62% to 42%: closing the leader-employee gap in 2026

Last week HR Dive surfaced a number from a fresh Dayforce report that should make every people leader in a 200- to 800-person company stop scrolling. Only 42% of front-line workers now believe their company's leaders understand the problems they face every day. In 2024 that number was 62%. Two years, twenty points down. That is not a slow drift. That is a hinge.

The headline framing on HR Dive leaned into AI-driven career disruption, but the more interesting subplot is the listening gap underneath it. Layoffs cooled in 2025. Wage growth flattened. Return-to-office settled. None of that is what's driving the disconnect. What's driving it is that leaders made a lot of decisions about AI rollouts, hybrid policies, and team restructures over the last 24 months without a feedback channel that actually surfaced what was happening to the people doing the work.

For a mid-size company, this is the moment to fix that. You cannot outspend the 5,000-person enterprise on perks. You absolutely can outpace it on listening cadence, because your loops are shorter and your decisions move faster.

Why annual engagement surveys stopped catching this

The default answer for the last decade has been an annual engagement survey. Once a year, 60 questions, an outside vendor, a sixty-page deck with quadrant heatmaps, a couple of action plans that maybe get traction by Q3, then another twelve months of quiet. By the time the 2025 results landed, the 2024-to-2026 collapse was already half-baked.

The problem is not the instrument. The problem is the cadence. A 60-question instrument tells you a lot about a moment in time. It tells you nothing about momentum. If you wanted to see a 62% number become a 42% number, you would need to be sampling something every four to six weeks, not every twelve months. Most HR teams know this. Few have rebuilt their stack around it.

The four-signal listening stack

Replace one annual instrument with four lightweight ones that each catch a different angle. None of them are new. Together they form a pattern.

Pulse surveys, monthly or bi-weekly. Five to eight questions, anonymous, rotating between core themes: clarity of direction, relationship with manager, workload, recognition, growth. The point is not the score. The point is the delta. A team eNPS that drifts from +32 to +14 over six weeks is a leading indicator. You can call the manager, you can probe, you can fix something. Wait for the annual and you'll be reading exit interviews instead.

eNPS as a steady backbone. One question, asked the same way every month: would you recommend this place as somewhere to work? It's a coarse instrument on its own, but its job is not nuance. Its job is to be the heartbeat that says "something changed" so the pulse questionnaire can ask why.

Structured 1-on-1s. Weekly or biweekly, owned by the report, not the manager. Thirty minutes, written agenda, action items that carry over until they're done. This is where the words come out that surveys never capture. "I don't think my work is visible to anyone above my manager." "I haven't grown in nine months." A 1-on-1 with a real agenda surfaces that conversation a quarter earlier than a survey can.

360 reviews, twice a year. Not for ranking, not for compensation gates. For the employee. A 360 done right gives each person a clean read on how they're seen by their manager, their reports, and their peers, across the competencies that matter for their next step. The trick is splitting the averages by relationship — a single mean across managers, peers, and reports is mathematically noisy and almost always misleading. Managers and peers usually weigh different competencies differently, and when they do, that information is where the growth conversation actually happens.

Used together, the four signals catch different time scales. Pulse catches the week. 1-on-1 catches the month. eNPS catches the quarter. 360 catches the year. When something starts to break, at least one of them sees it first, and the others corroborate.

What "leaders get us" looks like operationally

The Dayforce number isn't measuring whether the CEO is a nice person. It's measuring whether the operating system around the work transmits signal between layers. A leader who reads ten anonymized verbatim comments per week from a pulse survey, runs a 30-minute open-floor session once a quarter on the three themes that keep showing up, and walks into 1-on-1s with their direct reports having actually read their last self-assessment — that leader registers as "gets us" in a way that's hard to fake from a kickoff video.

The opposite is also operational. A leader who hasn't seen a single piece of unfiltered feedback in nine months, whose 1-on-1s with their reports are status updates, and who only sees engagement data through a quarterly board slide — that leader is going to score 42%, or lower, no matter how many town halls they do.

This is why the listening stack matters more than any single tool inside it. The four signals don't fix the disconnect on their own. They make the disconnect visible early enough that something can actually be done.

Where DTPulse fits

DTPulse is the HR portal that holds these four signals in one place for companies in the 20- to 1000-person range. Pulse and eNPS share a question bank with rotation and anonymity enforced at the schema level (no employee ID stored on responses; demographic slices aggregated only when group size meets the k-anonymity threshold). 1-on-1s have a per-pair agenda with carry-over action items, so the conversation builds. The 360 module splits scores by manager, report, and peer cohorts on the same axes you set up for the competency review.

The point of the integration isn't tooling cleverness. It's that when something starts to slip, the people-leader sees it in one place rather than chasing answers across three Google Forms exports and a Slack thread. The four-signal stack only works if the latency between "something changed" and "someone acted on it" stays short. Single portal, short latency.

If you want to read the framing for one of those signals on its own, our note on eNPS done right walks through the anonymity, cadence, and follow-up that separate a useful eNPS from a vanity metric.

What to measure if you start now

Pick a baseline before you change anything. Three numbers:

• Current eNPS, one question, asked the same way you'll keep asking it
• Voluntary attrition over the last 90 days for people in your top performance band
• Percentage of direct reports who can describe their next career step in two sentences without referring to a document

Run the four-signal stack for a quarter. Check the same three numbers. If eNPS moves up five points or more, top-band attrition falls, and the next-step-clarity number doubles, you have signal-to-noise. If nothing moves, one of the four practices is happening on paper but not in the work — usually 1-on-1s, sometimes the pulse follow-up loop.

Twenty points off "leaders get us" in two years is a serious number. The 5,000-person enterprises will respond to it with a consulting engagement and a new annual platform. The mid-size companies that win the next two years will respond with a faster loop. Smaller teams, shorter feedback paths, four signals that actually get read. That is the version of "leaders get us" that holds up.