What the data actually shows
Burnout is the most studied of the warning signs, and it has a specific shape. The World Health Organization classifies burnout as an occupational phenomenon defined by exhaustion, cynicism or mental distance from the job, and reduced effectiveness — drawing on the framework developed by Maslach and colleagues. When that state is chronic and the conditions producing it are unlikely to change, it is one of the clearest evidence-backed signals that the current situation is not sustainable as it stands.
Values and person-environment misalignment is another well-supported signal. Person-environment fit research (for example, Kristof-Brown and colleagues' 2005 meta-analysis) consistently links poor fit between a person's values, needs, and abilities and their work to lower satisfaction and higher intention to leave. A persistent sense that the work conflicts with what matters to you is not just a mood; it is the kind of mismatch the research connects to genuine, lasting dissatisfaction.
Work also has measurable effects on health, which is why health signals deserve weight. Research on job strain — the combination of high demands and low control described in Karasek's demand-control model — links chronically stressful work to worse mental and physical health outcomes. When a job is producing sustained effects on sleep, mood, or health that persist beyond ordinary busy periods, that is real data about the cost of staying, not a sign of weakness.
Why this feels different from how it actually is
Leaving feels harder than the situation may warrant because of sunk-cost thinking: the years, effort, and relationships you have invested feel like reasons to stay, even though they are already spent and cannot be recovered by staying longer. The mind treats past investment as a reason to continue, which is precisely the error the sunk-cost fallacy describes.
Status-quo bias compounds it. People systematically overweight the certainty of their current situation against the uncertainty of a change, even when the current situation is poor. The familiar bad is easy to picture; the unfamiliar better is not, so staying feels safer than the evidence often justifies — and inertia quietly wins by default.
And the decision is clouded by a genuine, two-sided uncertainty. The grass-is-greener pull can make any other option look brighter than it would prove to be, while fear of the unknown can make the current job look more tolerable than it is. Both distortions operate at once, which is part of why the timing of leaving feels so impossible to read from the inside.
The honest goal is not to find the perfect moment to leave but to make sure the decision is driven by an accurate reading of your situation rather than by the biases that distort it.
What the research says to do about it
Weigh the signals together and over time rather than reacting to a single bad week. Chronic burnout that rest does not relieve, a durable values mismatch, no growth or path to it, and sustained health effects are the patterns the research connects to genuine unsustainability. One signal is information; several, persisting across months, is a stronger case than any one moment of frustration.
Separate the fixable from the fundamental before deciding. Some of what drives the urge to leave is changeable inside the current role — a manager, a team, a workload, a remit — and job-design research suggests reshaping aspects of your work can meaningfully shift satisfaction. Testing whether the core problem can be addressed where you are gives you a cleaner read on whether leaving is actually necessary.
Name the biases explicitly when you decide. Ask what you would advise someone else in your exact position, which sidesteps sunk cost by removing your personal investment from the frame, and treat the years already spent as gone rather than as a reason to stay. The aim is a decision driven by an accurate reading of the present and future, not by the past or by inertia.
What the research says does not help
Waiting for certainty does not help, because it rarely arrives — there is no formula that will tell you unambiguously that now is the moment. Holding out for a clear sign often just means staying by default while status-quo bias makes the wait feel prudent rather than passive.
Staying because of what you have already put in is the sunk-cost trap, and it does not improve the decision. Past investment is spent regardless of what you choose next; the only thing that should drive the call is whether staying or leaving serves you from here forward.
Leaving purely to escape, without examining whether the core problem is portable, often disappoints. The grass-is-greener pull can make any alternative look brighter than it would prove, and if the real issue is a values mismatch or a pattern that follows you, a new job alone may not resolve it. Escape is a reason to investigate, not a guarantee of relief.
One signal is information; several, persisting across months, is a stronger case than any one moment of frustration.
What this looks like in real life
"I've already put in so many years"
The time, effort, and relationships invested feel like reasons to stay — but they are already spent and cannot be recovered by staying longer. That is the sunk-cost fallacy at work. The only thing that should drive the call is whether staying or leaving serves you from here forward, not what the past cost.
Testing whether the problem is portable
Someone dreads Mondays and assumes the answer is a new job. Before leaping, they separate the fixable from the fundamental: is it the manager, the workload, or the remit — things that might be reshaped where they are — or a values mismatch that would follow them anywhere?
Job-design research suggests reshaping aspects of a role can meaningfully shift satisfaction. Testing that first gives a cleaner read on whether leaving is actually necessary, rather than leaving purely to escape and risking the same disappointment elsewhere.
Real numbers in context
There is no statistic that tells you when to leave, and that is the honest first point — the signals are qualitative patterns to weigh, not a number to clear. What the data does establish is that the warning signs are real and measurable: burnout is a recognised occupational phenomenon (WHO), poor person-environment fit reliably predicts intention to leave (Kristof-Brown et al., 2005), and job strain is linked to worse health (Karasek's demand-control model).
It is also worth knowing that staying too long is a common, documented pattern rather than a personal flaw. Sunk-cost and status-quo biases are among the most replicated findings in decision research, which is why so many people remain in jobs past the point that serves them. Recognising the pull of those biases is part of reading your own situation accurately.