LinkedIn Post Ideas for HR Directors

Our exit interviews used to produce one of three sanitized answers. "It was time for a new challenge." "I got an opportunity I couldn't pass up." "Personal reasons." Then we changed one question. We now ask: "What did your manager not know that they should have?" What we get back: — A two-month performance issue with their peer the manager never saw. — A pattern of being skipped for stretch assignments that ended their interest in promotion. — A specific 1:1 conversation six months prior they wished had gone differently. None of this would have surfaced under "what could we have done better?" The reframing matters. What changed in our retention work: — We started running a similar conversation at month 18 with high performers. Not exit-style. Just "what does your manager not know that you wish they did?" — We caught three exits before they happened. Two of those three are still here. — Manager coaching improved because the feedback was concrete, not vague. If your exit interviews are producing nothing useful, change the question this week. Your move.

We were treating attrition as an employee problem. Then we cut the data differently. Our regrettable attrition was 14%, slightly above our band. We had a 47-page engagement survey explaining why. None of it was operationally useful. Then our People analyst cut the data by manager instead of by team. 62% of our regrettable exits, over 18 months, were under just 11 managers — out of 84. Further in: those 11 managers had three things in common. 1. They had been promoted into people-management within the prior 18 months. 2. Their direct-report 1:1 coverage was under 70% (vs ~92% company average). 3. Their teams' "would recommend manager" score was 30+ points below median. We weren't losing employees. We were losing them to a specific group of unsupported managers. What we did: — Built a structured manager-onboarding program. Six weeks. Mandatory for every new people manager. — Monthly skip-levels for the first six months of any new manager's tenure. — Quarterly "manager NPS" pulse with five questions, results back to the manager and their manager. Four quarters later, regrettable attrition dropped to 9%. New-manager team eNPS came up 26 points. "People leave managers" is a cliché. Cut your own data this quarter. You may find it's truer than you knew.

Unlimited PTO is a perk for the company, not the people. I'll defend that. We ran unlimited PTO for three years. The honest data: — Average days taken: 11. — Median for new hires: 6. — Top quartile (mostly senior leaders): 22. — Bottom quartile (mostly junior staff): 5. The people most likely to burn out were taking the least time off. The people most able to take time off were taking the most. The accounting benefit was real for the company. We didn't accrue PTO liability on the balance sheet. The cultural cost was real for the team. What we replaced it with: a minimum-PTO policy. 20 days required. Managers held accountable for ensuring their team takes them. Anyone who finishes the year under 20 days requires an explanation at the next manager review cycle — not from the employee, from the manager. First-year results: — Average days taken: 18. — Bottom quartile: 14. — Burnout self-report (annual survey): down 11 points. Unlimited PTO is a beautiful brochure. It's also a policy that systematically transfers permission from the employer to the employee, then watches the most cautious employees take the least. Minimum PTO is less marketable. It's also more honest. If you're inside a company running unlimited PTO, audit the usage distribution this quarter. If it looks like ours did, you have data to make a different decision.

Our time-to-hire was 52 days. Six months later it's 31. We did not lower the bar. Here's what changed. 1. Pre-booked interview slots. Each hiring loop has 4 interviewers; we used to email back-and-forth on scheduling, losing 5-9 days per loop. Now interviewers hold two 90-minute blocks on Thursday afternoons. We schedule into the blocks. Saved 7 days, instantly. 2. Structured debriefs within 24 hours. Same day if possible. Each interviewer submits a written scorecard before the debrief — no group-think bias. The hiring decision is made within one business day of the final interview, not three. 3. Killed two interview rounds. We had a "behavioral screen" and a "culture fit" round that produced overlapping signal. Merged them into one 45-minute conversation. Saved 6 days. 4. Pre-approved comp bands. Recruiters could verbally confirm a band at offer stage without re-loops to comp. Cut 3 days of back-channel. 5. Reference checks during, not after. We started references after the final interview, not after the verbal offer. If a reference surfaced a flag, we caught it before the candidate had moved on emotionally to us. Quality of hire metric (90-day manager rating): unchanged. The bar didn't move. The friction did. If your time-to-hire is over 40 days, audit your scheduling and debrief gaps first. Most of the loss is there, not in candidate quality.

I've delivered the layoff message dozens of times. Sharing what I've learned about doing it well — and what I regretted. The first time, fifteen years ago, I opened with "thank you for your time and dedication to the company." The person stopped me mid-sentence. "Are you firing me?" They knew. The thank-you came first because I was avoiding the sentence. What I do now: First sentence: "This is going to be a hard conversation. We're eliminating your role, and today is your last day." Direct. The clarity is the kindness. Then I stop. They speak first. Or they don't. Either is fine. I don't fill silence with explanation. When they speak, I answer one question at a time. I don't pile on context. Their working memory in this moment is two questions wide, not ten. I cover three things: the decision is final, the timeline of logistics, and who to contact in the next 48 hours. Then I ask: "Is there anything you need me to handle today before you leave?" What I regret about my first decade in this role: the apologetic tone. The "I know this is hard" before they'd told me it was hard. The leaking of my own discomfort into a moment that wasn't about me. The layoff conversation is the hardest thing in HR. Doing it well is the smallest dignity we can offer. If you're about to deliver one this week, take a walk first. Show up clear. Let them have the moment.

A manager came to me ready to terminate an employee for "not delivering." Eighteen months on the team. Performance review history was solid until six months ago. Then the wheels came off. My first question wasn't "what's the documented performance issue." It was "can I see how the role has been described in the last three review cycles?" Three versions of the job description. Same title. Different jobs. Version 1 (hire): individual contributor, technical execution, project ownership. Version 2 (one year in): "plus team leadership of two junior staff." Version 3 (six months ago): "plus stakeholder management for cross-functional alignment with sales and CS." The employee had been hired for one job. Through reorgs and growth, they were now doing three. The performance drop coincided exactly with the third version. This wasn't a performance problem. It was a role design problem. What we did: — Split the role. Hired a stakeholder-management lead for the cross-functional piece. — Rewrote the original role back to versions 1+2. — Coached the manager on what "adding scope" actually costs. The employee stayed. They got promoted nine months later for the work they were originally hired to do. HR's diagnostic seat is the most under-used asset in most companies. Managers don't always know what they're actually asking for. We do. Next time someone walks in wanting to fire — check the job description first.

Five signs your engagement survey is measuring fear, not engagement. 1. Struggling teams score higher than thriving teams. If your team with 18% attrition has a 9.1 engagement score, your team isn't engaged — they're hedging. The high score is performance, not signal. 2. Anonymity is technically true but practically false. "We don't share results below a team of 10" sounds private. In a team of 11, six people responding makes you identifiable. People know this. They write what's safe. 3. Response rates jump after the manager mentions the survey in the team meeting. If 40% respond before the all-hands and 90% after, the response is now performative. 4. Open-text comments are bland everywhere except in teams nobody is afraid of. If only your highest-trust team gives sharp open-text feedback, you're not getting data from the rest — you're getting compliance. 5. Scores improve over time without any operational changes. Engagement doesn't drift up on its own. If your trendline is climbing and you haven't done anything, your survey is measuring how well people learn to take the survey. What to do: — Use a third-party tool with credible anonymity guarantees. — Set a minimum responder threshold (we use 7) and never report below it. — Add one open-ended question whose answers you publish company-wide. — Stop incentivizing high response rates with team-level competitions. The survey isn't a thermometer if everyone is squeezing it. Fix the system first. The numbers second.

We hired her in October. By December she was running her own team. By April she'd been promoted. In February, on a hunch, I asked our talent ops lead to run her resume back through the AI screener we use for first-pass. She was filtered out. Rated "below qualification threshold." We audited 200 of the screener's recent rejections. What we found: — 23 had non-linear career paths the model interpreted as inconsistency. Three of those, we already knew were excellent candidates from referrals. — 14 had returning-parent gaps. The model treated them as departures. — 11 had military backgrounds where rank terms didn't map to corporate titles. — 8 were career changers whose former-industry experience the model couldn't weight. That's 56 of 200 rejections that needed a second look. We re-interviewed 12. We hired three. What we changed: — AI screening is now a sort, not a filter. Bottom-rated resumes still get a 30-second human glance before rejection. — We retrained the model on our actual high-performer resumes, not on a generic resume database. — Quarterly audits of the rejected pile, with a sample sent to a human reviewer. AI in hiring is going to compound. The companies winning will be the ones treating it as a tool the team supervises, not a system the team obeys. When did you last audit your rejected pile?

Inside a comp review cycle: how we handle the band collisions employees never see. Every year, we have employees whose current comp is higher than the band midpoint for their level, sometimes higher than the band maximum. This happens for three reasons: legacy comp from before we tightened the bands, equity grants made during scarcity periods, and individual deals struck during competitive offers. In a normal cycle, these employees create a problem: their merit raise would push them well above the band, but freezing them feels like a punishment for being valuable. Three principles we follow: 1. Above-band comp is a fact, not a moral judgment. We don't claw back. We don't "freeze until they grow into the band." We treat the band as the structure for new hires and promotions, not as a ceiling for existing employees. 2. Raises slow but don't stop. Above-band employees get half the percentage raise of below-band peers performing at the same level. We're explicit about why. 3. The conversation happens annually, not just when comp moves. Managers are coached to have the band conversation at goal-setting time, not at raise time. Surprises kill trust. What we communicate to the employee: — Their current comp relative to band. — The slowing-raise logic. — The path: if they want band-rate raises again, they need to grow into the next level. Here are the criteria. Most employees handle this fine when it's transparent. The disasters come from managers avoiding the conversation for two years and then delivering a 1% raise without explanation. Comp transparency is uncomfortable. Comp opacity is worse.

People leaders, low-judgment Tuesday question. What policy did you kill in the last 18 months that nobody missed? I'll start. We killed the formal performance rating scale. The 1-5 numbers. We had them for eight years. Managers spent hours calibrating. Employees obsessed over the digit. The actual feedback got buried under the score. We replaced it with three written paragraphs from the manager: what worked, what didn't, what's next. Nobody asked for the numbers back. Manager prep time dropped. Employee satisfaction with reviews went up 18 points. The rating scale survived because the rating scale survived. Not because anybody loved it. Your turn. What did you cut? The comments on this kind of post are the actual content. I'll learn more from your answers than from another HR conference. Kill the dead policy this quarter.