Stay Interviews: A Retention Playbook for Indian SMBs

By the time an employee resigns, you have already lost. The exit interview tells you what went wrong, months after it could have been fixed, and usually in a softened version because the person is being polite on their way out. A stay interview does the opposite: it asks the question while the answer is still actionable. If your retention numbers are worse than you would like and your exit interviews keep surfacing the same three themes, a structured stay interview program is the highest-return people process you can build this quarter.

This is a practical playbook for running stay interviews in Indian SMBs and mid-sized companies — what they are, how they differ from every other conversation you already have, exactly what to ask, how to train managers to run them without turning them into appraisals, and how to close the loop so employees see that talking actually changed something. Written for HR managers and founders running teams of 20 to 500 where every regretted exit is genuinely felt.

What a Stay Interview Is — and What It Is Not

A stay interview is a short, structured, forward-looking conversation with a current employee about what keeps them here, what frustrates them, and what would make them consider leaving. It is conducted with people you want to keep, before there is any sign of trouble.

That definition contains four load-bearing words.

Structured. It follows a consistent question set so that answers can be compared across people and across time. An unstructured "how are things going?" is a chat, not a stay interview.

Forward-looking. The subject is what happens next, not what happened last quarter. This is what separates it from a performance review.

Current employee. Not someone who has resigned, not someone on a performance plan, not a candidate.

Before there is trouble. Running stay interviews only with people you suspect are leaving turns the process into a retention panic signal. Employees notice immediately, and it stops working.

Here is how a stay interview differs from things you may already be doing:

The last column matters most. An engagement survey gives you a company-level theme. A stay interview gives you a person-level commitment. Both are useful; only one of them changes what happens to a specific individual on a specific team next month.

Why Stay Interviews Work

Three mechanisms, all of them fairly ordinary.

They surface problems while the cost of fixing them is low. Most resignations are not caused by a single event. They are caused by a slow accumulation — a project that never materialised, a promise about a role change that quietly disappeared, a manager relationship that cooled, a commute that got worse. Each of these is fixable in the first month and almost unfixable in the sixth, because by then the employee has already emotionally left and probably has an offer.

They make the employee's decision criteria explicit — to both parties. People often cannot articulate why they are unsettled until someone asks a good question. The interview does not just extract information; it helps the employee organise their own thinking, which frequently ends in "actually, what I want is X" — and X is often something you can provide.

They shift the relationship from transactional to negotiated. An employee who has been asked what they want, and has seen one of those things delivered, treats the next recruiter call differently. Not because of loyalty in the abstract, but because they have evidence that raising an issue here produces movement, and they have no such evidence about the new place.

There is a fourth, less discussed benefit: stay interviews make managers better. A manager who has to sit through eight of these a year and then own the follow-up learns more about people management than any workshop delivers.

The Cost of Getting Retention Wrong

Before building the program, it helps to have a number, because "retention matters" is not a budget argument.

Estimate the replacement cost of one regretted exit at a mid-level role:

• Recruitment cost. Agency fee or internal recruiter time, job board spend, referral bonus.
• Manager and interviewer time. Screening, interviewing, debriefs, offer discussions — often 20 to 40 hours of expensive time across the loop.
• Vacancy productivity gap. The role sits empty for the notice period plus the search period. Work either does not happen or lands on teammates.
• Ramp time. A new joiner is typically at partial productivity for the first quarter, sometimes longer for specialised roles.
• Knowledge loss. Undocumented context, client relationships, institutional memory. Hard to quantify, frequently the largest item.
• Contagion. Exits cluster. One departure in a small team measurably raises the probability of the next.

You do not need precision. A defensible estimate expressed as a multiple of monthly cost — many organisations land somewhere between three and nine months of the role's cost, depending on seniority and scarcity — is enough to justify a process that costs a few hours per employee per year.

Run this calculation for your own last four regretted exits. Present the total to your leadership team. That is usually the end of the debate about whether stay interviews are worth the time.

Designing the Program

Who gets a stay interview

Three defensible approaches:

Everyone. Cleanest signal, no stigma, most work. Feasible up to perhaps 150 employees if managers own the conversations.

Risk-weighted. Everyone gets one annually; higher-priority segments get two. Priority segments typically include: employees in their first year, employees in roles that are hard to refill, employees who have recently changed manager or role, high performers, and anyone in a team that has had a recent exit.

Segment-based pilot. Start with one function or one manager's team, prove it works, expand. This is usually the right answer for a first attempt.

Whatever you choose, be explicit and consistent. The one design that fails is "we do them for people we think might leave", because employees decode that instantly.

Timing and cadence

• New joiners: a light-touch version at 30 days and 90 days. Early-tenure attrition is disproportionately preventable and disproportionately about onboarding, clarity, and manager fit.
• Established employees: once or twice a year, deliberately spaced away from appraisal and increment cycles. If a stay interview happens in the same fortnight as a rating conversation, employees will read it as an appraisal, and they will answer strategically rather than honestly.
• Trigger-based: after a manager change, a reorganisation, a missed promotion, or a teammate's departure. These are the moments when the ground shifts under someone.

Avoid running them right before increments — the answers become salary negotiations. Avoid running them in the busiest operational month. A quiet period, roughly a quarter away from the appraisal cycle, works best.

Who conducts them

The default should be the direct manager, for a simple reason: most of what a stay interview surfaces can only be acted on by the direct manager. Work allocation, growth opportunities, recognition, workload, and team dynamics are all manager-controlled.

Exceptions where a skip-level or HR should conduct it:

• The employee's concern is the manager. Provide a stated alternative route so this is possible without drama.
• The manager is new to the team and has not yet built enough trust.
• The manager has demonstrated they cannot hold a non-defensive conversation. Fix this, but do not sacrifice the employee's experience while you do.

Some organisations run a hybrid: manager conducts, HR conducts a lighter version for a sample to validate that the manager version is producing honest answers. This is a good quality control if you have the capacity.

Confidentiality: be precise, not reassuring

Do not say "this is completely confidential" if it is not. Say exactly what happens:

• The notes go to HR and to your manager.
• Themes are aggregated and shared with leadership without names.
• Anything you specifically ask to be kept between us stays between us, with the standard exception of matters involving safety, harassment, or legal obligation.

Precision builds more trust than a comfortable overstatement, and an overstatement that turns out to be false destroys the program permanently.

The Question Set

A good stay interview runs 30 to 45 minutes and uses 8 to 12 questions, with the employee talking roughly 80% of the time. Here is a tested structure. Adapt the wording; keep the shape.

Opening — frame it (2 minutes)

Say something close to this:

"This is a stay interview. It isn't a review, it doesn't affect your rating or your increment, and there's no bad answer. I want to understand what's working for you here, what isn't, and what would make this a place you'd want to be in two years. I'll take notes, and at the end we'll agree on a couple of things I'll actually do."

The frame matters enormously. Without it, the employee spends the first ten minutes trying to work out what the conversation is really about.

Block 1 — What is working

1. What do you look forward to when you come to work? Softer than "what do you like", and it produces concrete answers.
2. Think about your best week in the last six months. What made it good? Specific memory beats abstract preference. You learn what conditions produce their best work.
3. What are you learning here that you value? Growth is the most common driver of both staying and leaving. This establishes the baseline.

Block 2 — What is not working

1. What do you dread, or what drains you? The mirror of question one. People answer this more honestly than "what don't you like about your job".
2. What's the most frustrating part of getting your work done? Aims at process and system friction rather than people, which lowers defensiveness and often surfaces genuinely fixable operational problems.
3. If you could change one thing about how this team works, what would it be? Bounded to one, which forces prioritisation.

Block 3 — The forward-looking core

1. What would make you seriously consider an offer from somewhere else? The single most important question in the set. Ask it directly. Most people answer it honestly if the frame has been set properly, and the answer is your actual risk register.
2. What would you like your role to look like in 18 months? Not "where do you see yourself in five years", which invites a rehearsed answer. Eighteen months is concrete enough to be honest and long enough to be aspirational.
3. What's one thing we could give you that would make the biggest difference to your work? Deliberately open. The answers are rarely money. Common responses: clearer priorities, a specific project, fewer meetings, a mentor, a piece of software, working from home on two fixed days, someone junior to delegate to.

Block 4 — About the relationship

1. What could I do differently as your manager? Requires a manager who can hear the answer without arguing. If they cannot, train them before they ask, or do not ask.
2. When you've done good work recently, did it get noticed? How would you want to be recognised? Recognition preferences differ enormously and are cheap to accommodate once known.

Block 5 — Close

1. Is there anything I should have asked you and didn't? Catches whatever the question set missed. Frequently the most valuable minute of the conversation.

Then close by summarising back what you heard, naming one or two things you will do, and stating when you will follow up. Do not promise more than two. Two delivered beats five discussed.

Questions to avoid

• "Are you happy here?" Yes/no, socially loaded, useless.
• "Are you thinking of leaving?" Puts the employee in a corner. Nobody says yes, and asking it damages trust.
• "How would you rate your engagement?" You are not running a survey. You are having a conversation.
• "Is the salary okay?" Turns the whole session into a compensation discussion. Let compensation come up naturally if it is a real driver; do not summon it.

Running the Conversation Well

Prepare, but lightly

Fifteen minutes before: review the employee's tenure, role history, recent work, last stay interview notes and commitments, and any known changes to their situation. If a commitment from last time was not delivered, open with that. Acknowledging a broken promise before the employee raises it is worth more than any question in the set.

Location and mode

Neutral, private, uninterrupted. Not the manager's office if the manager's office is where bad news gets delivered. Video is acceptable for distributed teams and often works fine; phone is not, because you lose too much. Some managers find a walking conversation produces the most honest answers, particularly with employees who find direct eye contact in a meeting room formal.

The three behaviours that determine success

Silence. After you ask a question, stop talking. The second answer — the one after the pause — is almost always the real one. Managers who fill silence get surface answers.

No defending. When an employee says something critical about a decision, a process, or you, the only acceptable responses are a follow-up question or an acknowledgement. Explaining why they are wrong ends the conversation, and word spreads that stay interviews are where you get lectured.

Follow the energy. If someone gets animated about something not on your list, abandon the list. The question set is scaffolding, not a script.

Note-taking

Take notes visibly and tell them you are doing it. Capture the employee's own words for key items, not your paraphrase — "I don't feel like anyone sees the work" is more useful six months later than "recognition concern". Send the summary to the employee within 48 hours and ask them to correct anything you got wrong. This single step converts the conversation from something that happened to something on record.

Turning Answers Into Action

This is where most programs die. The conversations happen, everyone feels good, nothing changes, and the second round has 40% attendance and 10% honesty.

Sort every item into one of four buckets

Fix now. Within the manager's control, low cost, quick. A recurring meeting removed. A tool licence approved. A clarification about who owns a decision. Target completion inside two weeks. These build the credibility that carries the whole program.

Fix with a plan. Requires time, budget, or coordination. A role change, a project reassignment, a training program, a hiring request to reduce someone's load. Give it a named owner, a date, and a check-in. Tell the employee the plan, including the realistic timeline.

Cannot fix — explain honestly. Some things genuinely cannot change. The office location is not moving. The client demands weekend availability. The senior role the person wants does not exist and will not for two years. Say so plainly, explain the constraint, and then have the harder conversation: given this, what else could we do that would matter to you? Employees leave over unmet expectations far more often than over unmet requests. Honest "no" preserves trust; vague "let me see" destroys it.

Escalate. Anything involving harassment, safety, discrimination, or serious misconduct leaves the stay interview process immediately and enters the proper channel. Say this out loud at the start so nobody is surprised.

Aggregate ruthlessly

Individual actions are the point, but the aggregate is where you find the systemic problems. After each cycle, code every item into themes and count them. When eleven of forty people mention that priorities change too often, you do not have eleven conversations to have — you have a planning problem to fix.

Track the aggregate over cycles:

• Top five themes this cycle versus last cycle
• Percentage of commitments delivered on time
• Themes by team, by tenure band, by function
• Items classified as "cannot fix" — if this bucket grows, something structural needs leadership attention

Report back visibly

Publish, to the whole company or the participating group: here were the top themes, here is what we are doing about each, here is what we decided not to do and why. Include the "not doing" list. Its presence is what makes the rest credible.

A 90-Day Implementation Plan

Days 1–15: Build the case and the design

• Calculate replacement cost for your recent regretted exits and get leadership sign-off on the time investment.
• Decide scope: who is in the first cycle, who conducts, what cadence.
• Finalise the question set and the confidentiality statement.
• Build the note template and the tracking sheet.

Days 16–30: Train the interviewers

• A two-hour session, mostly practice. Cover framing, silence, non-defensiveness, the four buckets, and note-taking.
• Run role-plays with hard scenarios: the employee who says nothing, the employee who criticises the manager to their face, the employee who asks for a promotion you cannot give, the employee who says they are already interviewing.
• Give managers a one-page cheat sheet with the questions and the closing script.

Days 31–60: Run the first cycle

• Communicate to employees a week before: what this is, why, that it does not affect ratings, and what will happen with the notes.
• Schedule with adequate spacing — no manager should do more than two in a day.
• HR checks in mid-cycle: are conversations happening, are notes being captured, is anything urgent surfacing.

Days 61–90: Act and report

• Deliver every "fix now" item. All of them. This is non-negotiable in cycle one.
• Assign owners and dates for "fix with a plan" items.
• Aggregate themes and present to leadership.
• Publish the summary and the action list to employees.
• Collect feedback on the process itself and adjust the question set.

Then set the next cycle date in the calendar before the current one closes.

Measuring Whether It Is Working

Retention is a lagging indicator with a long delay and many confounders. Do not judge the program on quarterly attrition alone. Track a mix:

Leading indicators

• Completion rate — percentage of planned interviews actually conducted
• Commitment delivery rate — percentage of "fix now" items closed within the promised window
• Depth score — a simple manager self-rating on whether the employee opened up; low scores flag training needs
• Employee-reported usefulness — one question after the conversation

Lagging indicators

• Regretted attrition rate, tracked separately from total attrition
• First-year attrition rate
• Attrition among employees who received a stay interview versus those who did not, in the first year of the program
• Exit interview themes — are they still the same themes, or are the fixable ones disappearing?
• Internal mobility rate, since a well-run program often converts an exit into a role change

The one comparison that convinces sceptics: take the themes from your last twelve exit interviews and the themes from your first stay interview cycle. If the overlap is high, you now have advance warning on problems you previously only learned about after the fact. That is the entire value proposition, demonstrated with your own data.

Common Failure Modes

Managers treat it as an appraisal. Symptom: the manager talks more than the employee, feedback on performance appears, ratings get referenced. Fix: training, plus deliberate calendar separation from the appraisal cycle.

No follow-through. Symptom: cycle two has poor attendance and thin answers. Fix: make commitment delivery a manager accountability with a visible tracker. One undelivered promise costs more trust than the conversation created.

Only run for flight risks. Symptom: employees ask "why me?". Fix: universal or clearly-stated-segment coverage.

Turned into a salary conversation. Symptom: every interview ends with a compensation request. Fix: framing at the start, plus a separate, credible compensation review process. If pay is genuinely uncompetitive, stay interviews will not save you — fix the pay.

Notes go into a drawer. Symptom: nobody can tell you the top three themes from last cycle. Fix: structured capture in a system, mandatory theme coding, aggregate review on the leadership agenda.

Over-promising. Symptom: a manager agrees to five things to keep the conversation pleasant, delivers one. Fix: cap commitments at two per conversation and train managers that "I don't think I can do that, here's why" is a strong answer, not a failure.

Anonymous or written versions. Symptom: someone suggests turning it into a form to save time. This is a survey, not a stay interview, and it loses the mechanism that makes it work — the negotiated, personal commitment. Run both if you like, but do not substitute.

Sample Note Template

Keep it short enough that managers actually complete it.

• Employee, role, tenure, date, interviewer
• What's working (their words):
• What's not working (their words):
• What would make them consider leaving:
• What they want their role to look like in 18 months:
• The one thing that would make the biggest difference:
• Commitments made — item, owner, date:
• Theme codes: (growth / manager / workload / recognition / compensation / clarity / tools / flexibility / team / role fit / other)
• Risk assessment: low / medium / high, with reason
• Follow-up date:

The theme codes are what make aggregation possible. Fix the list in advance and do not let managers invent new codes mid-cycle, or you will lose comparability.

Adapting for the Indian SMB Context

A few adjustments that matter in practice:

Hierarchy and directness. In many Indian workplaces, employees are reluctant to criticise a manager to their face, particularly across an age gap. Skip-level stay interviews, or an HR-conducted version for at least a sample, compensate for this. So does explicitly naming the norm: "I'd genuinely rather hear something uncomfortable now than read it in an exit interview."

Family and location factors. Commute, relocation, family responsibilities, and spouse employment are legitimate, frequently decisive drivers of exit decisions in Indian metros, and they are often left unsaid because they feel personal. A gentle, optional question — "is there anything outside work that affects how sustainable this role feels?" — surfaces issues that flexible arrangements can often solve cheaply.

Growth velocity expectations. Early-career employees in fast-growing Indian firms often benchmark their progression against peers who changed jobs. If someone's expectation of a title change in 18 months is unrealistic in your structure, say so, and then discuss what is realistic. Silence here reliably produces a resignation.

Counter-offer culture. Stay interviews reduce the number of situations where retention depends on a counter-offer, which is the worst possible mechanism — expensive, unfair to those who did not threaten to leave, and typically only delaying the exit by months. Make that argument explicitly to your leadership when justifying the program.

Frequently Asked Questions

How is a stay interview different from a one-on-one?

A one-on-one is about work in progress — this week's blockers, this project's decisions. A stay interview is about the employment relationship itself: what keeps this person here and what might push them out. Different subject, different cadence, and different output. Running stay interview questions inside a regular one-on-one usually means they get squeezed out by operational topics.

How long should a stay interview take?

30 to 45 minutes for an established employee, 20 to 30 for a new-joiner check-in. Longer than an hour usually means it has drifted into a different conversation. Schedule buffer afterwards so it can run over naturally without the manager watching the clock.

Should HR or the manager conduct them?

Manager by default, since most fixes are within the manager's control. Provide a skip-level or HR alternative for employees whose concern is the manager, and consider an HR-conducted sample to check that the manager-led conversations are producing honest answers.

What if an employee says nothing useful?

Usually a framing or trust problem rather than a person problem. Try more specific questions anchored in memory — "tell me about the best week you've had here" rather than "what do you like". If it is still thin, close gracefully, do something small and visible for them anyway, and try again in six months with a different interviewer.

What if someone asks for a promotion or a raise we can't give?

Answer honestly and immediately: here is the constraint, here is what would need to be true, here is a realistic timeline. Then pivot to what else could matter — scope, project, learning, flexibility, title within the current band. Employees accept "no" far better than they accept "maybe" that quietly becomes "no" six months later.

How often should we run them?

Once or twice a year for established employees, plus 30-day and 90-day check-ins for new joiners and trigger-based conversations after significant changes. More than twice a year for the same person tends to produce repetition without new information.

Won't asking "what would make you leave?" plant the idea?

No. Employees who are thinking about leaving are already thinking about it; the question surfaces it while you can respond. Employees who are not thinking about leaving are not prompted into it by being asked what they value. What does damage retention is a workplace where nobody ever asks.

How do we keep stay interview notes appropriately private?

Store them in your HR system with role-based access rather than in email or personal drives, be explicit with the employee about who can see them, retain them only as long as they are useful, and keep them out of performance and increment decisions. Handling personal data carefully is both a trust requirement and, increasingly, a legal one.

Do stay interviews actually reduce attrition?

They reduce preventable attrition — exits driven by fixable frustrations, unclear growth paths, manager friction, and workload. They will not retain someone leaving for a genuinely better opportunity, a relocation, or a career change, and they cannot compensate for uncompetitive pay or a broken culture. Judge the program on regretted attrition and on whether your exit interview themes are changing, not on total attrition.

The Short Version

Stay interviews work because they move the conversation forward in time. Everything an exit interview tells you, a stay interview could have told you six months earlier, when it was still a problem instead of a resignation.

The program itself is not complicated: a consistent question set, a manager who can listen without defending, notes that get captured, two commitments that actually get delivered, and an aggregate view that reaches leadership. What makes it hard is follow-through — and what makes it worth doing is that follow-through is exactly what your employees are currently unsure you are capable of.

Start with one team and one cycle. Deliver every small commitment. Publish what you heard and what you changed. Then expand. Within two cycles you will have something more valuable than any engagement score: a live, named, per-person understanding of why people stay, and a track record that proves speaking up here is worth the effort.

If you would like stay interview scheduling, structured notes, theme coding, commitment tracking, and attrition analytics to sit alongside your onboarding, performance, and payroll data rather than in a separate spreadsheet, CozyHR is built to keep the whole employee lifecycle in one place. Have a look and see how much lighter the administrative half of this becomes.