People treat their credit score like a grade in a class they never attended. They know it matters, they know it can be improved, and almost nobody can name what it’s actually made of — which is why “raise your score fast” advice is full of rituals that do nothing and misses the two inputs that do almost everything.
Here’s the thing: your score comes from the information in your credit reports, and the main scoring models weight five categories. Learn the five and you can predict your own score’s behavior — and more importantly, you can stop optimizing the wrong things. Quick framing: this is education, not financial advice, and the exact formulas are proprietary — but the public weightings below are well documented and stable enough to build a plan around.
The Five Inputs, Ranked by Actual Weight
| Input | Approximate weight | What it measures |
|---|---|---|
| 1. Payment history | ~35% | On-time vs late payments, collections, public records |
| 2. Amounts owed (utilization) | ~30% | Revolving balances vs limits; how much of your available credit you use |
| 3. Length of credit history | ~15% | Age of oldest account, average age, recency |
| 4. New credit | ~10% | Hard inquiries, recently opened accounts |
| 5. Credit mix | ~10% | Revolving vs installment accounts |
Read the first two rows again: roughly 65% of your score is payment history plus utilization. Everything the forums obsess over — the exact inquiry timing, the mix of loan types, the “credit-building hacks” — lives in the remaining third. If you only ever manage two things, manage those two.
Input 1: Payment History (35%)
This one is brutally simple and nearly impossible to fix after the fact: pay on time, every time, and the biggest input takes care of itself. Miss by 30+ days and the late payment can be reported and linger on your report for up to seven years, with the most damage in the first two.
The practical system isn’t willpower, it’s plumbing: autopay the full statement balance on every card (or at minimum the minimum), and keep one calendar reminder a few days before due dates as a backstop. A single forgotten $30 minimum payment is the most expensive sloppiness in personal finance measured in points per dollar.
One honest nuance: being a few days late before the 30-day mark usually just triggers a late fee, not a credit report entry. But autopay exists, so there’s no version of this where manual memory beats automation.
Input 2: Utilization (30%)
Utilization is the ratio of your reported card balances to your credit limits, and it’s the input with the weird superpower: it has no memory. Payment history is a seven-year file; utilization is basically a snapshot. The score this month cares mostly about the balances reported this month.
That’s the closest thing to a cheat code in credit scoring. Someone at 60% utilization can pay balances down (or down below ~10% before the statement closes) and see meaningful score movement within one or two reporting cycles — no waiting, no history required. The mechanics: card issuers typically report your statement balance once a month, so the number that matters is the balance on your statement date, not what you owe after paying it. Paying in full after the statement prints still leaves a high reported balance.
Practical moves that actually move this input:
- Pay balances down before the statement closes, not just by the due date.
- Spread spending across cards if one is carrying most of it (per-card utilization can matter alongside the overall ratio).
- Don’t close old no-fee cards — closing removes their limit from the denominator and your utilization jumps.
- Ask for credit limit increases on cards you already hold (a soft-pull at many issuers) — a bigger denominator does the same job as a smaller numerator. Don’t spend up to the new limit; the whole trick is the ratio.
Input 3: Length of History (15%)
This one is almost purely time. The levers: keep your oldest accounts open (that closure mistake hurts twice — utilization and age), and be slow about opening new accounts if you’re planning a mortgage in the next year or two. There is no fast path here, which is exactly why the score rewards people who just don’t churn their credit life.
Inputs 4 and 5: New Credit and Mix (10% each)
Hard inquiries from applications ding the score modestly and fade within months; they matter mostly in clusters. If you’re applying for a mortgage soon, avoid opening anything new during the months before — underwriters read recent applications as risk signals. Credit mix — having both revolving and installment accounts — is a small, mostly-passive factor. Don’t take out a personal loan to “improve your mix”; that’s paying interest for a rounding error.
The 90-Day Improvement Plan
If your score needs work and you have no derogs (collections, late payments) on file, here’s the sequence I’d run — it’s mostly one input, attacked correctly:
| Timing | Action | Why |
|---|---|---|
| Week 1 | Pull your actual reports (free weekly at annualcreditreport.com) and dispute any error in writing | Errors are more common than people think; fix the data before optimizing |
| Week 1–2 | Turn on autopay for every account, minimum full statement balance | Locks the 35% input at “perfect” going forward |
| Week 2–4 | Pay revolving balances to under ~10% of limit — specifically before statement close dates | Utilization is the fastest-moving lever |
| Month 2 | Request limit increases on 1–2 existing cards (skip new card applications) | Improves the ratio without an inquiry on every report |
| Month 2–3 | If you have a thin file, become an authorized user on a trusted, long-seasoned family card | Adds age and a payment record you didn’t have to build yourself |
| Month 3 | Re-check scores; set the calendar to re-pay before statement dates monthly | Confirm the snapshot is now routinely low |
Expected honest outcome: with clean reports and utilization dropped from high to single digits, meaningful movement typically shows up within one to two reporting cycles. Scores with actual derogatory marks take far longer — the plan still helps, but time is the dominant ingredient there. Anyone promising “100 points in 30 days” with derogatories on file is selling something.
What Doesn’t Move Your Score (And What People Get Wrong)
- Income. Not an input. A high earner with maxed cards scores worse than a modest earner who pays in full. (Income matters for approval, not the score itself.)
- Checking your own score. Soft pull, no effect. Check as often as you like.
- Carrying a balance to “build credit.” The most expensive myth in this space. You build history by using cards and paying them; interest is pure waste. Pay in full.
- Debit card usage. Builds nothing. Credit scores only see credit.
Keep the Score in Its Place
A credit score is a lender’s risk estimate, not a measure of financial health. You can have an excellent score while carrying expensive debt — the score is happy because you’re servicing payments. I’ve made this point before when we ignored the more important number in the pay-off-debt-or-invest decision: the interest rate you’re paying on existing debt usually matters more to your net worth than the score that debt happens to support. A 780 score with a 25% APR balance is a well-dressed leak.
The healthy framing: manage the five inputs so credit is cheap and available when you need it (mortgage, insurance pricing, some employers), and then put your actual energy where wealth is built — the debt paydown and investing decisions. If you’re carrying balances right now, the method matters: I compared avalanche vs snowball on the same debt stack, and our own payoff year is in the $62,000 debt-free story.
One forward step: tonight, find two numbers — the statement balance and credit limit on your most-used card. Divide. If utilization is over 20%, you’ve just found the single fastest lever available to you, and you now know exactly when to pull it: before the statement closes, not after.
