How do you calculate the ROI of learning a skill?
We compare the total cost (course fees + your time valued at your current hourly rate) against the expected salary increase over your remaining career. The break-even point is when cumulative salary gains equal total learning cost. IRR treats learning like an investment and calculates its annualized return.
Why does the online course tool adjust for completion rates?
Most MOOC research converges on 5–15% completion. Reich & Ruipérez-Valiente's 2019 Science paper analyzed 565 HarvardX/MITx courses and found median completion of just 3.1% across the broader sample, ~6% among certificate-track learners — and even paid Coursera/Udemy data sits in single digits to low teens. A $300 course you'll likely never finish has a real expected value close to zero. The course value calculator adjusts the cost by your estimated completion probability to show the true expected ROI.
How accurate is the learning time estimator?
It's based on deliberate practice research from K. Anders Ericsson (the original 'expert performance' research that Malcolm Gladwell popularized as '10,000 hours'), tempered by Macnamara, Hambrick & Oswald's 2014 meta-analysis in Psychological Science, which found deliberate practice explains ~20-25% of skill variance — not the near-100% the popular framing suggests. The estimates are ranges, not guarantees — your actual time depends on practice quality, prior knowledge, and instructional materials. Use it as a planning tool, not a precise prediction.
Is a language really worth learning for career purposes?
It depends heavily on the language, your field, and your location. A bilingual premium in English-speaking markets is typically 5–15% for high-demand languages — Saiz & Zoido's 2005 wage premium study (Review of Economics and Statistics) and follow-up labor economics research consistently land in this range for Spanish/Mandarin/German in the US. Premiums are larger in specific niches (legal translation, healthcare interpreters, international sales). The language ROI calculator lets you adjust salary boost estimates to match your specific situation.
Should I bother learning skill X if AI might replace it in 5 years?
Two reframings help. First: the threshold isn't 'will AI do it?' but 'will AI do it without expert supervision?' — most knowledge work is decomposing into 'AI executes + human verifies'. Recent productivity studies (Brynjolfsson, Li & Raymond 2023 on customer support; Dell'Acqua et al. 2023 on consulting tasks; Peng et al. 2023 on Copilot for developers) all show ~14-40% productivity uplift but a widening gap between expert and novice users — verifying experts are paid more than the executors they replaced. Second: even if a skill becomes 50% AI-augmented, the IRR over 5 years is still positive if your salary delta exceeds the learning cost. Use the Skill ROI tool with a haircut on the salary premium (e.g. assume 50% of historical premium persists) — if it still pencils out, learn it. If the premium goes fully to zero, the answer is no.
Are paid courses really worth it vs free YouTube/MOOC?
The honest answer: the content gap is small. The completion gap is huge. Free YouTube has ~95% of the information; the paid course's value is usually structure, deadlines, and accountability — which translate to higher completion. Course Value tool adjusts the headline price by your real completion probability. A $50 free course you'll finish beats a $2,000 paid course you won't, but a $2,000 course you'll actually finish probably beats a $50 course that never gets started.
Should I learn deep in one skill or broad across many?
Career data suggests T-shaped beats either extreme: one deep specialty (high IRR via expert salaries) anchored by 2–4 adjacent skills (high optionality across role transitions). Pure specialists get displaced when their narrow market shifts; pure generalists rarely make top-of-band money. The Skill ROI calculator works for either approach — just run it once for your depth investment, then again for each adjacent skill, and look at the combined IRR.