'Misalignment' Mentions Up 149% on Glassdoor — The Manager-IC Trust Crisis Needs More Than Town Halls

[vp_eng_keisha]

I’ve been studying the Glassdoor sentiment data that came out of SignalFire’s research, and one number stopped me cold: “misalignment” mentions in engineering reviews are up 149%.

Not 14.9%. Not 49%. One hundred and forty-nine percent.

That’s not a trend. That’s a structural failure in how engineering organizations communicate between leadership and individual contributors. And I don’t think most companies understand how deep the problem goes.

The Full Picture on Trust Erosion

Let’s look at the complete Glassdoor sentiment data for engineering-specific reviews:

• “Disconnect” mentions: up 24%
• “Miscommunication” mentions: up 25%
• “Distrust” mentions: up 26%
• “Misalignment” mentions: up 149%

The first three are concerning but expected — after years of layoffs, RTO mandates, and shifting strategies, some erosion in trust is predictable. But “misalignment” at 149% tells a different story. Engineers aren’t just losing trust in their leaders. They’re saying that leadership’s stated priorities don’t match actual priorities.

This is the gap between “we value engineering excellence” and then cutting the platform team. Between “we believe in work-life balance” and then mandating return-to-office. Between “people are our greatest asset” and then doing layoffs by algorithm.

What I’m Seeing in EdTech

I run engineering at an EdTech startup, and we went through our own trust crisis after a round of layoffs in 2024. Here’s what happened and what we did about it:

The Problem

After we let go of 15% of engineering (driven by runway concerns, but communicated poorly), our internal pulse survey showed:

• Trust in leadership dropped from 78% to 41%
• “I understand the company’s strategic direction” dropped from 82% to 39%
• “My manager advocates for my team” dropped from 85% to 52%

The engineers who survived the layoff weren’t grateful — they were traumatized and suspicious. Every ambiguous Slack message from leadership was interpreted through the lens of “what are they not telling us?”

What Didn’t Work

• Town halls: We did three in two weeks. Attendance was high, but the anonymous Q&A was brutal. Engineers didn’t want presentations — they wanted accountability.
• Manager talking points: We gave managers scripts to use in 1:1s. Engineers immediately recognized the corporate language and trust eroded further. “When my manager sounds like a press release, I know they’re not being honest with me.”
• “Our values” messaging: Referencing company values felt tone-deaf when the values had just been violated. You can’t cite “we put people first” two weeks after a layoff.

What Actually Worked

1. Radical financial transparency

I started sharing actual financial data in engineering all-hands. Not “we’re in a strong position” — actual runway numbers, revenue trends, unit economics. When engineers could see the math themselves, they could evaluate decisions rationally instead of through a lens of fear.

2. Decision architecture documentation

For every significant decision (reorgs, project cancellations, policy changes), we now publish a decision document that includes:

• The options we considered
• The data we used
• The tradeoffs we accepted
• What we got wrong in hindsight

This isn’t about making every decision democratic. It’s about showing the reasoning. Engineers are fundamentally logical people — they can accept a decision they disagree with if they understand the reasoning was sound.

3. Skip-level trust audits

Every quarter, I do skip-level conversations (not meetings — conversations) with 10-15 ICs across the org. No managers present. No agenda. Just: “What’s working? What’s broken? What are you worried about?”

These conversations surfaced problems that weren’t showing up in surveys. One engineer told me their team lead was taking credit for their work. Another said a project was six weeks behind but no one wanted to tell the product team. A third said they were job hunting because their manager had promised a promotion that never materialized.

4. Manager accountability with teeth

This is the hard one. We started measuring managers on trust metrics — not just output. If your team’s trust scores drop below a threshold, you go into a coaching program. If they don’t recover, you’re moved off the management track. We’ve moved three managers in the last year. Each time, the team’s trust scores recovered within one quarter.

The Post-Layoff Trust Playbook

After 18 months of intentional work, here’s where we are:

• Trust in leadership: 41% → 73% (not back to pre-layoff levels, but trending right)
• “I understand strategic direction”: 39% → 81%
• “My manager advocates for my team”: 52% → 79%
• Voluntary attrition: dropped from 28% (crisis period) to 13%

The 149% increase in “misalignment” mentions on Glassdoor tells me most companies aren’t doing this work. They’re hoping that if they stop talking about the layoffs and push forward with new initiatives, engineers will forget. Engineers don’t forget. They update their priors.

The EdTech Advantage

One thing that helps in EdTech: our mission is inherently meaningful. We’re building tools that help students learn. When trust wobbles, we can re-anchor to purpose. Not every industry has that luxury, but every company has some version of customer impact that can serve as a north star.

For those of you managing through trust crises — what’s worked? And more importantly, what looked like it would work but didn’t? I think the failure stories are more valuable than the success stories here.

[cto_michelle]

Keisha, your post is important and I want to address the executive accountability angle directly — because I think that’s where the trust chain actually breaks.

The Executive Credibility Gap

Here’s what I’ve learned after 25 years and multiple leadership roles: trust doesn’t flow from town halls. It flows from consistent behavior over time. And the reason “misalignment” is up 149% is that executive behavior has been wildly inconsistent over the past 3 years.

Consider what many engineering organizations experienced between 2021 and 2025:

• 2021: “We’re going remote-first! The future of work!”
• 2022: “We’re hiring aggressively! Growth mode!”
• 2023: “We’re doing layoffs for efficiency. But we still value you.”
• 2024: “Return to office. Collaboration is essential.”
• 2025: “Another round of cuts. But trust us, this is the last one.”

Engineers are pattern-matching on this sequence, and the pattern they see is: leadership says whatever is convenient for the current quarter’s narrative. That’s not misalignment. That’s credibility destruction.

What Executive Accountability Actually Looks Like

When we did layoffs at my previous company, I insisted on three things:

1. Executives take the first hit. Before any IC was affected, the C-suite took a 20% salary reduction. Not because it materially changed the budget — but because it demonstrated shared sacrifice. Engineers notice when the CEO’s compensation doesn’t change while their colleagues are being walked out.

2. Named accountability for decisions. Every layoff decision had a specific executive’s name attached to it. Not “the company decided” — “I, Michelle Washington, made this decision based on this data.” When you strip away the passive voice, you force honest evaluation of whether the decision was sound.

3. Post-decision retrospectives. Six months after our layoff, I published an internal retrospective that included what we got wrong. We over-cut in platform engineering and had to re-hire 3 months later. I named that as my mistake publicly. Not to self-flagellate, but because if leaders can’t admit errors, engineers correctly conclude that feedback only flows one direction.

The Trust Recovery Timeline

Your data showing trust recovery from 41% to 73% over 18 months matches what I’ve seen. But here’s the uncomfortable truth: that remaining gap might be permanent. Some engineers will never fully trust leadership again, and that’s a rational response given what they experienced. The goal isn’t getting back to 100%. It’s building a system where trust can function at 75% because the accountability structures are strong enough to compensate.

Your decision architecture documentation approach is exactly right. Transparency isn’t about making everyone happy with every decision. It’s about respecting engineers enough to show your work.

[eng_director_luis]

Keisha, your skip-level trust audit approach is something I want to adopt, and I want to share the middle management perspective because I think it’s the most compressed layer in this trust crisis.

The Middle Management Squeeze

Here’s what most discussions about manager-IC trust miss: middle managers are often the most misaligned people in the organization, and it’s not entirely their fault.

As a Director, I sit between VP/C-suite strategy and IC execution. When leadership makes a decision I disagree with — like the time our CEO pushed for an aggressive RTO timeline — I have three options:

1. Champion it authentically (hard when I genuinely disagree)
2. Relay it neutrally (engineers immediately detect the lack of conviction)
3. Push back upward (which sometimes works but often just delays the inevitable)

The 149% increase in “misalignment” partially reflects this squeeze. When engineers say their leadership is misaligned, they often mean their direct manager is saying one thing in 1:1s and the company is doing another. That’s because their manager is navigating the same dissonance they are.

What I’ve Changed

After reading similar data internally, I made three changes to how I operate:

1. I stopped defending decisions I disagree with.

When leadership mandated something I thought was wrong, I used to try to put a positive spin on it. Now I say: “Leadership made this decision. Here’s their reasoning as I understand it. Here’s what I agree with and here’s where I have concerns. I’ve raised my concerns upward, and the decision stands.”

Engineers respect this dramatically more than the corporate spin. They can see I’m being honest about my own position while still executing on the organizational direction.

2. I publish my “disagree and commit” log.

Every quarter, I share with my managers a list of decisions I disagreed with but committed to executing. This does two things: it normalizes disagreement, and it shows that commitment doesn’t require agreement. The “misalignment” engineers are detecting is often the absence of this honest acknowledgment.

3. I hold my managers accountable for trust, not just delivery.

Echoing what you’re doing, Keisha — if a manager’s team trust scores decline, we have a real conversation. Not “try harder” — a structured examination of what behaviors are eroding trust and what changes need to happen.

The hardest conversation I’ve had this year was with a manager who was technically excellent but had been promising promotions he couldn’t deliver. His team’s trust was at 38%. We co-created a recovery plan, and he’s now at 62%. The key was treating trust erosion as a skill gap to be coached, not a character flaw to be punished.

Your approach of moving managers off the track when coaching fails is brave and necessary. Most companies don’t have the courage to do it.

[data_rachel]

Keisha, the data scientist in me is fascinated by the Glassdoor sentiment analysis, and I want to dig into how to measure trust systematically — because I think most organizations are flying blind on this.

Why Traditional Trust Measurement Fails

Most companies measure trust through quarterly or annual engagement surveys. Here’s why that’s insufficient:

Frequency problem: Trust erodes in days and weeks, not quarters. By the time your Q3 survey shows a trust decline, you’ve already lost people. The Glassdoor data is actually a lagging indicator of problems that started 6-12 months before the reviews were written.

Honesty problem: Even “anonymous” surveys aren’t trusted by engineers who’ve seen how data gets used. At my company, I ran an experiment: we asked the same trust questions through our official survey tool and through a truly anonymous external tool. The external tool showed trust scores 12 points lower on average. Engineers self-censor when they think their manager might see the results.

Aggregation problem: “Trust in leadership” as a single metric hides enormous variation. Trust in your direct manager, trust in your skip-level, trust in the C-suite, and trust in the board can all be at wildly different levels. Lumping them together makes the data actionable for no one.

A Better Trust Measurement Framework

Here’s what I’ve built and what I’d recommend based on combining behavioral data with targeted micro-surveys:

Leading Indicators (Weekly/Biweekly)

• 1:1 cancellation rates by managers — consistent cancellations correlate with 2x higher team attrition within 6 months
• Anonymous “traffic light” check-ins — single question, red/yellow/green, tracked weekly. Changes in pattern are more informative than absolute levels
• Code review turnaround time — when trust erodes, engineers stop investing in thorough reviews. If your average PR review drops from 4 hours to 30 minutes, that’s a signal
• Internal job board activity — anonymized tracking of which teams’ engineers are browsing internal transfers

Mid-Cycle Indicators (Monthly)

• Skip-level meeting request rates — when engineers start requesting skip-levels, it often means they’ve lost faith in their direct manager
• Documentation contribution — high-trust teams document more because they believe the organization will still exist and benefit from the documentation
• Cross-team collaboration frequency — trust correlates with willingness to work across boundaries

The NLP Approach

What I’d love to see is companies applying the same NLP analysis that was done on Glassdoor reviews to internal communications (with proper consent and anonymization). The linguistic markers of trust erosion — hedging language, passive voice, decreased future-tense statements — show up in Slack messages and documents before they show up in surveys.

Your 41% to 73% recovery is impressive, Keisha. I’m curious: what was your measurement cadence? Monthly pulse surveys? And did you see any leading indicators that predicted the recovery before the survey numbers moved?