You check three apps and get three different scores. Nothing is broken. The gaps come from two separate causes, and only one of them is worth your attention.
Cause One: Three Bureaus, Three Different Files
The first reason is that there is no single credit file. Equifax, Experian and TransUnion each maintain their own, built from whatever creditors choose to send them. Reporting to the bureaus is voluntary, and plenty of creditors report to two rather than all three.
So your files genuinely differ. A card you have held for eight years might appear in two files and be missing from the third. A collection account might have been furnished to one bureau only. A limit increase might have reached one bureau a cycle earlier than the others.
Timing compounds this. Creditors report on their own schedules, generally monthly but not on the same day. A score calculated on Tuesday from one bureau and Thursday from another can be working from different balances even when nothing about your behaviour changed.
This is why a serious review means pulling all three reports rather than one. An error you cannot see is an error you cannot dispute.
Cause Two: There Is No Single Scoring Model
The second reason surprises people more. “Credit score” is not one number with one definition. FICO and VantageScore are different companies with different models, and each has multiple versions in active use at once.
Lenders do not all upgrade to the newest version the moment it ships, so older versions stay in circulation for years. There are also industry-specific versions — variants tuned for auto lending or card issuing that weigh the same file differently because they are predicting a different kind of risk.
Run the same file through two models and you get two numbers. Neither is wrong. They are answering slightly different questions with slightly different weights, and their ranges do not always line up either.
What Free Score Apps Are Actually Showing You
Most free apps show a VantageScore, or a FICO variant, built from one bureau’s data. That is a legitimate number, and it is genuinely useful — but it is one model reading one file, which is why it rarely matches what a lender quotes you.
The term for this is an educational score. It exists to let you monitor direction, and for that purpose it is good. If your app score climbed forty points over six months, your file improved. That signal is reliable even when the absolute number is not the one a lender will use.
Where these apps earn their keep is monitoring. A new account you did not open, a sudden balance jump, a new collection — the alert matters far more than the score on the dashboard, because it is how you catch a problem in days rather than at your next application.
Which Number a Lender Actually Uses
That depends on the lender and the product. Many mortgage lenders pull all three bureaus and work from the middle value. Auto lenders frequently use a version tuned for auto lending. Card issuers vary by issuer and sometimes by product.
You generally cannot know in advance which version will be used, and chasing the exact number is a poor use of effort. What you can know is that every model reads the same underlying file. A file with no late payments and low utilization scores well under every version; a file with a recent collection scores poorly under all of them.
The middle-score convention in mortgage lending has one practical consequence worth planning around. If a lender takes the middle of three bureau scores, then your weakest file is not what sinks you — but your strongest one cannot rescue you either. The bureau to work on is the middle one, which means you need all three reports in front of you to even identify it. This is the clearest case where pulling a single report leads you to fix the wrong file.
If you are applying for something significant, you are entitled to know the score used in a decision that went against you — an adverse action notice discloses it, along with the reasons. That is the number that mattered in that decision, and the reasons listed are a more precise repair list than any app dashboard.
How to Stop Worrying About the Spread
Pick one app and one score version to track over time, and ignore the others for trend purposes. Comparing your VantageScore in March to a FICO variant in June tells you nothing, because you changed the ruler as well as the file.
Then shift your attention from the score to the report. The score is a summary of the file; the file is the thing you can actually change. Every point of improvement comes from something concrete — a balance reported lower, an error removed, another month of on-time payments, an old account left open.
A twenty-point gap between two apps is normal and requires nothing from you. A collection account appearing in one of your three reports requires action today. Sorting one from the other is most of what good credit management is.
One last practical note: a dispute resolved at one bureau fixes that bureau only. If the same error sits in two files, you file twice. People who dispute once, see the score move in their app, and assume the job is done are frequently still carrying the error at the bureau their next lender happens to pull.
