The Real Cost of Paying Only the Minimum
A $5,000 credit card balance at 22% costs $8,680 in interest if you pay $100 a month. Raise it to $150 and the same balance costs $2,800 and clears seven years sooner. Here is the arithmetic.
Five factors, with published weightings, explain almost everything about a credit score. Most of the advice circulating about them is wrong, and some of it actively lowers your score.
Credit scoring is unusually transparent for a financial system that affects so much. FICO, whose scores are used in the large majority of US lending decisions, publishes the weighting of the five factors that make up the score. VantageScore uses similar inputs with different weightings, and credit reference agencies outside the US run their own models on comparable data.
The numbers below are FICO's published weightings. The principles behind them apply broadly, even where the specific model differs.
| Factor | Weight | What it measures |
|---|---|---|
| Payment history | 35% | Whether you have paid on time, how late, how recently, how often |
| Amounts owed | 30% | Balances relative to limits, on each card and overall |
| Length of credit history | 15% | Age of oldest account, newest account, and the average |
| New credit | 10% | Recent applications and recently opened accounts |
| Credit mix | 10% | Whether you handle both revolving credit and instalment loans |
The single largest factor, and the one with the least nuance: pay everything on time, every time.
A payment is generally not reported to credit bureaus until it is 30 days late, so being a few days behind will typically cost you a late fee from the lender without appearing on your credit file. Once reported, a late payment stays on the file for years — commonly seven in the US, six in the UK — although its impact fades as it ages.
Severity matters too. A single 30-day late payment is damaging; a 90-day late, a default, a charge-off or a court judgment is far worse and takes far longer to recover from. If you can only pay one thing this month, pay the minimums on everything before paying extra on anything.
This is mostly credit utilisation: your balances as a percentage of your limits.
Most card issuers report your balance to the bureaus on your statement date, not your payment due date. So you can pay in full every month, never owe a cent of interest, and still show 80% utilisation — because the statement was cut before you paid. The fix is to pay down the balance before the statement date, or make a mid-cycle payment. This single adjustment moves some people's scores by tens of points without changing their spending at all.
Both overall utilisation and per-card utilisation are considered, so one maxed card is a problem even if your total is low. Below 30% is the conventional guidance; below 10% is better still. Reporting 0% on every card is marginally worse than reporting a small balance, because the models want evidence that credit is being used and repaid.
Three measures: the age of your oldest account, the age of your newest, and the average across all of them.
This is why closing an old card is often a mistake. It removes the limit immediately, which raises your utilisation, and eventually removes the account's age from the calculation. A card you never use costs you nothing to keep open — put one small recurring payment on it and set it to pay automatically in full.
Each formal application leaves a hard search visible to other lenders, typically for a year or two, with most of the score impact in the first few months. One application is a minor effect. Six in a month reads as financial distress.
Rate shopping is treated differently: multiple enquiries for the same type of loan — a mortgage, a car loan — within a short window are usually counted as a single enquiry, so comparing lenders is not penalised. The window varies by model, commonly 14 to 45 days.
Models like to see that you can handle both revolving credit (cards) and instalment credit (loans with fixed payments). This is the least actionable factor, and the correct response to it is to do nothing: taking out a loan you do not need to improve your mix costs more in interest than the points are worth.
A surprising amount, and knowing it saves worry:
Lenders do consider income and employment. They simply assess those separately from the score, using your application and their own affordability rules.
No company can remove accurate, correctly-reported negative information from your credit file. Anything that can be done — disputing genuine errors, adding a statement of correction, negotiating with creditors — you can do yourself, free, in an afternoon. Be especially wary of any firm asking for payment before doing anything, which is illegal in some jurisdictions.
A "thin file" is not the same as a bad score — there is simply not enough data to score you, which can be just as much of an obstacle. The standard routes in:
Expect six to twelve months before a meaningful score exists, and do not open several accounts at once trying to speed it up.
Because you have many scores, not one. Different models — FICO, VantageScore, an agency's own — run on data from different bureaus, which do not all hold identical records. Lenders may also use industry-specific versions tuned for car loans or credit cards. Treat any single number as an indicator of the direction you are moving in, not as a precise figure a lender will see.
Usually the opposite. Closing an account removes its limit from your utilisation calculation immediately — if you close a card with a $5,000 limit and no balance, your utilisation across everything else rises overnight — and over time the account's age drops out of your history. Close a card only if it charges a fee you do not want to pay, or if keeping it genuinely tempts you to spend.
A single 30-day late payment typically has its largest effect in the first year, fades noticeably after two, and drops off the file after six or seven depending on the country. Serious defaults, charge-offs and judgments take the full reporting period. During that time, consistent on-time payments on everything else are what rebuild the score — there is no faster route.