The Real Cost of Burnout—SelfWare Consulting

Some minor adjustments have been made in this version (v1.2). See what changed →

ⓘ How to use this tool — Conservative, research-anchored estimates — a floor at the default settings. Click to learn more.

How to use and interpret this tool: Burnout, stress, and disengagement are different experiences whose costs do not fully overlap; the World Health Organization defines burnout as a syndrome resulting from chronic unmanaged stress, which is why we also measure stress. At the default settings, the cost of stress and burnout shown here is a deliberately conservative floor—decreased work quality, decision errors, lost innovation, eroded confidence, and healthcare are excluded, so your true cost is almost certainly higher.

The headline figure is the annual cost of poor wellbeing, shown as a range and built from two pieces: lost productivity and burnout-driven turnover. Below the headline you will find a cost-over-time chart, a savings & ROI section, and collapsible sections detailing the costs we cannot calculate and why we believe burnout is often the best place to start to improve workplace wellbeing.

Want to tailor or pressure-test what you see? The advanced settings & detailed methodology at the bottom of the page let you adjust every assumption, data source, and reduction target in addition to providing complete transparency into the formulas and studies used to create your cost estimates. When your settings reflect your organization, use the Copy shareable link button at the top to share this page—numbers intact—with others in your organization.

Where does the range come from?

Headcount, salary, prevalence, turnover rate, and replacement cost all feed in as single values for calculating the cost of burnout. To increase accuracy and account for a greater variety of org sizes and industries, each cost model includes one deliberately-uncertain assumption that opens the band:

Productivity—AJPM lens: low end uses 0.75× and the high end 1.25× the published AJPM average cost. The band says: based on your size, industry, and pay, the relative impact may run a bit below or above the average firm in AJPM's study.
Productivity—Colonial lens: low end prices lost hours at 1.0× salary (base wages only); the high end at 1.4× salary (total comp incl. benefits + overhead). The question is how much a lost hour actually costs—payroll only, or fully-loaded. (The band widens to include your choice if you select a basis outside 1.0×–1.4×)
Turnover: the band opens over the actual-departure relative risk R, set by your costing stance—low end 1.5× (an extra-conservative buffer), high end 2.1× (Hamidi et al.'s measured two-year relative risk, undiscounted). The default point is 1.8×: Hamidi's 2.1× measures departures within two years, so it is discounted for this tool's one-year horizon. The band widens to include your override in either direction (for example SHRM's 2.8× above, or a lower value below 1.5×).

So the band moves one honest assumption per model—everything else is held fixed. At the built-in stances, even the high end prices only what published research supports; if you override an assumption beyond the built-in bands, the band widens to include your value and the estimate reflects your assumption, not the published one.

Executive summary

The annual cost of poor wellbeing in your workplace.

The ranges below update in real time as you adjust headcount, salary, and severity.

Total Employees

Average Annual Salary ($)

Workforce Severity

This field only applies to the Colonial and burnout-attributable-turnover models.

View settings at each level

Level	Burnout prevalence	Voluntary turnover	Stress hrs/wk
Lower	29%	10%	2.0
Typical (default)	44% (SHRM)	13%	3.5
Higher	59%	18%	5.0

The Typical row is the SHRM-measured 44%. Lower (29%) and Higher (59%) are symmetric ±15-point brackets around it for sensitivity testing—not separately published figures.

Costing Stance

This field only applies to the Colonial and burnout-attributable-turnover models.

View settings at each level

Level	Stress cost basis	Replacement cost	Departure risk R
Most conservative	0.75× salary	0.5× salary	1.5×
Conservative (default)	1.0× salary	1.0× salary	1.8×
Aggressive floor	2.0× salary	2.0× salary	2.1×

Lost-Productivity Measure

Two research-backed ways to price the same lost productivity—only one feeds the total at a time, never both. AJPM is peer-reviewed and role-tiered; Colonial values hours lost to stress.

SelfWare Recommendation

SelfWare's estimate for a software firm. This preset takes a more aggressive stance on the cost of burnout. Use the default (Typical + Conservative) when you need the most defensible number.

View settings at this configuration

Setting	Value	Setting	Value
Productivity lens	AJPM (whole-workforce)	Replacement cost	3× salary
Stress hours/wk	4.0	Actual-departure risk R	1.8×
Stress cost basis	2× salary	Burnout reduction	15%
Burnout prevalence	50%	Program investment	$125,000 (incl. internal labor)
Voluntary turnover rate	14%	Delay before benefit	3 months
AJPM cost multiplier	1.25×	Break-even horizon	1 year
Role mix	5% Hourly NM / 83% Salaried NM / 10% Mgr / 2% Exec

Combined estimate · Conservative floor

Total annual cost of poor workplace wellbeing

The range shown above reflects deliberate modeling-assumption bands—not statistical uncertainty. See the "Where does the range come from?" dropdown in the "How to use and interpret this tool" note above for details.

What this number leaves out

The figure above excludes major cost categories regardless of your settings. The AJPM, Colonial, and turnover models do not price team contagion, error and quality rates, slowed decisions, lost institutional knowledge, or added healthcare costs. More information on this is included in the immeasurable costs section below.

Lost productivity Burnout + disengagement

Showing: AJPM · whole-workforce lens

—

Turnover Compounding

Burnout-Driven Turnover

—

SHRM (2024) & Hamidi et al. (2018)—actual-departure relative risk · McKinsey (2022)—supporting context (stated intent)

Cost over time

The cost of burnout & disengagement over time

The chart below shows your three-year cumulative cost. The shaded band in blue is your calculated floor in today's dollars. Everything above it in red shows the realm of possibility—factoring in all of the costs that cannot be reliably estimated.

Why this chart includes an optional "compounding" slider. The real costs of poor wellbeing do not stay flat because burnout breeds burnout, departures trigger further departures, and effects like eroded quality or missed opportunities can make a significant impact on long-term growth. No published source gives a single defensible compounding rate, so this optional slider lets you model a rate you find plausible in your scenario (see the immeasurable costs section for more information).

Measured floor (real $, your range)

With compounding

Unmeasured range (actual cost > floor)

Measured floor · 3-yr (flat)

—

Today's annual range × 3

Added by compounding

—

(your assumption)

Projected 3-year cost

—

Floor + your compounding

Real annual compounding, compounded monthly (your assumption) 0%

Drag to model compounding. At 0% the dotted line sits exactly on the top of the measured band.

Show in nominal dollars (not inflation-adjusted) escalate at % / yr

Off by default—the band above is in today's dollars, inflation removed from both sides, which is why the floor is flat. Turn this on to escalate both the floor and your compounding curve at a nominal wage rate. Source: BLS Employment Cost Index—wages & salaries rose 3.4% over the year ending March 2026 (long-run trend ≈ 3%); real, inflation-adjusted wages rose ~0.1%.

The opportunity

Your savings & ROI

The money recovered each year if an intervention improves wellbeing—shown as a range using the same conservative-to-less-conservative values as the cost cards above. Because these savings are a percentage of those costs, at the built-in conservative stances, the dollars recovered shown here are also a floor. Set your target reduction in burnout and, optionally, what you plan to invest.

A note on how savings scale: productivity savings move linearly with the reduction target (a 10% reduction recovers 10% of the productivity cost). Turnover savings move slightly nonlinearly via the attributable fraction (at the default settings, a 10% prevalence cut would reduce burnout-attributable departures by ~7.6%, not 10%). The difference is by design—it is not an error.

Reduction target (%)

Applies to stress and burnout together. This is a relative reduction.

Meta-analytic evidence reports burnout drops of roughly 10–18 percentage points (an absolute measure—e.g. West et al., Lancet, 2016). Because this field is a relative reduction, the 15% default sits conservatively under the lower end of the evidence. Effects can fade within 6–12 months without ongoing support.

Program investment ($/year)

Direct spend on the intervention over a year (fees, software, materials). This includes an estimate for your internal labor cost—coordinator hours, review meetings, and evaluation work—so the payback and break-even figures reflect the full cost of your initiative.

Delay before benefit (months)

Time before results appear—including any time to create and roll out the program.

Break-even target (years)

Used to calculate the reduction you would need for savings to cover the program's cost within x years. The program is paid every year of the horizon; savings begin only after your selected delay.

Estimated annual dollars recovered

—

Break-even reduction

—

What intervention buys

The chart below shows a single year's recoverable cost set against the price of an intervention. "Recoverable" is the same figure as the estimated annual dollars recovered above—what the model estimates a program could win back if your assumed reduction holds—not a guarantee. Everything below moves live with the reduction target and program investment you set above.

Net first-year benefit

—

if your assumed reduction holds

Payback

—

if your assumed reduction holds

Return

—

if your assumed reduction holds

How each figure is calculated

Annual dollars recovered
Productivity savings + Turnover savings
The two non-overlapping models; no separate disengagement figure is added on top, so nothing is double-counted.

Productivity savings—AJPM lens
AJPM productivity cost × burnout-reduction target
AJPM prices burnout and disengagement as one bundled construct, and the two are heavily interconnected. Reducing burnout therefore recovers a significant—if not exactly proportional—slice of the whole bundled cost, which is why this term scales linearly with the reduction target. See the "why we target burnout" card below for more info. Note: because the AJPM cost applies to the whole workforce, this savings term scales with the reduction target but not with your burnout-prevalence setting; prevalence drives only the turnover savings.

Productivity savings—Colonial lens
Colonial stress cost × stress-reduction target
Used instead of AJPM when the Colonial lens is selected; never added to it.

Turnover savings
Salary × C × T × Headcount × [ AF(B) − AF(B×(1−r)) ]
where C = replacement multiple, T = voluntary turnover rate, r = reduction target, and the attributable fraction AF(B) = B(R−1) / (1 + B(R−1)). The bracket is the exact drop in burnout-attributable departures, so the nonlinearity in B is handled precisely rather than approximated. This holds the observed turnover rate fixed, which slightly understates the savings (as burnout falls, total turnover itself falls) — a deliberate conservatism.

Net first-year benefit
Dollars recovered × (months of benefit in year one ÷ 12) − Program investment
Blank until you enter an investment; months of benefit in year one = max(0, 12 − delay before benefit). The delay shrinks year-one recovery—a program that delivers sooner nets more in year one.

Payback
First month t where (Dollars recovered ÷ 12) × max(0, t − delay) ≥ Program investment × (⌊t ÷ 12⌋ + 1)
The month at which cumulative savings first cover cumulative program cost. How it is computed internally: the tool steps through time in quarter-month increments (t = 0.25, 0.50, 0.75, … up to 600 months). At each step it computes cumulative savings = (annual dollars recovered ÷ 12) × the months elapsed after the delay, and cumulative cost = the annual investment × the number of year-start payments billed so far (the ⌊t ÷ 12⌋ + 1 term—the fee is re-paid at the start of every service year). The first step where savings ≥ cost is the payback month. A program whose annual recovery does not exceed its annual fee never pays back (shown as n/a)—consistent with a return below 1×.

Return
Dollars recovered ÷ Program investment
The steady-state annual return once benefits are flowing.

Break-even reduction

Smallest r where  Productivity cost × r + Salary × C × T × Headcount × [ AF(B) − AF(B×(1−r)) ]  =  Program investment × horizon ÷ effective years

where effective years = horizon − delay ÷ 12 (the delay eats into the horizon), the program is paid every year of the horizon, and Productivity cost is the active lens's point estimate at your settings. How it is computed internally: define S(r) = annual savings at reduction r (the left side of the equation), and target = the right side. Because the productivity term is linear in r and the turnover term rises monotonically with r, S(r) only increases as r grows—so the tool solves S(r) = target by bisection: it starts with the bounds r = 0% and r = 100%, evaluates S at the midpoint, keeps the half of the interval that still contains the answer, and repeats 60 times, narrowing the interval by half each pass (final precision far beyond the 0.1% shown). The midpoint of the final interval is the break-even reduction. If even S(100%) falls short of the target, the result is shown as > 100%.

The totals above only count what is measurable from published research. But burnout and disengagement also erode quality, judgment, health, and discretionary effort — costs that are real and large, yet impossible to put a defensible per-company dollar figure on. The statistics below are not added to your total. They are here to show that the effects of poor wellbeing extend far beyond the costs shown above.

Keep in mind

These costs compound over time. Higher wellbeing produces better managers, who attract and keep stronger people, who innovate and deliver sooner. Improving wellbeing captures advantages that grow year over year. The same works in reverse: the gap between acting and not acting widens every year you wait. The figures here are only a snapshot. Consider how your own organization is impacted as these factors multiply.

When employees are burned out

63%

More likely to take a sick day — more absenteeism means more coverage gaps and lost output that rarely show up as a line item.

23%

More likely to end up in the emergency room — burnout drives real health costs that land back on the employer through claims and lost time.

13%

Less confident in their performance — hesitation and second-guessing slow decisions and reduce innovation.

50%

Less likely to discuss performance goals with a manager — development slows and misalignment compounds when people stop engaging with their own growth.

32%

Less likely to feel great responsibility for the quality of what the organization delivers — ownership erodes exactly where it is hardest to measure.

58%

Less likely to say coworkers always do what is right for customers — internal trust frays, and collaboration suffers with it.

56%

Less likely to say the organization delivers on its customer promises — customer experience and reputation take the hit.

51%

Of employees feel "used up" at the end of the day — Relationships and quality of life outside of work take damage which can cause resentment.

45%

Of employees feel "emotionally drained" by their work — emotional exhaustion drives people to mentally check out.

29%

Less likely to go above and beyond (40% of burned-out workers say they go above and beyond, vs 56% of others) — the extra effort that separates good teams from great ones quietly disappears.

Disengagement: Only partially captured in your cost estimates

The AJPM productivity lens above prices burnout and disengagement together (the study measures them as a bundled construct, dominated by presenteeism). The statistics below show the scale of disengagement at the workforce level, plus performance, quality, and safety gaps that AJPM's dollar figure does not fully capture.

23%

Engaged

Thriving at work — psychological "owners" who drive performance and innovation. (Global engagement has since edged down to 20% in 2025, its lowest since 2020.)

62%

Not engaged

"Quietly quitting" — psychologically unattached, putting in time but not energy or passion.

15%

Actively disengaged

"Loudly quitting" — resentful and acting out, undermining what engaged coworkers accomplish.

Where it says "top vs bottom quartile," the figure compares business units with the highest employee engagement against those with the lowest, from Gallup's Q12 meta-analysis. These are differences between high- and low-engagement teams — strong evidence of what is at stake, but correlational, not a direct "disengagement costs you X."

Disengagement and stress feed each other

54%

Of actively disengaged workers report a lot of daily stress, vs 41% of employees overall — and stress is the main driver of burnout, so disengagement and burnout reinforce one another.

The performance gap: most- vs least-engaged teams top vs bottom quartile

23%

Higher profitability in the most-engaged units — engagement shows up where the CFO is already looking.

18%

Higher sales productivity in the most-engaged units — engaged sellers convert more, lifting revenue without adding headcount.

14%

Higher productivity measured by production records and evaluations — the output gap shows up in hard operational metrics, not just surveys.

78%

Lower absenteeism in the most-engaged units — fewer gaps, less scramble, more continuity.

10%

Higher customer loyalty and engagement — engaged teams create customers who stay.

Quality, safety, and loss top vs bottom quartile

28%

Lower shrinkage (theft and loss) in the most-engaged units — people protect the assets they feel ownership over.

63%

Fewer safety incidents and accidents — disengagement is a safety problem, not only a morale one.

32%

Fewer quality defects — disengaged hands make more mistakes.

58%

Fewer patient-safety incidents in healthcare settings — in some industries, engagement is a matter of human safety.

The scale of it

$228–355M

Annual lost productivity from disengagement and attrition at a median-size S&P 500 company (McKinsey) — at least $1.1 billion over five years. (Includes attrition, which overlaps with the turnover model above, so it is not additive to your total.)

~$10T

Estimated global cost of low engagement in lost productivity — about 9% of global GDP. At any scale, disengaged time is paid-for time that produces little.

Burnout figures: the sick-day, emergency-room, confidence, and performance-goal figures are from Gallup's 2018 study of ~7,500 full-time U.S. employees (Employee Burnout, Part 1); the responsibility-for-quality and customer-promise figures are from Gallup, Why Americans Are Working Less; the "used up," "emotionally drained," and above-and-beyond figures are from SHRM's Employee Mental Health in 2024 series (1,405 U.S. employees). All compare employees who say they are burned out "very often / always" with those who are not. Disengagement performance, quality, and safety figures: Gallup's Q12 meta-analysis (11th edition), comparing top- vs bottom-quartile engagement business units. Stress-vs-disengagement (54% / 41%): Gallup, State of the Global Workplace 2024 — 54% of actively disengaged employees report a lot of daily stress, vs 41% of all employees globally. Tier shares (23% / 62% / 15%): Gallup, State of the Global Workplace 2024 (2023 data). The 20% engagement reading and the ~$10 trillion lost-productivity estimate (roughly 9% of global GDP) are from the 2026 report (2025 data). The $228–355M per-company figure — about $1.1 billion over five years — is from McKinsey, "Some employees are destroying value. Others are building it." (2023). Gallup's global figure counts lost productivity only and is conservative by design: it does not monetize the turnover, safety, theft, and healthcare effects shown separately above. Sources: gallup.com, shrm.org, Q12 meta-analysis, mckinsey.com.

We will not pretend burnout is your largest cost. By most available data, disengagement has the bigger price tag. We focus on burnout because it is the problem you can act on the fastest—and because the people who are burned out are genuinely unwell, not simply underperforming.

The distinction matters. Disengagement is largely a function of management quality and workplace design—role clarity, recognition, growth, feeling heard. Burnout is a distinct state of exhaustion, cynicism, and reduced capacity. They share some causes and consequences, but a well-managed person can still burn out from sustained load, and a person who is burned out can still feel their opinions count.

Burnout responds quickly. Reducing stress means changing workloads, staffing, and systems—actions that are measured in quarters. Improving engagement involves rebuilding management capability, role clarity, and culture—often measured in years. Burnout can respond to intervention in weeks to months. A single workshop can shift mindset, and relevant skills can be practiced immediately. The people suffering now should not have to wait for the slower structural work to conclude before they get relief.

It is actionable at three levels at once.

HR: Reducing burnout in the HR team makes HR more effective—and HR delivers most of the other interventions that help with stress and disengagement, so the effect compounds across everything they touch.
Managers: The Gallup Q12 behaviors that drive engagement—recognition, development, progress conversations—are largely manager-led, and a burned-out manager cannot deliver them well, so easing manager burnout improves burnout and engagement on their teams.
Individuals: Many of the factors that can reduce burnout—mindset shifts, understanding how it works, energy management—are things people can apply on their own, without waiting for organizational change.

It fills a gap your current spend leaves open. Most wellness and engagement efforts improve the conditions around a person—the workplace, the management, the perks. Few teach the person how to recognize burnout and manage their own energy. That is one of the reasons organizations can invest heavily in engagement and still watch capable people deplete. The missing piece is rarely the environment alone—it also includes the individual's skill in protecting and restoring their energy. This is new ground, not a duplicate line item.

People actually want this help. Re-engaging a deliberately disengaged employee can feel like asking them to give more to an organization they have already decided does not deserve it—so some resist. Burnout is different. Nobody wants to be burned out, and the people who are tend to be the ones who cared most and worked hardest to get there. Being told "you are disengaged" can land as an accusation about commitment. "You are burned out" carries implicit recognition of effort. That asymmetry means burnout programs may see more voluntary participation and faster individual change.

The honest framing: this does not replace the structural work your organization needs to do on management, workloads, and engagement—it complements it by addressing the layer where individuals and managers can change things now. We do not claim burnout is the biggest cost. We focus on it because it delivers relief and return on a timeline that does not require waiting years—and because teaching people to manage their own energy is a skill that keeps paying off long after any single intervention ends.

Advanced settings & model verification

The exact dollar figures for your specific inputs are shown in the cost cards near the top of the page, the cost-over-time chart, and the Savings & ROI chart. The controls below let you adjust and pressure-test every assumption behind those figures. Each model shows its formula so you can trace any number back to its inputs.

Every figure in this tool traces back to a published source. Here is what each study measures, how it was conducted, where it is strong, and where it is limited.

Productivity—lost output

AJPM (Martinez et al., 2025) is the primary productivity number: a peer-reviewed, role-tiered estimate of the annual cost of lost productivity per employee, published in the American Journal of Preventive Medicine. How it was conducted: rather than a survey of a fixed number of people, it is a computational simulation model built by the PHICOR team at the CUNY Graduate School of Public Health (with Baruch College, Johns Hopkins, and the University of San Diego), parameterized from national U.S. data on wages, burnout prevalence, absenteeism, presenteeism, and turnover. The model outputs a per-employee cost by role tier ($3,999 hourly nonmanager, $4,257 salaried nonmanager, $10,824 manager, $20,683 executive) and ~$5.04M per 1,000-employee company. About 89% of the cost is presenteeism (people at work but impaired) and roughly 11% is absenteeism, which is why presenteeism dominates the headline. Limitation: it is a modeled population average, so we cannot adjust for scenarios with lesser or greater personal wellbeing.

Why does the calculator show $5,038.92, not $5,040? The tool doesn't use the study's headline per-employee average directly—it rebuilds it as a weighted average of the four published role-tier costs ($3,999 / $4,257 / $10,824 / $20,683) times your role-mix sliders, so the total updates correctly when you adjust the mix. At the study's own default mix (59.7% / 28.6% / 10% / 1.7%), that reconstruction lands at $5,038.92, about $1.08 off the paper's $5,040 headline. This isn't a slider or calculation error: the paper rounds the headline average and the four tier costs/mix percentages independently, so multiplying the already-rounded tier costs by the already-rounded mix can't perfectly reproduce a number the authors derived from unrounded underlying data.

Colonial Life (2019) is the cross-check on lost productivity. How it was conducted: a survey of 1,506 U.S. adults working full-time, fielded by the research firm Dynata on behalf of Colonial Life between January 29 and February 1, 2019. It asked how many hours per week people spend at work thinking about their stressors: 28% under 1 hr, 50% 1–5 hrs, 16% 5–10 hrs, and 6% over 10 hrs (weighted mean ≈ 3.5 hrs/wk). It also asked what those stressors are: 29% job, 24% finances, 17% their own health, 9% the health of a spouse/partner/children, 8% family, and 5% the health of an elderly family member. Note that most of these stressors originate outside of work, yet they still play out on the clock—the hours are lost at work regardless of where the stress comes from. A principles-based program is well suited to this because it helps people make changes across both their work and their personal lives, reducing and better managing stress from every source rather than only the work-related share. Limitation: self-reported survey hours, and the data predates the pandemic, so today's figure is likely higher.

Turnover—burnout-driven departures

SHRM (2024) supplies the default burnout prevalence (44%). How it was conducted: SHRM's Employee Mental Health in 2024 research series surveyed 1,405 U.S. employees; 44% reported feeling burned out, and burned-out workers were nearly three times more likely to be actively searching for another job (45% vs 16% of those not burned out). Hamidi et al. (BMC Health Services Research, 2018) supplies the actual-departure relative risk. How it was conducted: the team used de-identified records from 472 physicians who completed a 2013 wellness survey at two Stanford-affiliated hospitals, then tracked who actually left over the following two years (turnover compiled in 2015). This produced a measured 2.1 relative risk (odds ratio 2.68) of leaving, based on people who actually left rather than people who said they might. Why the default is 1.8, not 2.1: Hamidi's 2.1 measures the risk of leaving within two years, while this tool prices annual (within one year) voluntary departures. How that risk is distributed across the two years was not measured, so the one-year figure is genuinely uncertain. Rather than guess in the model's favor, we apply an additional conservatism margin and discount the measured 2.1 to a 1.8 default; the aggressive stance applies Hamidi's 2.1 undiscounted, and the most-conservative floor is 1.5. SHRM measures a burned-out population actively looking for new employment—more serious than mere intent, but still not 1:1 with actual departures—so even though the SHRM data is much more recent, its 2.8 ratio is kept as an upper cross-check (reachable as a manual override) rather than the default. Limitation: Hamidi studied physicians (statistics for other industries could differ) and while the study is a bit newer the data is over 10 years old.

Disengagement—shown for context, excluded from the total

Gallup and the McKinsey Health Institute quantify the cost of disengagement and the productivity-plus-attrition drag of distressed employees. How they were conducted: Gallup's engagement figures come from its Q12 meta-analysis pooling thousands of business units across many companies (comparing top- vs bottom-quartile engagement), alongside its large global workforce survey; McKinsey's figures come from large multi-country employee surveys. These figures bundle productivity loss and attrition, so they are kept as a labelled reference card, explicitly outside the headline total.

Intervention effectiveness—how much you can actually move

West et al. (Lancet, 2016) and Panagioti et al. (JAMA Internal Medicine, 2017) find that interventions cut burnout prevalence by roughly 10–18 percentage points (an absolute measure; West's pooled estimate is about 54%→44%, ≈18% in relative terms), and that organization-directed and blended programs outperform individual-only approaches. Because this tool's reduction target is expressed in relative terms, its 15% default sits just under the relative equivalent of that evidence—conservative by design. How they were conducted: both are systematic reviews and meta-analyses of controlled trials in physicians—West screened 2,617 articles and pooled 15 randomized trials covering 716 physicians (plus cohort studies); Panagioti pooled controlled intervention trials and is where the organization-directed advantage is clearest. A 2023 meta-analysis in the American Journal of Medicine tempers this: the measured improvements are real but may not always be clinically meaningful, and effects can fade without ongoing support.

Sources

This is the AJPM lens of the productivity lost card: the per-employee annual cost of burnout by role tier (Martinez et al., American Journal of Preventive Medicine, 2025). AJPM measures burnout and disengagement as a bundled construct, so the figure already covers both. Following AJPM's published method, the cost applies to your whole workforce—their headline figure ($5.04M per 1,000 employees) is not multiplied by a burnout-prevalence rate. The cost multiplier lets a company that pays more, or earns more per employee, scale the published average to its own economics.

AJPM Cost Multiplier (×)

The AJPM study uses an average across many companies. If your firm pays above market, or earns more per employee (so lost output is worth more), set this above 1.0; below 1.0 if the reverse. The card's range spans 0.75×–1.25×, but will widen to include your value. SelfWare's software-company default is 1.25×.

Role Mix (must total 100%)

Hourly ($3,999) 59.7%

Salaried ($4,257) 28.6%

Managers ($10,824) 10%

Executives ($20,683) 1.7%

Your weighted role cost: — vs. $5,040/employee—AJPM published average

The role mix is an anchored refinement around AJPM's published role-tier costs. To scale the whole figure to a company that pays more or earns more per employee, use the AJPM cost multiplier above—that is the primary "scale to your org" lever, not the role mix.

Formula Annual burnout cost = Headcount × Weighted role cost × Cost multiplier Weighted role cost = (%Hourly-NM × $3,999) + (%Salaried-NM × $4,257) + (%Manager × $10,824) + (%Exec × $20,683), from AJPM's four role-tier figures (Martinez et al., 2025). No burnout-prevalence term—the cost applies to 100% of the workforce. The default mix is AJPM's published U.S. distribution (59.7% hourly NM / 28.6% salaried NM / 10% managers / 1.7% executives), which matches the study's published inputs and lands within $2 of its published $5,040/employee average ($5,038.92 weighted). Manager and executive weights move the total far more than the hourly/salaried split.

This is the Colonial lens of the productivity card: a measure of presenteeism based on hours spent at work thinking about stressors. The AJPM lens above prices the same presenteeism cost a different way, so only one feeds the total at a time. The default 3.5 hrs/wk is the weighted mean of Colonial Life's four published time buckets (n=1,506; derivation in Studies & Methodology).

Two things shape how to read it. First, these are hours lost to stress of any origin—work or personal—that the employer already pays for; a principles-based program reduces stress across both domains. Second, the figure is an average across every employee, not just the burned-out—so the reduction target applies to the whole workforce too. If burnout is severe in your org, a single case can lose 500+ hours/year (10+ hrs/wk); raise the hrs/wk figure or use the Higher Severity preset (5.0) to account for it.

Avg Hours Lost / Employee / Week

Default: 3.5 hrs/wk—the weighted mean from Colonial Life's four published buckets. Severity preset moves this ±1.5 hrs.

Working Days / Year

Used only in this model to set working weeks (days ÷ 5). The hourly rate is set separately on the standard 2,080 paid hours/yr. Default 230 ≈ 46 weeks.

Cost Basis

An employee's value to the company is higher than their salary. 0.75× is the most-conservative basis (discounting survey overestimation and partial productivity); 1× is the default payroll floor; 1.4× is HR-loaded cost; 2× reflects productive surplus. The card's range spans 1×–1.4× but widens to include your choice if you pick a value outside it. Two deliberately different year bases: exposure uses your actual working weeks (fewer weeks = fewer hours lost), while the hourly rate spreads salary across all 2,080 paid hours (the CFO-standard rate, which is lower than a worked-hours rate). Both choices reduce the estimate — conservative twice, on purpose.

Formula Annual stress cost = Headcount × (Hours/week × Working weeks) × Hourly rate Working weeks = Working days ÷ 5. Hourly rate = (Salary × Cost basis) ÷ 2,080 paid hours/yr. Cost basis ranges 1×–1.4× salary for the card range.

A separate cost—added on top of the productivity lens: what you spend replacing people who quit because they were burned out.

Burnout Prevalence (B) %

The share of your workforce that is burned out. Drives the burnout-attributable departures here and the burnout-reduction savings. The 44% default comes from SHRM 2024 data.

Voluntary Turnover Rate (%)

Your org's voluntary annual turnover (people who choose to leave)—not layoffs or RIFs. The attributable-fraction logic models burnout-driven choices to leave. U.S. voluntary avg ≈ 13%. Part of the Severity preset.

Replacement Cost (× salary)

The cost to replace one departing employee, as a multiple of salary. Enter your own figure if you have calculated it. For reference: estimates commonly run 0.5–2× salary; the default is 1× (recruiting and onboarding), the most-conservative floor is 0.5×, and the aggressive stance is 2×. It is reasonable to assume that senior or specialized roles could reach 3–4× or more once lost domain knowledge and ramp-up are counted. Part of the Costing Stance preset.

Departure Relative Risk (R)

How much more likely a burned-out employee is to actually leave within a year. R is set by your costing stance (most conservative 1.5×, default 1.8×, aggressive 2.1×). 2.1× is Hamidi's measured relative risk of leaving within two years; the 1.8× default discounts it because this tool prices annual departures, as an added conservatism margin. The card's range spans 1.5×–2.1×, widening to include your override in either direction (for example SHRM's 2.8× above, or a lower value below 1.5×). Editing R puts the costing stance into Custom.

How R is set & why

Anchored on actual departures, not just intent. R is grounded on actual job-seeking and departure data—Hamidi et al. (BMC Health Services Research, 2018: relative risk 2.1, odds ratio 2.68; link)—cross-validated by SHRM's actively-searching ratio of 2.8 (45% vs 16%). Hamidi measured the risk of leaving within two years; this tool prices departures within one year, so the default working point is discounted to 1.8, as an added conservatism margin. The costing-stance band runs from 1.5 (an extra-conservative floor below the annualized default) to 2.1 (Hamidi's measured two-year value applied undiscounted); 2.8 itself is reachable as a manual override.

Why not a higher ratio? Stated-intent ratios run far higher—intent-attributable risk is roughly 3.7× actual turnover-attributable risk (Hamidi). For context, McKinsey (2022) found burned-out employees 6× more likely to intend to leave within 3–6 months—a stated-intent measure. R is deliberately anchored on actual departures instead, so the model is not inflated by what people merely say.

Disclaimer: the Hamidi data comes from health care and may not transfer exactly to other industries, but coupled with the SHRM figure and a conservative mindset it remains a valuable benchmark for reasonable estimation.

Formula Annual turnover cost = Salary × Replacement multiple × Turnover rate × AF × Headcount AF (attributable fraction) = B(R−1) ÷ (1 + B(R−1))—the slice of all departures attributable to burnout. B = burnout prevalence, R = actual-departure relative risk (default 1.8×, band 1.5×–2.1×, or your override). Replacement multiple is the ×-salary cost to replace one person.

Changelog — what has changed since v1

Version 1.2

A follow-up audit pass. No figures change at any built-in configuration; the fixes below affect edge cases and copy accuracy.

Savings range corrected for large R overrides. The recoverable-dollars range could display backwards (low above high) when the departure-risk override was set very high (roughly 8×–14× and above). The range now always brackets correctly. Built-in stances (1.5×–2.1×) were never affected.
Shared links now preserve a custom-stance R exactly. Previously a link saved from a hand-tuned costing stance could reopen with the departure risk reset to the 1.8× default.
The software-preset settings table now shows R = 1.8×, matching what the preset applies (the v1.1 change had updated the preset but not its table).
Print/PDF export now includes the intro note and this changelog.
Wording tightened where copy overstated the code: the "conservative floor" claims in the intro and ROI section are now scoped to the default settings, the role-mix rationale no longer claims to reproduce $5,040 "exactly," and a few stale references (heatmaps, hours-per-year arithmetic) were corrected.

Version 1.1

A small round of accuracy and transparency refinements. The overall estimate is essentially unchanged at the default settings — the only figure that moves is the AJPM productivity model, and only by about a dollar per employee (see the last note).

Default role mix now matches AJPM's published U.S. distribution (59.7% hourly / 28.6% salaried / 10% managers / 1.7% executives). The sliders now allow one decimal of precision so the tool now uses the study's own published inputs rather than a calibrated approximation of them ($5,038.92 weighted, within $2 of the published $5,040/employee average).
The "Conservative floor" label is now dynamic. It reads "Conservative floor" only when the costing levers (stance, cost multiplier, departure risk R, replacement cost) are at or below the conservative defaults; severity describes your workforce rather than the costing stance, so it does not affect the label. If you dial the costing assumptions up, it reads "At your chosen stance" so an aggressive configuration is never mislabeled as a floor.
Colonial cost-basis range now widens in both directions. Previously, selecting the most-conservative 0.75× basis could push the point estimate below the displayed range; the range now always contains the point.
Turnover risk (R) can now be set as low as 1.0×. Setting R = 1 (burnout does not raise departure risk) now correctly zeroes the turnover model, mirroring how a 0× replacement cost already worked.
Prevalence and turnover-rate fields now accept 0. A typed 0 is honored as "0%," matching the underlying math.
Role-mix sliders that don't total 100% are now normalized so the weighted cost stays a true per-employee average (the on-screen warning still prompts you to fix them).
SelfWare software-company preset now uses the default one-year departure-risk R of 1.8× (was 2.1×), aligning it with the tool's conservative stance.
Methodology wording clarified in several places: the turnover band widens in either direction, the AJPM savings term scales with the reduction target but not with prevalence, the R discount is framed as an added conservatism margin, and the Colonial model's two-year-base choice is disclosed as deliberately conservative.

Effect on the headline at default settings (500 employees, $80,000, Typical + Conservative, AJPM lens): the AJPM productivity point moves from $2,519,905 to $2,519,458 (−$447 total, −$0.89/employee). The turnover model, ROI, and payroll-percentage figures are unchanged.

From SelfWare Consulting

Your workforce is not a spreadsheet.

You have just pressure-tested the numbers. The next step is a conversation about your situation: where burnout exists, what is driving it, and the fastest path to solutions that last. We are happy to give you honest, actionable advice regardless of whether we work together in the future.

Book a Free Consultation