The retention interview protocol

Exit interviews tell you why people left. By the time you read them, the decision is six months old and the cost is sunk.

Stay interviews, when run seriously, tell you why people are likely to leave six to nine months from now. The cost to run them is low. The impact, when leaders act on the signal, is the single highest-ROI retention lever we see.

But most "stay interviews" are not interviews. They are a manager asking "so what keeps you here?" over coffee. The protocol below is the structured version we use with clients and teach to HR business partners.

When to run them

Three triggers, in priority order:

1. High-value employees in year two to three of tenure. This is the window when external recruiter interest peaks and internal growth plateaus.
2. Post-major-change moments. Reorg, leadership change, acquisition, major system migration. Anything that resets the psychological contract.
3. Flight-risk signals in the data. Drop in engagement score, performance rating movement, unusual PTO patterns, LinkedIn profile refresh (if you track it via a credible source).

Avoid running them defensively, "I heard a rumour so I'm checking in", that is about you, not them, and they know it. The best retention interviews are booked proactively on a 12-month cycle.

The protocol

The conversation is 45 minutes, run by a manager who is one layer above the person's direct manager. The skip-level distance matters. It makes the conversation safer to be honest in.

The protocol has four parts.

Part 1. Frame (5 minutes)

"I wanted to book time because I want you here for the long run. This is not about performance. It is about understanding what is keeping you, what is frustrating you, and what would make staying an obvious choice. Nothing you say here goes into your HR file. I will follow up with two or three concrete commitments, or honestly tell you if I cannot."

The frame matters. Without it, the person defaults to a professional, evasive script.

Part 2. The five questions (30 minutes)

Ask these five, in this order, and leave space between answers:

1. "Walk me through the last 90 days. What have you loved? What has been frustrating?"
2. "If you were offered a role elsewhere tomorrow, what would it need to have for you to take it seriously?"
3. "What is one thing about working here that you would change tomorrow if you could?"
4. "Where do you want to be in 18 months, and what capability gap stands between now and then?"
5. "What would make the next 12 months great enough that leaving would be a hard decision?"

Do not rush. The most useful answers come in the 15 seconds of silence after the polite answer.

Part 3. The commitment (5 minutes)

"Based on what I heard, here is what I am going to do in the next 30 days: [two or three specific commitments]. Here is what I cannot change: [one or two honest limitations]. Is there anything missing?"

The commitments must be specific, time-bound, and actually within your power. A vague "I will look into it" is worse than "I honestly cannot change that."

Part 4. The follow-up (5 minutes)

Book the 30-day check-in before leaving the room. This is the single highest-signal discipline.

What to do with the output

The protocol produces three kinds of data:

Individual actions. The commitments you made. Track them. Close them.

Team-level patterns. If five people in one function name the same issue (workload, a specific leader, unclear career paths), the problem is structural. Escalate it.

Flight-risk signals. A flat "nothing is frustrating, everything is fine" from a two-year high performer is often a louder signal than a list of complaints. They have already decided.

The rules that make it work

Four non-negotiables:

• Skip-level runs it. Not the direct manager, not HR alone.
• No performance content. Mentioning performance in this conversation collapses the safety. Performance lives in a different room.
• No notes taken during, synthesis after. Real honesty shuts down when someone is typing.
• Commitments or honest limits. The conversation is worthless without one of the two.

Why it works

Three reasons.

First, it reframes retention as a manager-owned responsibility rather than an HR programmatic one. The skip-level manager leaves the conversation owning two or three specific moves. That is where retention is actually won.

Second, it produces a paper trail of commitments. Over time, that trail becomes a defensible case for organisational investment, comp adjustments, flexible work policy changes, career path redesign, because the data is specific and repeated.

Third, it catches flight risks in the window when intervention is still possible. Most retention interventions fail not because the offer was wrong, but because the conversation happened two months too late.

Where to start

Pick your top 30 people across the business (by critical role, not by level). Run the protocol in the next 60 days, with skip-level managers briefed on the protocol. Aggregate the patterns, commit to three organisational-level moves, and measure attrition in the named cohort over the following 12 months.