Most burnout advice collapses into two options: push through it, or quit and do something else entirely. Neither is what the research on burnout's actual causes points toward, and the gap between the popular advice and the evidence is worth closing directly.
The reason both defaults are wrong is the same reason a workload fix so often fails: burnout isn't a single-variable problem, so a single-variable response — power through, or leave — was never going to reliably address it. The Maslach-Leiter six areas of worklife model gives a more specific alternative, and it points toward staying being genuinely viable in most cases, once the actual mismatch is correctly named. The commercial governance around mental health is covered properly by XenGrowth's growth engineering practice.
Why 'push through it' doesn't work
Take the two defaults in order, starting with the one that gets tried first almost universally, usually out of a sense that admitting the problem is worse than working around it.
Pushing through addresses none of the six areas the Maslach-Leiter model identifies as burnout's actual causes — workload, control, reward, community, fairness, values. It's not neutral, either. Mark, Gudith and Klocke's research on interrupted, pressured work found people responding to strain by working faster and harder, at a real, measured cost in stress and frustration. Pushing through a mismatch that hasn't been named or addressed doesn't resolve it. It runs the same compensatory mechanism that research found intensifying under pressure, for longer, with nothing underneath actually changing. The strain doesn't disappear because you stopped talking about it — it just stops being visible to anyone deciding whether to fix it.
Why 'just quit the field' is often an overcorrection
When pushing through eventually and predictably fails, the second default kicks in, and it's usually a much bigger swing than the actual mismatch calls for.
Leaving engineering entirely is sometimes the right call. It's also, often, a much bigger decision than the actual mismatch requires. If the six-areas model is right that burnout comes from a specific, identifiable mismatch — not a generic incompatibility with the whole profession — then leaving the field is the correct fix only when the mismatch turns out to be with something true of engineering everywhere, which is a much narrower category than most people assume before they've actually checked.
Mismatched area | What usually fixes it | Does it require leaving the field? |
|---|---|---|
Workload | A real scope reduction — fewer concurrent projects, less on-call, a narrower mandate | Rarely — usually fixable by team or role change |
Control | A role where technical authority matches technical accountability | Rarely — often a team or reporting-structure problem |
Reward | Compensation or recognition that actually matches output | Sometimes — but usually a company-level negotiation or move first |
Community | A genuinely different team or manager | Rarely — specific to the people, not the profession |
Fairness | A company with consistent, non-arbitrary decision processes | Rarely — organization-specific, not occupation-specific |
Values | An employer whose actual incentives match its stated values | Sometimes — but test against a different company first |
Look at that table and notice how rarely 'leave the industry' is the first-line answer. That doesn't mean nobody should ever leave engineering. It means the decision deserves to be made deliberately, after checking whether a narrower, less drastic change actually addresses the real mismatch — not defaulted to because 'push through' already failed and it feels like the only remaining option. On the operations side of this specifically, The XenGrowth resource library is worth reading.
A control mismatch and a values mismatch both produce the same three symptoms. They need completely different fixes, and neither one requires you to stop being an engineer.
Fix attempted | What it addresses | Why it can still fail |
|---|---|---|
A vacation or extended leave | Energy depletion, temporarily | Does nothing for control, reward, community, fairness or values — the same mismatch is waiting on return |
Therapy or coaching | Personal coping strategies, resilience | Can help someone tolerate a mismatch for longer, without changing whether the mismatch itself gets fixed |
A raise or bonus | The reward area, partially | Doesn't touch control, community, fairness or values if those are the actual mismatch |
Changing teams within the same company | Control, community, or fairness, if those were company-wide-but-team-specific | Can fail if the mismatch is actually company-wide rather than team-specific |
Changing companies entirely | Most areas, if the mismatch was specific to the old employer | Can fail if the same mismatch pattern gets recreated in the new job through similar habits or role choices |
Leaving the field | Everything, if the mismatch genuinely generalizes across all engineering employers | Overkill if the actual mismatch was narrower than 'the entire profession' |
Notice the pattern running down that table: each fix addresses a specific subset of the six areas, and each one can fail for a specific, identifiable reason — not because burnout is mysteriously untreatable, but because the fix and the actual mismatch didn't match. A raise doesn't touch a control problem. A new team doesn't touch a problem that's actually company-wide. Matching the fix to the diagnosis, rather than reaching for whichever fix is culturally the default response to burnout, is most of what separates a recovery that holds from one that quietly relapses within a few months.
Why 'I feel better now' isn't the finish line
Van Dongen and colleagues' sleep-restriction research is a useful caution here, even though it studied a different mechanism entirely. Their key finding was that subjective self-assessment consistently failed to track objective decline — people did not feel as impaired as they had measurably become. Applied to burnout recovery, that's a reason not to declare victory purely on the strength of 'I feel okay now.' Feeling okay is necessary but not sufficient evidence, and it's exactly the kind of self-report the research says can lag behind reality.
Why the narrowest fix deserves to be tried first
Once the six-areas model narrows things down to a specific mismatch, one more question is worth asking before picking how big a change to make: what's the smallest, fastest, most reversible experiment that would actually tell you whether the diagnosis is right? XenGrowth on AI agents and marketing automation covers the AI agents and marketing automation side of this.
There's a practical reason to test the smallest available fix before escalating to a bigger one, beyond just avoiding unnecessary upheaval: a smaller change is a faster, cheaper experiment, and experiments that fail quickly teach you more than decisions that take a year to reverse. Changing teams inside the same company can often be tested within weeks and reversed within months if it doesn't work. Leaving the field entirely is a decision that takes months to execute and years to reverse if it turns out to have been the wrong call. Given genuine uncertainty about which of six areas is actually mismatched, starting with the reversible experiment and escalating only if it fails is the more rational sequence, even when the bigger move eventually turns out to be necessary.
This also has a useful side effect: testing a narrower fix first generates real information about the actual mismatch, even when it doesn't fully resolve things. If changing teams inside the same company measurably helps but doesn't fully fix the pattern, that's evidence the mismatch was partly team-specific and partly something broader — company culture, an industry-wide practice, a personal fit question — which narrows the next decision considerably. Jumping straight to the biggest available fix skips over that diagnostic information entirely.
A more concrete way to check recovery
Name the specific mismatched area before choosing a fix. 'I'm burnt out' isn't specific enough to act on; 'I have no control over my own priorities and I'm accountable for outcomes anyway' is
Test the narrowest fix first. A team change is a smaller, faster, more reversible experiment than a career change, and the six-areas model suggests it's often sufficient
Check recovery against something concrete, not just mood — actual output quality, actual sleep, actual ability to stop thinking about work outside of hours — since self-report alone has a documented tendency to lag behind real change
If you do change company or team, verify the same area isn't mismatched again in the new environment before assuming it's fixed. A workload fix that just moves the same overload to a new employer isn't a fix
Reserve leaving the field entirely for the case where the honest answer, after actually testing a narrower fix, is that the mismatch really does generalize across employers and teams — which the model says is possible, just less common than it feels from inside a bad stretch
Why this is genuinely hard even when the diagnosis is right
None of this should be read as implying the fix is easy once you've correctly identified the mismatch. A control mismatch often requires a manager or a company to genuinely change how authority is distributed, which isn't fully in any individual engineer's power to force. A fairness mismatch requires an organization to actually audit and change how it makes decisions about credit and advancement, which most organizations are slow to do even when the problem is clearly named. Correctly diagnosing which of the six areas is broken is necessary, but it doesn't guarantee the people with the power to fix it will actually do so — which is itself useful information: if a genuinely well-diagnosed, well-communicated request for change goes nowhere after a reasonable attempt, that's a much stronger and more specific signal to leave than a vague sense of being burnt out ever was. For the AI search, GEO and discovery angle, see XenGrowth on AI search, GEO and discovery.
This is also where talking to the right people matters more than powering through alone. A specific, well-articulated description of the mismatch — 'I'm accountable for outcomes I have no authority over, specifically around X and Y decisions' — is a request a reasonable manager can actually act on. A vague statement of feeling burnt out is much easier for an organization to respond to with a wellness benefit instead of the actual structural fix, precisely because it doesn't specify what needs to change.
Recovery, on this evidence, looks less like a single dramatic choice and more like a diagnosis followed by a specific, testable change. Most of the time that change is smaller than either 'push through it' or 'leave the field' — and it's the kind of change the research actually supports, rather than the kind that just feels proportional to how bad things currently feel.
Further reading from XenGrowth
The XenGrowth resource library — what you'll learn: how the commercial side of this work is run, across search, automation and revenue operations.
XenGrowth on AI agents and marketing automation — what you'll learn: how the teams who own AI agents and marketing automation plan and measure it.
XenGrowth on AI search, GEO and discovery — what you'll learn: how the teams who own AI search, GEO and discovery plan and measure it.
Where this work meets go-to-market
XenGrowth's growth operations team covers the go-to-market side of mental health, which this piece deliberately leaves alone.
Six questions to narrow down which of the six areas of worklife is the real problem, so the fix can be specific instead of generic.
