Compare interest rates. Debt costing more than you'd reasonably expect to earn investing should be paid off first — that's a guaranteed, risk-free return equal to the interest rate, which is better than almost any investment on a risk-adjusted basis.
**The practical order:**
1. **A small emergency buffer first** — one month of expenses. Without it, the next unexpected cost goes back on the credit card and you're running in place.
2. **Employer retirement match**, if offered. An instant 50-100% return beats paying off almost any debt. Never skip this.
3. **High-interest debt aggressively** — credit cards, payday loans, anything above roughly 10%. Credit card interest in the 30-40% range is financially urgent; paying it off is equivalent to a guaranteed return no investment can match.
4. **Complete the emergency fund** to three months.
5. **Moderate-interest debt** (personal loans, car loans, roughly 7-10%) — a judgement call between this and investing.
6. **Invest**, while paying only the minimum on low-interest debt (below about 6-7% — many student loans and home loans).
**The reasoning behind the threshold**: long-run equity returns have historically been somewhere around 7-10% nominal, but with real volatility and no guarantee in any given decade. Paying off a 15% debt returns exactly 15%, guaranteed, immediately. Paying off a 4% subsidised loan while markets might return more is a reasonable gamble in the other direction.
**Two methods for the debt itself:**
- **Avalanche** — highest interest rate first. Mathematically optimal, saves the most money.
- **Snowball** — smallest balance first. Mathematically worse, psychologically better, and evidence suggests people stick with it more reliably. If you've abandoned debt plans before, the snowball's early wins are worth the extra interest.
**Factors beyond the maths**: tax treatment (some debt interest is deductible, some investment accounts are tax-advantaged), job security, and how debt makes you feel. Someone who can't sleep because of a loan may be better off clearing it even when the spreadsheet disagrees — a plan you follow beats an optimal one you abandon.