure type labelled | The organisation can learn without calling a planned completion a resignation |
| Individual raises a formal safety, conduct, or legal concern | Designated case route | A research conversation is not the correct process |
| Leaders want to know who will resign | Neither | Conversations should not be turned into individual predictions or scores |
Do not invite only people a manager believes are likely to leave. That selection is opaque and can distort the evidence. Use transparent moments or populations, such as a role transition, onboarding checkpoint, team change, open invitation, or all eligible departures.
Ask paired questions for different decisions
The same topic needs a different framing before and after resignation.
| Topic | Stay question | Exit question |
|---|---|---|
| Work design | Which part of the work makes good performance harder now, and what could we test? | Which part of the work most shaped your experience, and what should we examine? |
| Manager support | What support or clarity would help in the next work cycle? | Where did manager support help or fall short during your time here? |
| Schedule | What current schedule friction could we reasonably address? | How did scheduling affect your experience and departure decision, if at all? |
| Development | Which skill or responsibility would you like a realistic route to use? | Which development opportunities were visible and realistic to you? |
| Voice | Is there something you cannot comfortably raise with me, and would another route help? | Was there something you did not feel able to raise while employed? |
| Close | What action may I record, who owns it, and when should we review it? | What may be retained as an agreed note, and does anything need another route? |
The stay question should not become “What would make you leave?” followed by a hidden risk label. The exit question should not become “What could we offer to make you stay?” after the employee has set a clear boundary. Listen to the answer given rather than steering it toward a preferred outcome.
Build different follow-up records
A stay interview record should focus on commitments:
- the employee's agreed summary;
- the action, explanation, or referral;
- the owner;
- what is within and outside the owner's authority;
- the follow-up date;
- what the employee agreed may be recorded.
An exit interview record should keep evidence and interpretation separate:
- departure type and relevant workforce context;
- the employee's agreed examples;
- the interviewer's interpretation, clearly labelled;
- any matter routed to another process;
- thematic codes;
- coverage limitations;
- the owner and review date for any wider action.
Do not copy sensitive stay-interview notes into a manager dashboard simply because they may relate to retention. Do not circulate raw exit accounts as entertainment for leaders. Access should follow a defined purpose and role.
A combined listening calendar
Using both does not require constant interviewing. Choose moments linked to a decision.
| Moment | Conversation | Decision it can inform |
|---|---|---|
| After onboarding, once the employee has enough experience to compare expectation with reality | Stay | Improve role preview, induction, or early support |
| After a material team or operating change | Stay | Adjust responsibilities, handovers, schedule, or manager support |
| At an employee-requested point | Stay or another appropriate route | Address a current issue without waiting for a survey cycle |
| After notice is given | Exit | Learn from the experience and resolve outstanding matters |
| At a periodic people review | Aggregate both sources | Compare current-employee themes with departure evidence |
The organisation should choose the cadence that fits the workforce and decisions. A monthly ritual is not inherently better than a well-timed conversation with credible follow-through.
At the periodic review, ask:
- Who was invited and who was missing from each source?
- Are the questions and populations comparable enough to place themes side by side?
- Which themes appear only among current employees or only among leavers?
- Which individual matters require restricted handling?
- Which hypotheses need operating, workforce, or case evidence before action?
- Which proposed action has an owner, measure, and review date?
The CIPD turnover and retention factsheet recommends examining turnover by factors such as occupation, location, and tenure rather than relying on a crude total. Apply the same care when comparing stay and exit evidence. Different participation patterns can create an apparent difference that reflects who spoke rather than what the whole workforce experiences.
Fictional worked example
Morrow Digital is a fictional US and UK software services business. A delivery unit changes how weekend support is assigned. In the next stay-interview window, several participating employees describe unclear escalation ownership. Managers agree a six-week rota test and a follow-up date.
Two months later, three employees resign. Two participating leavers also mention weekend escalation, while one describes an unrelated career move. The accounts do not establish that the rota caused the departures. They are associated with the same work condition, but the evidence is small and other explanations remain plausible.
The combined review separates the decisions:
| Evidence | Decision |
|---|---|
| Current employees describe unclear ownership | Clarify the escalation role and test the rota |
| Exit interviews repeat the same example | Keep the hypothesis open and examine coverage and operating data |
| One leaver describes a career move | Record it in the appropriate theme, without forcing it into the rota story |
| Managers completed the test but employees saw no update | Improve follow-through and report what changed |
The team reviews missed handoffs, rota changes, voluntary comments, and the next check-in. It may continue, revise, or stop the test. It does not score employees by likelihood of leaving or claim that the intervention retained anyone.
Common design mistakes
Avoid these substitutions:
- Stay interview as retention negotiation: a conversation should not pressure the employee to promise they will remain.
- Exit interview as investigation: route allegations and urgent concerns through the process designed to handle them.
- Manager-only channel: offer another interviewer when the manager is part of the concern.
- Absolute confidentiality promise: explain recording, access, aggregation, and escalation limits accurately.
- Theme as cause: an observed pattern is a question to test, not a causal conclusion.
- Conversation as score: group evidence should not become an individual employment verdict.
- Listening without response: report what was accepted, declined, changed, and scheduled for review.
Use the stay interview guide and question bank to prepare a current-employee conversation. Use the complete exit interview guide to build the departure process, coding method, and action log.
When a defined population needs a guided conversation and consistent follow-up, review the Lontra product approach against the purpose and access rules above. For exit use, verify leaver access after account closure and the connection to the organisation's existing offboarding process. For stay use, decide who owns each agreed action and review date.
The choice between a stay and exit interview should follow the decision in front of you. If the employee is still in role, make the conversation useful now. If the employee is leaving, learn carefully from the experience. In both cases, preserve choice, avoid overclaiming, and close with a named next step.
