The People Data Trust Gap Nobody Talks About
Marsh's 2026 People Risk report found that employee thriving in the workforce dropped from 66% in 2024 to 44% in 2026. A 22-point fall in two years. Most of the coverage went to enterprise, but the story is more useful for small businesses, where the same problem shows itself a little differently.
Most small companies do not have a people analytics function. They have a founder, a head of operations, and maybe one HR generalist. So when the thriving number drops by a third in two years, the small business version reads like this. The founder does not trust the survey. The team does not trust the survey. Nobody believes the engagement score on the dashboard, and the dashboard keeps reporting it anyway. And this is even if they are measuring engagement and employee experience.
That is the trust gap. It is the reason most employee engagement surveys turn into theater, especially in smaller companies.
Why the Trust Gap Exists in the First Place
Mercer's 2026 Inside Employees' Minds report and Marsh's 2026 People Risk report were published within weeks of each other. Different samples, different questions, but they agree on one thing. People are staying, but they are not okay.
Mercer found that 14% of US workers are no longer actively looking for a new job, even though they would like to, because they are worried about job security with a new employer (10). Marsh found that 26% of employees globally are unsatisfied at work but believe leaving is not an option (22). Different studies say the same thing. People feel trapped.
The Mercer engagement breakdown puts a number on it. 25% highly engaged. 43% moderately engaged. 25% passive. 7% actively disengaged (21). The 25% passive bucket is your group of employees who feel trapped. They give your survey a 10/10 and quit two quarters later, except they don't quit, because they can't.
Let's say a 60 person company runs a quarterly pulse. The CEO opens the dashboard, sees a 4.2 out of 5 on "I feel safe raising concerns to my manager," and walks into the leadership meeting feeling good. Three months later, two senior engineers quit, and the exit interviews revealed manager issues the survey never listed. The survey did not lie, but the team did. They lied because the team did not trust the anonymity, and the people feeling stuck did not see a reason to spend the energy telling the truth.
What the Trust Gap Costs a Small Business
I have seen the cost spread and grow across three aspects of the business.
The first is retention. Low turnover in 2026 is not the win it used to be. Mercer's data shows people are staying put because the labor market is tight, not because they are happy with their experience. In the first three quarters of 2025, US employers announced 1.10 million job cuts, up 65% from the same period the year before (11). Disengaged but trapped employees cost more than disengaged who leave, because they sit in the role, miss the goals (or just meet them), and watch the team underperform around them. If your survey is not catching them, you have quite a blind spot.
The second is decision quality. Mercer found that 53% of US workers agree that new technology will impact their job security (15). Marsh found that 40% of employees globally are concerned about losing their job to AI (16). Companies making AI policy on fake engagement data are going to make worse policy. Small businesses do not have the cushion to absorb a wrong move and recover. The founder needs to know whether the team is optimistic about the rollout or if they’re working with an LLM to revamp their résumé while saying it's fine.
The third is the founder's own credibility. If the team has answered an engagement survey twice and watched nothing change, they will not answer the third one honestly. That is not cynicism on the team's part. They just recognize a pattern and assume nothing will change so why even bother? Then the data goes flat, the founder is no longer aligned with the team, and the team stops telling them the truth.
What to Do Instead, in Order
Start with anonymization that the team believes and can trust. Most surveys claim anonymity and then route responses through a tool that the founder can filter by team, tenure, and department until there are three people left in the bucket. Everyone on the team knows this. Run a survey where the anonymization is verified by a human, not assumed by the platform, and show the team exactly how it works. The trust comes from the process being explainable and the transparency in what leadership sees.
Ask fewer questions and ask them more often. A 40-question annual survey at a 60-person company is a research project the team treats like a homework assignment. A 5 to 8 question quarterly pulse is a conversation. Pulse surveys give you trend data, which is what people analytics is supposed to deliver in the first place. The shorter format reduces survey fatigue enough that response quality goes up.
Show the team what changed because of what they said. This is where the majority of surveys fall apart. The data comes in, the leadership team reads it, and the team never hears about it again. Pick one finding per quarter, name the change, and report back. If the survey said meetings are killing focus time, and you killed two recurring meetings, say so, specifically, by name.
Pair the engagement data with behavioral data. The engagement number tells you what the team feels. A DISC profile of the team tells you how to talk to each person about it. Marsh found that inadequate leadership skills trigger or worsen more downstream people risks than any other factor they tested. And that’s across all 25 risks they measured (20). So the founder reading a low score on psychological safety from a team full of high S and high C profiles is not just misreading the data. They're sitting on the single biggest cascade variable in the business, and the survey on its own won't tell them how to act on it.
A related Mercer finding worth noting. Employee confidence in which skills matter has dropped (16). "I have a good understanding of the skills I need to progress" went from 89% favorable in 2023 to 83% in 2025. Six points in two years, on a question employees can usually answer with conviction. If people don't know what to learn next, they don't know what to flag on a survey either. They answer what's in front of them and leave the bigger worry off the page.
Finally, do not run people analytics from inside the company alone. Asking the people who manage the team to also interpret how the team feels about being managed is a conflict. A second source of analysis, even a light one, gives the data a kind of independence the team can see and feel. And that will get you honesty.
Before You Read the Next Dashboard
Before you read the next dashboard, look at how many people wrote something in the free-text boxes. Clicking 4 out of 5 or 9 out of 10 takes three seconds. Writing a sentence takes a minute and could expose you. A survey with 95% completion and 8% comments is the team showing up to satisfy the mandate and going home. The radio buttons are the filler, but the comments are the truth.
What's one thing your team has heard back from leadership because of an engagement survey this year?
Sources: Mercer, Inside Employees' Minds 2026 (survey of 4,500+ US employees, fielded September–October 2025). Marsh, People Risk 2026 (survey of 4,517 HR and Risk professionals across 26 markets, fielded October–November 2025). Page references throughout refer to each report's published PDF.
Michael Franco runs Quokka Hub, an anonymous engagement survey and DISC platform for 50-500 person companies. He writes about people ops for founders and HR leaders. https://www.linkedin.com/in/michaelrfranco/