How CareerROI does the math.
Every number on this site comes from somewhere. Here's the somewhere — and what we deliberately leave out of the calculation, and why.
Last updated: January 1970 · Methodology version 1.0
The core formula.
Career ROI isn't just "salary after minus salary before." That math ignores the cost of the program, the time you weren't earning, taxes, and the probability you actually finish and land the job. Here's what we actually calculate.
- Lifetime Salary Lift
- Post-program salary minus a realistic counterfactual salary, projected over a defined horizon (typically 5 or 10 years). We discount future earnings only when explicitly noted in a calculator.
- Completion Probability
- The share of cohort members who finish the program. Pulled from CIRR or program-reported outcomes when available, otherwise flagged as unknown rather than assumed 100%.
- Program Cost
- Tuition, required fees, books, software, and exam costs. Living expenses are excluded unless the program is full-time and residential.
- Opportunity Cost
- Income forgone during the program, scaled to your actual baseline salary and program length. Part-time programs apply a partial offset.
- Tax Impact
- Federal + state marginal tax on the salary delta, not the full salary. We use a flat marginal rate per income bracket; we do not model itemized deductions.
Where the numbers come from.
BLS Occupational Employment Statistics
Federal salary data for 800+ occupations, broken down by metro area and experience level. This is our baseline for 'what does this job actually pay?' Updated annually by the Bureau of Labor Statistics.
bls.gov/oes →Levels.fyi
Self-reported total compensation for tech roles, including base, bonus, and equity. We use this for software engineering, data, and product calculations where BLS underrepresents equity comp.
levels.fyi →Glassdoor
Self-reported salary data with broader role coverage than Levels. We use this as a triangulation source — never as a single point of truth.
glassdoor.com →Program-reported outcomes
When bootcamps and programs publish CIRR-audited or self-reported outcomes data, we include it — and we tell you when we don't trust it.
cirr.org →What goes into the math.
Direct program cost
Tuition, fees, materials, and required exam costs.
Opportunity cost
Income forgone during the program, scaled to your baseline.
Time to employment
Median weeks from completion to first relevant role.
Salary lift
Post-program salary minus what you'd realistically earn without it.
Geographic cost-of-living adjustment
Salaries normalized to your metro using BLS area data.
Federal + state tax impact
Marginal tax on the salary delta, not the full income.
Completion probability
Cohort completion rates, when available.
Job-placement probability
Share placed in a relevant role within 12 months, from audited or reported outcomes.
What we don't hide from you.
Most career calculators show you the median outcome and stop there. The median is a lie if you don't see the variance. Here's what we surface that others bury:
Failure rates
Every calculator output shows the % of cohort members who don't complete the program or don't land a relevant role within 12 months. Median outcomes assume you're average. You might not be.
Variance, not just median
We show you the 25th, 50th, and 75th percentile outcomes. The gap between them is often more important than the median itself.
Scenario robustness, not a forecast
We model three scenarios — pessimistic (p25), median (p50), and optimistic (p75) — and report how many produce a positive 5-year ROI as 'X of 3'. This is a robustness check, not a probabilistic forecast. If only the optimistic scenario clears, the program's upside depends on getting lucky.
Loan interest is an upper-bound estimate
Loan interest in our calculator is modeled as simple interest accrued during the payback period only. Real student loans accrue interest over the full loan term, which means actual financed cost is higher than our calculation. We're working on a more accurate amortization model — for now, treat financed-program ROI as an upper bound.
When we don't have data
If a program won't share outcomes data and it's not in CIRR, we say so. We don't make up numbers.
What we deliberately leave out — and why.
Some things matter for career decisions but don't belong in an ROI calculator. Here's what we leave out on purpose:
- Job satisfaction
- Real, but unmeasurable in dollars. Don't let a calculator tell you whether you'll enjoy the work.
- Network effects
- The relationships you build in a program have real value, but it's too variable to model. We mention it qualitatively in reviews.
- Personal circumstance
- A bootcamp's ROI depends on your starting salary, location, family situation, and risk tolerance. We give you the inputs; you bring the context.
- ‘Passion’
- We don't model it. If a career change is right for you at negative ROI, that's a values decision, not a math decision — and you should make it with eyes open about the cost.
On bias and conflicts of interest.
We earn affiliate commissions when readers enroll in programs through our links. This is real money and it could absolutely bias us. Here's how we handle it:
- We complete the review and assign a verdict before pursuing affiliate partnerships.
- We publish negative reviews of programs we have affiliate relationships with when the math says we should.
- We disclose every affiliate relationship on every review page.
- We do not accept payment, free access, or sponsorships in exchange for coverage.
- We will publish our affiliate revenue annually starting 2027.
Methodology changelog.
When we change how we calculate something, we say so. Major methodology updates will be logged here.
v1.0 (current)
Initial methodology published January 1970.
Questions about the methodology?
Email us via the contact form — we'll answer it.