Credit scoring models differ in detail and agree on the broad weighting: payment history is the largest single factor, credit utilisation the second, and the length of your history the third. Between them those three account for the great majority of a score, which is why the actions that move it are fewer than the advice suggests.
The most common self-inflicted damage is closing an old, unused card — it removes available credit, which raises utilisation, and it shortens the average age of accounts. Both move the wrong way at once.
What weighs how much?
Approximate weightings, consistent across the major models.
| Factor | Roughly | What it measures |
|---|---|---|
| Payment history | ~35% | Whether you pay on time |
| Utilisation | ~30% | Balance against available credit |
| Length of history | ~15% | Average and oldest account age |
| Credit mix | ~10% | Types of credit held |
| New applications | ~10% | Recent hard searches |
Payment history dominates and it is also the slowest to repair. A single missed payment can stay on a file for years, and nothing accelerates its removal — which is why the practical advice reduces to paying the minimum on time even in a month when paying more is impossible.
What is utilisation actually measuring?
The balance reported to the bureau against the limit, and the reporting date is the part people miss. A card paid in full every month can still report high utilisation if the statement is generated before the payment lands.
Thirty per cent is the figure usually quoted as a ceiling, and lower is better — but zero across every card is marginally worse than a small reported balance, because the model wants evidence of use rather than dormancy.
The lever most people overlook is the limit rather than the balance. Requesting a higher limit and not using it lowers utilisation immediately, and it is one of the few changes that acts within a single reporting cycle.
Why does closing a card hurt?
Two mechanisms at once. It removes that card’s limit from the utilisation denominator, so the same balances now represent a higher percentage. And it eventually removes the account’s age from the average, which is the third-largest factor.
The exception is a card with an annual fee you are not getting value from, where the cost is real and the score effect is a few points. Paying a fee to protect a score is usually the wrong way round.
Keeping an old card open with a small recurring charge and a direct debit is the standard way to hold the history without managing it.
Does a thin file behave like a bad one?
Differently, and often just as awkwardly. Someone with no credit history is not scored as low-risk — they are unscoreable, which many automated systems treat as a decline rather than as a neutral.
That is why the standard advice for building a file is to hold one modest line of credit and use it lightly for a sustained period. The history is what is being built, and there is no way to build it faster than time passes.
Do applications matter?
A hard search knocks a few points off temporarily and falls away over months. What does real damage is a cluster of them: several applications in a short period reads as distress, whatever the reason.
Rate shopping for one product is usually treated more gently — multiple mortgage or car loan searches within a short window are often counted as one — but the window and the rules vary by model and jurisdiction, and a soft eligibility check is the safer route where one exists.
What about a credit freeze?
It blocks new applications in your name entirely, which is the strongest protection against identity fraud and has no effect on the score itself. Lifting it takes minutes and it has to be lifted before any legitimate application, which is the whole inconvenience.
It is worth distinguishing from a fraud alert, which asks lenders to verify identity more carefully rather than refusing outright. The freeze is the stronger of the two and the one worth having in place by default if you are not applying for anything.
What does none of it capture?
Income, savings, and whether you can afford the thing you are applying for. A score is a measure of how you have handled credit, not of financial health — it is entirely possible to have a high score and no assets, or substantial assets and a thin file.
Lenders know this, which is why affordability is assessed separately. The score gets you considered; the affordability check decides the amount, and the paycheck and net worth figures are what the second one runs on.
Questions people ask
How long does it take to improve? Utilisation changes within a reporting cycle. History and missed payments take years, and there is no legitimate way to accelerate them.
Does checking my own score hurt it? No — that is a soft search and it is invisible to lenders. Only applications generate hard searches.
Do I need to carry a balance? No. Paying in full is better in every respect; the myth that carrying debt helps costs interest for nothing.
Why do two agencies show different scores? Different models, different data, different scales. The direction of travel matters far more than the number, and comparing across agencies is comparing two different measurements.
Does being on the electoral roll matter? In the UK, substantially — it is used to verify identity and address, and being absent from it causes declines that have nothing to do with creditworthiness.
Pay on time, keep utilisation low, and leave old accounts open. The credit card payoff calculator shows what a balance actually costs to clear, and the net worth calculator covers the position a score says nothing about.