Credit scoring is treated as mysterious, which suits nobody except the people selling credit repair. The weightings are published. FICO groups the data into five categories, and the split is not close to even.
| Factor | Weight | What it measures |
|---|---|---|
| Payment history | 35% | Whether you paid on time, and how badly you missed |
| Amounts owed | 30% | Balances against limits — credit utilisation |
| Length of credit history | 15% | How long accounts have been open |
| New credit | 10% | Recent applications and newly opened accounts |
| Credit mix | 10% | Revolving versus instalment accounts |
Two categories are 65% of the score. If you want to move the number, that is where the work is.
Payment history: the one that cannot be repaired quickly
Paying on time is the largest single input, and its asymmetry is what makes it matter. Years of perfect payments raise the score slowly. One missed payment reported at 30 days past due drops it sharply and stays on the file for years.
The practical implication is that automating minimum payments is the single highest-value action available. Not the full balance — the minimum, as a floor, so that a distracted month never becomes a reported default. Pay more manually on top whenever you can, but never leave the floor to memory.
Worth knowing: most lenders do not report a payment as late until it is 30 days overdue. Paying at day 20 damages nothing on your credit file, though the lender may still charge a fee. If you have realised you are a week late, you are almost certainly fine — pay it now and stop worrying.
Amounts owed: the one you can fix this month
This is utilisation — balances as a proportion of limits — and it is the fastest-moving component of a score. The conventional threshold is to stay under 30%, with better scores generally seen under 10%.
The critical detail most people miss: utilisation is measured on the balance reported on your statement date, not on whether you pay in full afterwards. Someone who spends ₹80,000 on a ₹1,00,000 limit and clears it in full every month is never in debt, pays no interest — and looks like an 80% utilisation borrower to every scoring model, every month.
The fix is timing, not spending less. Pay the card down before the statement generates, so a low balance is what gets reported. Or ask for a limit increase, which lowers utilisation without changing behaviour at all.
Note also that utilisation carries no memory. Unlike payment history, it is recalculated each cycle. A high balance this month is fixed next month — which makes it the only major lever that works quickly before a loan application.
The move that actively hurts you
Closing an old credit card feels responsible. It is usually a mistake, and it damages two categories at once.
It raises utilisation, because the card’s limit disappears from the denominator while your balances stay the same. Close a ₹1,00,000-limit card and every other balance you carry instantly represents a larger share of your remaining credit.
It shortens history — 15% of the score — particularly if the card you close is one of your oldest.
Unless the card charges an annual fee you are not getting value from, or the account genuinely tempts you into spending you cannot control, leaving it open with a small recurring charge and autopay is almost always better than closing it.
Things that do nothing
Checking your own score. That is a soft enquiry and has no effect. Applying for credit creates a hard enquiry, which does — but the two are entirely different events, and the myth stops people from monitoring their own file.
Carrying a balance to "build credit". This is the most expensive myth in the category. Paying interest does not improve your score. Using the card and clearing it works exactly as well and costs nothing.
Your income, savings or job. None of it appears in the score. Lenders consider income separately when assessing an application, but it is not an input to the number.
Multiple rate quotes for one loan. Rate-shopping for a mortgage or car loan within a short window is treated as a single enquiry precisely so that comparing lenders is not penalised. Do compare.
India: CIBIL and the same underlying logic
In India the dominant score is CIBIL, on a 300–900 scale, with roughly 750 and above treated as good by most lenders. The exact weightings are not published as openly as FICO’s, but the drivers are the same: repayment history dominates, utilisation follows, and length of history and enquiry frequency contribute.
One difference worth noting is that in India a single reported default has an outsized effect on loan approval specifically, not merely on pricing — lenders are more likely to decline outright rather than approve at a higher rate. That makes the automated-minimum-payment advice more important rather than less.
In priority order
- Automate at least the minimum payment on everything. 35% of the score, and the damage from missing is disproportionate and long-lived.
- Get reported utilisation under 30%, ideally under 10% — by paying before the statement date or raising limits. 30% of the score and it moves within a cycle.
- Stop closing old cards. Costs you on two factors at once for no benefit.
- Space out applications. Each hard enquiry is minor; several in a short period is a pattern.
- Check your report for errors. Genuine mistakes are common, and disputing one is the only route to a fast, large, legitimate increase.