Application Data Is Scored Alongside The File

Assessment does not read the credit file alone. The application itself carries scoreable information, and in some models it contributes as much as the bureau data.

Two data sources, one decision

A typical decision combines bureau characteristics with application characteristics: time at address, employment status, housing tenure, household composition and the amount requested relative to income.

Both sets are weighted by the same statistical method, based on how earlier applicants with similar declarations performed after the money was lent.

The applicant sees only one outcome, so a decline driven by application data can look inexplicable to someone who has just reviewed a clean file.

Stability characteristics do heavy work

Length of time at an address and in employment are among the oldest predictors in consumer lending, because instability correlates with disruption to payment.

Recent movers and recent job changers therefore score lower on those characteristics regardless of income, and the effect fades as the position endures.

This is why an applicant can be assessed less favourably immediately after a move that improved their circumstances in every other respect.

The amount requested is itself a variable

Requesting close to the maximum available, or an amount large relative to income, is a characteristic in its own right rather than only an affordability question.

Stated purpose can also be scored, since some purposes are associated with different outcomes, which is part of why lenders ask what the money is for.

Applicants sometimes adjust these fields to obtain an approval, but misstatement is a misrepresentation that can invalidate the agreement if discovered.

Verification catches inconsistency

Declared information is checked against the file and other sources where possible: addresses against residency records, income against documents or bank data, existing commitments against the bureau.

Mismatches between what is declared and what is held do not merely correct the record; they can trigger review or refusal because inconsistency is itself a risk signal.

Applicants who round figures generously or omit a commitment often trip this, having intended nothing more than to simplify a form.

The mix shifts over the customer relationship

Application data matters most at the start, when the lender has no direct experience. After the account opens, internal behaviour progressively displaces it.

Which characteristics may lawfully be used, and how far a decision may be automated, differ between jurisdictions and have been the subject of continuing regulatory change.

Because the mix differs by lender and product, two firms reading the same application and the same file can weight them very differently.