In 2025, global employee engagement fell to 20% — its lowest level since 2020, and the second consecutive year of decline. Manager engagement dropped even more sharply, falling five points to 22%. Daily stress sat above pre-pandemic levels. Gallup put the cost of this disengagement at $10 trillion annually, roughly 9% of global GDP.
And yet, walk into most HR functions and you’ll find dashboards full of green numbers. Engagement scores holding. Training completion rates climbing. Time-to-hire shrinking. eNPS within target. Wellness program participation up year over year.
Both pictures are true at the same time. That is the problem.
HR has never had more data, more dashboards, or more measurement infrastructure than it does today. It has also never had less confidence that the numbers reflect reality. Something has gone wrong in the gap between what we measure and what we actually need to know — and the reason is older than any of the platforms producing the dashboards.
The Law That Explains the Gap
In 1975, the British economist Charles Goodhart observed a pattern in monetary policy: “Any observed statistical regularity will tend to collapse once pressure is placed upon it for control purposes.” Two decades later, the anthropologist Marilyn Strathern reformulated it more memorably in a 1997 paper on audit culture in British universities: “When a measure becomes a target, it ceases to be a good measure.”
Goodhart’s Law has since traveled far beyond economics. It explains why standardized testing distorts education, why hospital wait-time targets distort emergency care, and why every system that gets serious about measuring something eventually finds the measurement gaming itself.
It also explains, with uncomfortable precision, what is happening across HR right now.
How Goodhart Lives in HR Dashboards
The gaming rarely looks like cheating. It looks like sensible behavior under pressure.
Consider time-to-hire. When it becomes the dominant measure of recruiting success, the entire system reorganizes around speed. Roles that fill quickly become more visible than roles that fill well. Pipelines get optimized for velocity. The pressure to close shifts the conversation from fit to availability — not because anyone wants that, but because the measurement architecture rewards it. The metric improves. The quality of hiring, measured on a longer horizon, often doesn’t.
Consider engagement scores. Once quarter-over-quarter increases become a leadership expectation, attention naturally concentrates on the survey itself rather than the conditions it was meant to read. Question framing gets refined. Timing gets optimized. Response patterns shift in subtle ways — the most engaged respond first, the most disengaged often opt out altogether. None of this is dishonest. It is what happens when any instrument becomes the thing being managed. The score moves. Whether the underlying experience moved with it is a separate question, and one the score itself can no longer answer.
Consider performance ratings. When ratings drive compensation, promotion, and visibility, the rating itself becomes the focus of the conversation rather than the performance it was meant to describe. Calibration sessions concentrate on defending or adjusting numbers. Forced distributions create pressure to rate within bands rather than within reality. The annual rating ends up reflecting the political and procedural realities of the organization at least as much as it reflects what an individual actually contributed. McKinsey’s recent organizational research has flagged the broader pattern: many organizations are now becoming “more rigorous about measuring what works” precisely because single-metric approaches captured only one part of a much larger picture, leaving the rest of the picture invisible to the dashboards leadership was reviewing.
Consider training completion rates. They climb steadily across most organizations, supported by platforms that make completion straightforward to track. Whether completion translates into capability — whether attendance becomes learning, and learning becomes practice — sits in a different measurement system entirely, often a manual one. The gap between the two is where the real question lives, but it rarely makes it to the dashboard.
In each case, the metric is doing exactly what it was designed to do. The problem is that what it was designed to do has come apart from what the organization actually needs.
Why HR Is Especially Vulnerable
Four structural conditions make HR a particularly fertile ground for Goodhart failure.
Single-number reporting. Boards want one number. CHROs are asked to summarize complex human systems in dashboards small enough to fit on a slide. The compression is brutal. A workforce of 10,000 people, each with their own context, manager, team, life circumstances, and trajectory, gets reduced to an engagement score, a turnover percentage, and an eNPS. Whatever survives the compression becomes the target. Whatever doesn’t, disappears.
Fixed measurement instruments. Most organizations run the same engagement survey, in the same form, for years at a time — partly for trend comparability, partly because changing the instrument is expensive and disruptive. But fixed instruments are exactly the kind that Goodhart’s Law degrades fastest. The longer a question stays the same, the better the organization gets at producing the answer it wants to see, often unconsciously.
Short time horizons. Quarterly reporting rhythms are inherited from finance. They suit revenue and cost. They suit human systems poorly. Burnout, disengagement, cultural drift, and trust erosion compound over months and years, not quarters. Measuring them quarterly is like checking the tide every fifteen seconds — you see ripples, not the underlying movement.
The separation of measurement from meaning. As HR tech proliferated, measurement infrastructure was largely procured separately from the people who interpret it. Surveys come from one vendor, learning data from another, sentiment analysis from a third, exit interview platforms from a fourth. Each one optimizes its own metric. Nobody owns the question of whether the metrics, taken together, tell a coherent story about the workforce.
The result is an HR technology stack that is comprehensive, expensive, and quietly retrospective. Most dashboards tell leaders what already happened, often in the form of metrics that have already lost their meaning by the time they’re reviewed.
The Cost of Getting This Wrong
It would be easy to treat this as an academic problem if the financial cost weren’t so visible.
Gallup’s $10 trillion figure for global disengagement is the headline number, but the operational waste sits closer to home. Recent analysis from major consultancies suggests that intervention waste — money spent on wellness, engagement, or development programs not targeted to specific, evidence-backed needs — can account for 15 to 20% of total HR budgets in mid-to-large enterprises. That is real money, deployed against metrics that have come unmoored from outcomes.
Wellness specifically has been a cautionary case. MIT Sloan researchers have noted that despite the corporate wellness market’s projected trajectory toward $100 billion by the early 2030s, rigorous independent evidence of what actually works remains thin. Randomized controlled trials repeatedly find that platform-reported ROI numbers — based on engagement metrics like app logins, session completions, or self-reported satisfaction — have little to no correlation with actual clinical health outcomes or productivity gains. The metrics look good. The outcomes don’t follow.
This is Goodhart at scale. An entire category of investment, measured by proxies the platforms themselves define, producing rising numbers without the underlying outcomes those numbers were meant to represent.
What Better Measurement Looks Like
The response isn’t to measure less. It’s to measure differently.
The shift that the most thoughtful people analytics functions are now making is from single-number reporting to what could be called measurement triangulation: reading workforce health across multiple independent signals, over multiple time horizons, and treating coherence between them as the actual evidence of progress.
In practice, that means three layers working together.
Immediate signals capture present state. Daily or weekly sentiment pulses, micro-feedback after key moments, behavioral indicators of engagement with work. These are noisy individually and powerful in aggregate. Their value is timeliness, not precision.
Leading indicators capture trajectory. Frequency and quality of manager-employee one-to-ones. Patterns in internal mobility interest. Early signs of disengagement that precede attrition by months. These are the signals that, read well, allow intervention before a problem reaches the outcome metric.
Outcome measures capture what actually shifted. Retention of high performers. Involuntary turnover. Productivity. Claims patterns. Manager effectiveness over time. These are slower and lagging by nature, but they are the proof that the earlier signals were real.
The discipline is in how the three layers cross-validate each other. When immediate signals say morale is high but leading indicators show rising attrition risk, something is masked. When leading indicators predict a problem and outcome measures later confirm it, the prediction model proves itself. When interventions move leading indicators but not outcome measures, the intervention is doing less than it claims.
This is the architecture Deloitte has been pointing toward in recent Human Capital Trends research, with its argument for moving beyond what it calls “big moments” — annual surveys, periodic check-ins — toward connected analytics that read the workforce continuously and across dimensions. McKinsey’s parallel framing of organizational capital makes a similar point: workforce health is a long-term asset, not a monthly score, and it has to be measured on the timescale at which it actually changes.
Three Disciplines Worth Adopting
Three principles distinguish organizations that have started doing this well.
Triangulation over single source. No signal worth acting on rests on a single input. What people say, how they behave, and what leading indicators show all need to align before a number is treated as evidence. One source is information. Three sources, in coherence, is signal.
Rotation over fixation. The instruments themselves evolve. The same question, asked the same way, year after year, eventually measures the response style of the organization rather than the underlying reality. Varying the instrument — different question framings, different timing, different sampling — keeps measurement honest.
Coherence over optimization. The goal is not to maximize any single number. It is to keep the picture coherent, with signals, interventions, and outcomes telling a consistent story. When the layers diverge, that divergence is the alert. When they converge, that convergence is the proof.
The Measurement Conversation HR Needs to Have
HR is being asked to prove its value at exactly the moment its primary measurement tools are losing credibility. That tension is uncomfortable, but it is also clarifying. The functions that get this right — that move from chasing single numbers to reading coherent pictures — will be the ones that earn the seat at the table everyone has been talking about for the past decade.
The functions that don’t will continue producing dashboards that look healthier than the workforces underneath them, and the gap between the two will keep widening.
Goodhart’s Law isn’t a warning about measurement. It’s a warning about measuring without thinking. The discipline is to keep the measurement, and add the thinking back in.
What Better Looks Like
Picture an organization where the measurement actually matches the reality. Where signals are read continuously, not extracted in annual moments. Where leading indicators reach the right people early enough to act, not packaged into reports that arrive after the moment has passed. Where interventions are matched to what the workforce actually needs, not selected from last year’s catalog. Where proof emerges naturally from the coherence between what people feel, how they behave, and what outcomes follow — rather than being manufactured from whichever number happens to look best.
In that organization, employees are heard before they have to raise their voices. Managers see what’s happening on their teams while there’s still time to respond. HR leaders walk into board meetings with stories the numbers actually support. And the workforce, the dashboards, and the strategy all describe the same reality.
That is the workplace worth building toward. The measurement architecture is how you get there.
Workbliss is building the platform that makes this possible. Join the waitlist to be first in line.
