Here's the version of this conversation you've probably had with yourself at the kitchen table at 11 p.m. You have five or six debts. A couple of credit cards, maybe an Affirm balance, a medical bill, and (if you're being honest) a payday loan you keep rolling over. You read a Dave Ramsey post that said pay the smallest one first. Then you read a Bankrate post that said no, pay the highest interest rate first. Both posts ended with the same wishy-washy line: "pick the one that works for you."
That advice is useless. So let's actually pick.
Pure math fails real humans. Pure psychology costs real money. And neither method, in its classic form, was designed for the debt mix most people actually carry in 2026 (revolving cards plus buy-now-pay-later plus a payday loan plus a medical bill). What you need is a sequencing rule that handles all of that. I'll show you mine after I explain what the actual research says, not the watered-down version that gets repeated in every article on the topic.
The Two Classic Methods, In Plain English
The snowball method: List your debts from smallest balance to largest. Pay minimums on everything. Throw every extra dollar at the smallest balance until it's gone. Then roll that payment into the next smallest. Repeat.
The avalanche method: List your debts from highest APR (Annual Percentage Rate, the all-in yearly cost of a loan including fees) to lowest. Pay minimums on everything. Throw every extra dollar at the highest-rate debt. Repeat.
On paper, avalanche always wins on dollars. You're killing the most expensive debt first, so the interest you pay over the life of the plan is lower. By how much? Usually a few hundred dollars on a $15,000 mixed-debt load, sometimes more, occasionally less.
On paper, snowball is always more expensive. In practice, snowball finishes more often. That gap is the whole argument.
What the Research Actually Says
The study everyone cites (and almost nobody quotes correctly) is from David Gal and Blakeley McShane at Northwestern's Kellogg School, published in the Journal of Marketing Research in August 2012. They tracked consumers in a debt settlement program and found that borrowers who paid the smallest balance first were:
- 14% more likely to have eliminated their debt after one year
- 43% more likely to have eliminated their debt after four years
Compared to borrowers who paid the highest-rate debt first.
The mechanism the researchers identified is what they called the "small wins effect." Closing one account, no matter how small the balance, predicted whether someone would finish the whole plan. The psychological win of clearing a line from the list outperformed the math.
One caveat worth being honest about: the study population was people already in a debt settlement program, which is not the same as the general population. So generalize with care. But the underlying behavioral finding (early wins drive long-term completion) has been replicated across enough behavior-change research that we can trust it as a real effect, not a fluke.
Avalanche is mathematically correct. Snowball is humanly correct. If you finish, you save money. If you quit, you don't. That's the trade-off.
The Four Things Classic Methods Miss
Both snowball and avalanche assume a specific kind of debt: a small handful of installment loans and credit cards with stable balances and stable APRs. That's not the debt mix most people walk in with anymore. Four real-world problems break the classic playbook.
1. Payday loan rollovers. A payday loan has an average APR around 391% according to CFPB data, and more than 80% of payday loans get rolled over within two weeks. Every rollover is another finance charge, typically $15 to $30 per $100 borrowed. Even your worst credit card APR (call it 29%) is a discount compared to that. Snowball says "pay the smallest." Avalanche says "pay the highest APR." Both end up at the same place for once: a payday loan eats everything until it's dead. The exit plan, if you're already in one, is in how to break the payday rollover cycle.
2. BNPL stacking. Affirm, Klarna, and Afterpay split a purchase into 4 to 12 installments. Late fees run $7 to $10 per missed payment. They often don't show up in your traditional scoring picture until something breaks (Equifax began accepting BNPL data in 2022 and FICO 10T now incorporates it). The classic methods don't even know what to do with three Affirm payments running in parallel.
3. Medical debt statute of limitations. Medical bills sit on a different clock. Paid medical collections were removed from credit reports starting in 2022, unpaid medical under $500 was removed in 2023, and the CFPB's January 2025 rule to remove medical debt entirely has been tied up in litigation. So a $700 medical bill might be worth disputing first, then negotiating, before you pay a dime. It's not always in the same priority bucket as a credit card.
4. Motivation cliffs. Pure avalanche puts you on the highest-APR debt, which is often also the biggest balance. So you grind for six months and nothing visible happens. That's where 30% of avalanchers quit, switch to snowball, and finally make progress. The Kellogg numbers are basically capturing that quit rate.
The Hybrid Rule
This is the sequencing rule I give clients who walk in with a 4 to 7 debt mix. It's not an academically validated method, it's an editorial position, but it's grounded in the underlying research. Here's the rule:
- Step 1: Stop the bleeding. If you have an active payday loan, kill it first, regardless of balance. The rollover fee compounds faster than any credit card APR. Use a Payday Alternative Loan (PAL, from a federal credit union, capped at 28% APR, see our PAL guide) or an extended payment plan if your state requires lenders to offer one (16 states do). Just don't roll over again.
- Step 2: One quick win. Snowball the smallest non-payday debt. One account closed. The psychological boost the Kellogg study identified. This is the "small wins" deposit.
- Step 3: Switch to avalanche. Now sort the remaining debts by APR, highest first, and grind. You've banked the momentum. Now you collect the dollar savings.
- Step 4: Re-evaluate every 90 days. Balances shift. APRs shift (promo rates expire, cards reprice). Life shifts. Resort the list every quarter and adjust.
That's the whole framework. Three principles in trench coats: cut the worst structural debt first, get one early win, then optimize the rest with math.
Worked Example: $8,400 Across Five Debts
Let's run a real scenario. You have:
- Payday loan: $400 balance, $60 per two-week rollover (roughly 391% APR)
- Credit card A: $2,800 balance, 24% APR, $70 minimum
- Credit card B: $1,400 balance, 18% APR, $40 minimum
- Affirm payment: $600 balance, 0% APR, $150 monthly
- Medical bill: $3,200, no interest, $50 monthly negotiated
Total: $8,400. You have $400 a month above minimums to throw at debt.
Pure snowball attacks Affirm or payday first based on smallest balance. Most snowball worksheets would put Affirm first (because the payday "balance" is technically smaller, but rolls over). Let's say snowball clears them all in roughly 22 months and you pay about $1,950 in interest and fees.
Pure avalanche attacks the payday loan first (391% APR), then credit card A (24%), then card B (18%), then Affirm (0%), then medical (0%). Roughly 21 months to clear, with about $1,580 in interest. Saves you around $370 vs snowball.
The hybrid kills payday in month 1 (stop the bleeding), then snowballs the Affirm in month 2 to 4 (quick win, closes one line), then switches to avalanche on cards and medical. Roughly 20 to 21 months to clear, with about $1,620 in interest. You give up roughly $40 vs pure avalanche, but you get one closed account by month 2, which is the difference between finishing and quitting.
The hybrid isn't always the dollar winner. It's the completion winner.
When to Deviate From the Hybrid
Three situations where you should bend the rule.
Near-cap utilization. If a credit card is at 95%+ of its limit, your credit score is taking a hit beyond the dollar cost. Paying that card down to under 30% utilization first can lift your FICO 20 to 40 points in one billing cycle, which matters if you're trying to refinance other debt in the next year. The mechanics are in why did my credit score drop.
Looming statute of limitations. If you have an old debt approaching the SOL in your state (3 to 10 years depending), do not voluntarily pay it without legal advice. Paying or even acknowledging a time-barred debt can restart the SOL in some states.
Tiny balances on a closed account. A $30 balance sitting on an account that's been closed for collections? Just kill it. Don't make it part of the strategy. Cut the noise.
Tools That Actually Help
A free undebt.it account or a basic Google Sheet runs the math for any method you pick. The National Foundation for Credit Counseling (NFCC) offers free debt counseling through nonprofit member agencies if you want a human to look at your full picture. The CFPB publishes plain-language debt repayment tools and isn't selling you anything.
Your single action this week: list every debt with balance, APR, and minimum payment. Pick which step in the hybrid rule you're on. Set up the extra payment for next payday before you can talk yourself out of it.
Frequently Asked Questions
Should I pay off the payday loan or the credit card first?
Payday loan, almost always. A payday loan at 391% APR rolling over every two weeks is more expensive per dollar than any credit card you have. Kill it, then use a PAL or credit union loan to refinance any remaining balance you can't pay outright.
Does paying off small debts first hurt my credit score?
Not usually. Closing an installment loan (like Affirm or medical) doesn't hurt much. Closing a credit card can hurt by reducing your available credit and bumping your utilization ratio. The fix: pay the card down to zero but keep the account open. Cut up the physical card if temptation is the issue.
Where does BNPL fit in snowball vs avalanche?
Most BNPL loans are 0% APR if you pay on time. So in pure avalanche, they're last. But they have short terms (6 to 12 weeks), so they often finish themselves before you have to worry. The exception is Affirm's longer-term financing, which can carry 10 to 30% APR, and those should be sorted by APR like any other installment loan.
Should I close a credit card after I pay it off?
Generally no. Closing a card lowers your total available credit, which raises your utilization ratio, which lowers your score. Keep it open with a $0 balance. Exception: if the card has an annual fee you can't get waived, downgrade to a no-fee version with the same issuer rather than closing.
What about my 401k loan or auto loan: should those be in the plan?
401k loans deserve their own analysis. The interest you pay goes back to you, but if you leave your job, the balance is often due fast, with tax consequences. Auto loans are usually low APR and secured, so they go near the bottom of the avalanche list unless they're upside-down on value. Both belong in your full debt picture, just not always at the top.
Is debt consolidation better than either method?
It can be, if you qualify for a personal loan at a significantly lower APR than your weighted average. A 12% consolidation loan beats a stack of 24% credit cards. But consolidation does not solve the behavior problem. If you consolidate and keep using the original cards, you've just added a sixth debt. Treat consolidation as a tool inside the hybrid plan, not a replacement for it.
