Your Credit Score Is a Product Someone Sells

You Don't Have a Credit Score

You have somewhere north of forty of them, calculated by two competing companies across three bureaus in a couple dozen actively-used versions, and the one on your phone is almost certainly not the one your lender pulled.

This is not a consumer protection failure. It's a business model, and it's one of the best ones in American finance. Understanding it changes what you do about your credit, which is why it's worth twenty minutes.

The Royalty That Never Appears On Your Closing Disclosure

Here's what's actually happening in the mortgage market. Fair Isaac Corporation licenses its scoring algorithm to the three credit bureaus, who bundle it into the reports lenders pull. FICO collects a royalty per score.

That royalty was around $0.60 per score in 2018. It went to roughly $2.75 in 2023, $3.50 in 2024, and $4.95 in 2025.

Run the increase: $0.60 to $4.95 is a 725 percent price hike over seven years.

Now here's the part that makes it remarkable. The scoring models used in mortgage lending are FICO Score 2, 4, and 5, known collectively as Classic FICO. They date to the late 1990s. Fannie Mae and Freddie Mac require them. So the price of the product went up more than sevenfold while the product itself did not change at all, because the buyer is legally required to purchase it.

A mortgage requires a tri-merge: one score from each bureau. Three scores per borrower, six on a joint application. At $4.95 that's $29.70 in FICO royalty on a two-person application. The credit report line item you'll see at closing is usually $100 to $175, because the bureaus and the reseller each mark it up on the way through.

The Margin Tells You What Kind of Business This Is

FICO's Scores segment generates roughly a billion dollars a year in revenue at operating margins that have run north of 85 percent. Software companies dream about 85 percent. Utilities would be delighted with 25.

That margin exists because of where FICO sits. It doesn't collect the data (the bureaus do). It doesn't furnish the data (your lenders do, for free, because they're required to). It doesn't generate the data (you do, by paying bills). It owns the three-digit number at the end of the pipeline and charges a toll to everyone who needs to cross.

Every party in that chain has a revenue relationship with FICO except the one who produced the underlying material.

The Score You're Watching Isn't the Score They're Pulling

This is the practical part, and it's where the confusion costs people real money.

• Free dashboards (Credit Karma, most banking apps) typically show VantageScore 3.0, a model built by the three bureaus as a competitor to FICO. No mortgage underwriter uses it.
• Mortgage lenders use Classic FICO: Score 2 from Experian, Score 5 from Equifax, Score 4 from TransUnion. Three different models producing three different numbers, and the middle one is what prices your loan.
• Auto lenders use FICO Auto Score 8 or 9, on a 250 to 900 scale, not 300 to 850. A different range entirely, weighted toward your auto payment history.
• Card issuers use FICO Bankcard Score, also 250 to 900, weighted toward revolving behavior.

A 40 to 80 point spread between what Credit Karma shows you and what your mortgage lender pulls is normal. It is not an error, and calling the bureau to complain about it will not accomplish anything. Two different companies are answering two different questions about the same person.

One development worth tracking: in July 2025 the FHFA approved VantageScore 4.0 for Fannie Mae and Freddie Mac loans and permitted bi-merge reporting instead of requiring all three bureaus. That is the first genuine competitive pressure on Classic FICO's mortgage monopoly in about thirty years. Watch what happens to the royalty. Monopolies price like monopolies right up until the week they can't.

Free Is a Customer Acquisition Cost

Intuit paid $7.1 billion for Credit Karma in 2020. Credit Karma does not charge you anything. Its revenue, running somewhere near $1.8 billion annually, comes almost entirely from referral fees when you click through to a card, loan, or refinance offer.

Do the math on the acquisition. Roughly $7.1 billion for a base of about 110 million members works out to around $65 per member. That is what your attention, and your willingness to check a number twice a month, was valued at.

The score isn't the product. The score is the reason you open the app. The pre-approved offer underneath it is the product, and you are the inventory.

This isn't a reason to stop using it. Free monitoring is genuinely useful and I use it myself. It's a reason to understand why the app is engineered to make you check compulsively, and why the "you could be approved for" module sits directly under the number.

What Actually Moves the Number

The FICO weighting has been public for years. Payment history 35 percent, utilization 30 percent, length of history 15 percent, credit mix 10 percent, new credit 10 percent.

Most of the advice built on that is fine and boring. Two things are worth more than everything else combined.

Utilization Is a Snapshot, Not an Average

Your card issuer reports your balance to the bureaus once a month, on your statement closing date. Not your due date. The statement date.

Which means someone who charges $4,000 a month on a $10,000 limit and pays in full every single month still reports 40 percent utilization, every month, forever. Perfect payment behavior, mediocre utilization number.

Pay the balance down a few days before the statement closes rather than after, and that same person reports 5 percent. Same spending, same money, same perfect payment record. Frequently a 20 to 40 point swing, achieved by changing a calendar reminder.

This is the highest-leverage move available to most people and almost nobody does it, because the entire consumer education apparatus teaches "pay your bill on time" and stops there.

Rate Shopping Windows Differ By Model

Multiple hard inquiries for the same loan type get bundled into one, so shopping around doesn't shred your score. The window is 45 days on newer FICO models. On the older Classic FICO models that mortgage lenders actually use, it's 14 days for some versions.

So the advice you read (45 days) can be wrong for the exact transaction where it matters most. Compress mortgage rate shopping into two weeks and the question resolves itself.

The Part That Argues Against Optimizing At All

Above roughly 760, the score stops doing much for you.

Fannie Mae's loan-level price adjustment grid, which sets what your credit costs you on a conforming mortgage, tops out at a 780-and-above tier. There is no better bracket. An 820 and a 785 get identical pricing on the same loan.

Auto lenders generally cap their tiers around 720 to 740. Card issuers approve their best products well below 800.

So the person on a forum comparing an 811 to an 826 is engaged in a hobby. There's nothing wrong with hobbies. But the marginal dollar value of moving from 790 to 820 is approximately zero, while the value of moving from 640 to 700 is tens of thousands of dollars over a mortgage term. The effort is worth spending where the curve is steep, and the curve flattens hard right about where most people start caring the most.

Know which score your lender pulls. Pay before the statement date. Compress your rate shopping. Stop past 780. That's the whole thing, and none of it requires paying anyone a monthly fee to watch a number that belongs, in every sense that matters, to somebody else.

If you're picking a card to build history with, our Discover it vs. Chase Freedom Unlimited comparison and Sapphire Preferred vs. Venture X cover approval profiles and reporting behavior.