I’ve been VP of Engineering for 18 months now, and I’m constantly navigating this tension: be the empathetic leader everyone says drives great outcomes, while also delivering the business results that keep the company alive.
The narrative is everywhere—human-centered leadership prioritizes well-being, growth, and empowerment. Care about your people, and they’ll care about the work. Sounds great. And honestly, there’s data to back it up:
- Research shows 415-700% ROI on human-centered leadership practices
- 18% higher team engagement
- 23% improvement in organizational performance
- 37% increase in productivity from empathy and empowerment initiatives
Those numbers are compelling. And in my own experience? The most empathetic moves I’ve made—flexible work arrangements, robust learning budgets, mental health days, structured career development—have genuinely improved retention and velocity on my teams.
But here’s what keeps me up at night:
If human-centered leadership is so effective, why is global employee engagement stuck at 23%? It went from 21% in 2024 to 23% in 2025. That’s… barely moving.
Are these ROI studies measuring correlation or causation? Are companies with better cultures just better at everything, including leadership? Is there selection bias in who participates in these studies?
When I sit in board meetings, I can’t just say “we should invest in empathy because it feels right.” I need to show metrics. But which metrics actually matter?
The questions I’m wrestling with:
-
How do we prove impact beyond anecdotes? Retention is easier to track, but how do you measure psychological safety, trust, or sense of belonging as leading indicators?
-
What separates evidence-based human-centered practices from feel-good virtue signaling? Not all “empathy” initiatives are equal.
-
How do you balance high empathy with high accountability? I’ve seen leaders swing too far toward “caring” and lose the ability to make hard calls.
-
What would actually convince a skeptical CFO or board member? What data would make them say “yes, invest here”?
I believe in this approach—I’ve seen it work. But I also know belief isn’t enough. We need to get better at translating “human-centered leadership” from a philosophy into a measurable performance strategy.
What evidence has worked for you? What metrics have you used to demonstrate impact?
I hear you on this, Keisha. I’ve been wrestling with the same tension, but I’ve learned that not all “empathy” is created equal.
A few years ago, I tried unlimited PTO thinking it would show we trusted our team. Disaster. People didn’t take time off because there were no norms, and managers weren’t tracking burnout. It became performative—we said we cared, but the structure didn’t support it.
Then we tried structured learning time—4 hours/week for skill development, with quarterly presentations on what people learned. That actually worked. Productivity went up 31% over six months because people were solving problems with new skills instead of brute-forcing with old approaches.
Here’s what I think the research misses:
The 415-700% ROI numbers lump together all “human-centered” practices. But which ones actually drive performance? I’d bet most of that ROI comes from a small set of high-leverage interventions:
- Structured 1:1s with clear development plans (not venting sessions)
- Psychological safety that enables fast escalation of problems
- Autonomy on HOW, but clarity on WHAT and WHY
- Real career progression, not just titles
The stuff that doesn’t move the needle? Pizza parties. “Wellness” initiatives that ignore systemic workload problems. Empathy theater.
My question back to you: How do we separate evidence-based practices from virtue signaling? I think we need to stop treating “human-centered leadership” as a monolith and start asking which specific interventions drive which outcomes.
Keisha, this is exactly the conversation I’ve been trying to have with our executive team.
From a CFO lens, here’s the problem: The ROI studies on wellness programs cite $1.65-$1.88 return per dollar invested. That sounds great until you actually audit the measurement methodology. Most studies:
- Cherry-pick participants (people who opted in are already healthier/more engaged)
- Measure “productivity” through self-reported surveys (bias much?)
- Attribute savings to wellness programs that might have happened anyway
- Use 3-5 year timeframes that make causation nearly impossible to prove
What I actually track:
-
Attrition costs - We spend ~$150K to replace a senior engineer (recruiting, onboarding, lost productivity). Teams with better manager relationships have 20% lower attrition. That’s real money.
-
Time-to-productivity - New hires on teams with strong onboarding/mentoring are productive 40% faster.
-
Revenue per employee - Our best-performing teams have high autonomy + high clarity. That translates to measurable output.
Here’s my controversial take: Maybe the problem is that most “human-centered” programs aren’t designed with ROI in mind from the start. They’re designed to make leadership feel good about caring.
If you want CFO buy-in, you need to speak our language:
- What’s the unit economics of this intervention?
- What’s the time to payback?
- How do we measure leading indicators (not just engagement surveys)?
- What’s our confidence interval on the causal relationship?
I’ve approved every empathy-driven investment Luis and his team have proposed—because they came with a financial model, not just feelings.
Keisha, you’re asking the right question, but I think the framing is slightly off.
The question isn’t “does human-centered leadership enhance performance?” The question is “why are we treating this as a trade-off?”
I’ve been in this industry for 25 years. I’ve seen toxic high-performers burn out entire teams, and I’ve seen “nice” organizations plateau because they couldn’t make hard calls. Neither extreme works.
The best teams I’ve built have high expectations AND high support.
Here’s what 25 years has taught me:
-
Change management quality predicts engagement more than anything else. Research backs this up. When people understand why changes are happening and have agency in how they’re executed, engagement follows. It’s not about being “nice,” it’s about being clear and involving people in problem-solving.
-
The leaders who struggle with “does empathy work?” usually haven’t defined clear outcomes. They confuse empathy with low standards. That’s not human-centered leadership—that’s abdication.
-
Psychological safety isn’t about comfort, it’s about candor. Best-performing teams can surface problems early, disagree constructively, and hold each other accountable. That requires trust, not niceness.
My tactical approach:
- Set ambitious goals (high bar)
- Give people resources, autonomy, and context to achieve them (high support)
- Remove blockers ruthlessly (high agency)
- Hold people accountable to outcomes, not activities (high clarity)
On your metrics question: Stop tracking “engagement” as the goal. Track:
- Cycle time (how fast can we ship value?)
- Incident recovery time (how fast do we bounce back?)
- Internal promotion rate (are we growing our people?)
- Voluntary attrition of high performers (are our best people staying?)
Human-centered leadership isn’t a performance strategy—it IS performance strategy. But only when it’s operationalized correctly.
Keisha, I’m watching this from the product side, and I have a different lens on your question.
Your team ships faster AND with fewer production incidents than the previous VP’s team. I know because I track product velocity and quality closely, and there’s a clear inflection point when you took over 18 months ago.
What changed? People actually surface blockers early instead of hiding problems until they explode.
When your engineers say “this technical debt is going to slow us down,” they’re not punished for “negativity”—they’re listened to, and we make trade-offs together. That’s psychological safety in action, and it directly impacts our roadmap execution.
The cross-functional impact of human-centered engineering leadership:
- Product-eng partnerships are smoother because trust is higher
- Feature scoping conversations are more honest (less sandbagging, less overpromising)
- Post-mortems actually identify root causes instead of scapegoating
But I’ve also seen the failure mode: “Empathetic” leaders who can’t make hard calls. Performance issues that drag on for months. Projects that should be killed but aren’t because “the team worked so hard.” Technical directions that everyone knows are wrong but no one wants to hurt feelings by saying so.
So here’s my question back: How do you measure psychological safety as a leading indicator of velocity? Because I think that’s the link. Empathy without safety is just being nice. Safety without accountability is chaos. But safety + accountability = speed.