The exit interview nobody read
She wrote four pages. HR filed it. Two months later, three more people quit for the reason she'd already named. The warning was right there.
An HR director showed me a year's worth of exit interviews from her growth-stage company. There were 31 of them, ranging from one paragraph to four pages. The patterns were visible by interview three. In her stack, about 60% named the same manager, 40% named the same career-development gap, and 30% named pay disparity with newer hires. The same complaints, repeated, all year.
I asked who had read them. She had. The COO had skimmed a few. The CEO had read none. The patterns had been filed but never aggregated. The issues that produced 31 resignations last year were producing this year's wave too. The company kept learning the same lesson and never acting on it.
This is the normal state of exit-interview programs at growth-stage companies. The interviews happen because HR was told to run them. The forms get filled out, sometimes thoughtfully. Then they're filed. The patterns inside never get aggregated, never reach leadership, never turn into action items. The exit interview becomes a ritual the company performs without committing to act on what it produces.
The structural problem is a gap. The people who read the interviews aren't the people with authority to act. HR reads them. HR usually can't remove a bad manager, restructure career paths, or fix pay disparity. The COO and CEO have that authority, but they often don't read the interviews, because reading them is HR's job. So the information stays in the file. The patterns continue. Next quarter produces more exits citing the same causes.
Bridge the gap mechanically. The head of people reads each exit interview monthly and produces a themed summary every quarter. The summary aggregates the patterns, anonymizes the individual content, and names the top three to five recurring issues. It goes to the leadership team in a quarterly review. Each theme gets a named owner, a specific intervention, and a timeline.
Named ownership is what produces action. Themes without owners are observations. Themes with owners become accountable work, not blame. The manager named in most of the interviews needs a specific response: coaching, a role change, or an exit. The career-development gap needs a specific program. The pay disparity needs an audit and a remediation. Each of these is uncomfortable. All of them cost less than losing people to the same issue, quarter after quarter.
The anonymization step matters. Individual interviews name specific people and specific grievances. The aggregated themes carry the signal without exposing the source. The team being criticized doesn't need the raw complaints. It needs the pattern and a path to fix it.
The deeper habit is to read exit interviews as data about the company's current state, not as commentary on the person leaving. That person has no remaining stake, so they give you the most honest feedback you'll ever get — the flip side of the loyalty that kept them quiet while they stayed.
Read the interviews. Aggregate the themes. Name the owners. Track the interventions.