Per-card vs. overall utilization: why one maxed card still hurts

Two people each owe $3,000 across cards with $20,000 in combined limits. Both are at 15% overall utilization. One of them has the entire $3,000 sitting on a card with a $3,000 limit, maxed out. The other has it on a card with an $8,500 limit.

Those two files don’t score the same, and the reason is that scoring models don’t stop at the overall number. They also look at every card on its own.

The two ratios, side by side

Overall (aggregate) utilization is every reported card balance added up, divided by every card limit added up. Per-card utilization is one card’s reported balance divided by that card’s limit, computed separately for each revolving account.

VantageScore states that its models add up balances and limits for the aggregate ratio, and that “scoring models also consider percent of credit limit used on a card-by-card basis.” It adds that models “will penalize you for having too many cards with high percentages of credit limit used.” myFICO’s 2022 release explains utilization the same way, starting with a per-card calculation and then rolling it up.

Here are the two people from the opening, as an example:

Person 2’s highest card is $3,000 divided by $8,500, or about 35.3%. Neither file is perfect, but Person 1 has a maxed-out card, and that’s a stronger risk signal. FICO doesn’t publish how much stronger.

What the scoring companies’ own examples show

FICO and VantageScore have each walked through a version of this.

VantageScore’s example follows a consumer with four cards: $1,500 on a $3,500 limit (about 43%), $750 on a $1,500 limit (50%), and two unused cards with $5,000 and $3,750 limits. Overall, $2,250 divided by $13,750 is about 16%, which VantageScore calls “very respectable.” But two of the four cards are above 40% on their own. VantageScore then shows what happens if he closes the two unused cards: the limits drop to $5,000 and the same $2,250 becomes 45% overall.

myFICO’s example uses three cards: $1,000 on a $5,000 limit (20%), $4,000 on a $10,000 limit (40%) and $750 on a $1,000 limit (75%). That’s $5,750 over $16,000, roughly 36% overall. The smallest card has the smallest balance and the worst ratio.

That last detail is the trap. People tend to pay attention to whichever card has the biggest dollar balance. The model is paying attention to percentages.

Which number should you fix first?

When you have a fixed amount of money to put toward cards, the order matters. This table is a rough guide.

“Brackets” here is a loose idea. Forum members talk about score jumps at certain per-card percentages, but FICO hasn’t confirmed any specific cutoffs, so don’t spend extra money chasing a rumored line.

What common moves do to each ratio

Take Person 1 from the opening and run four moves through the math. Each row starts from the same place: $3,000 on Card X, $0 elsewhere, $20,000 in total limits.

Paying down wins on every column, which shouldn’t be surprising. The others each fix one thing and cost something else:

• A balance transfer spreads the debt so no card is maxed, but it usually charges a fee and adds a second card with a balance. FICO counts the number of accounts with balances as its own signal.
• A limit increase helps both ratios without spending anything, though it doesn’t fix the underlying balance, and some issuers check your credit before granting one.
• Closing a card does nothing for the maxed card and pushes the overall ratio from 15% to about 26%.

A quick check on your own file

1. Pull a free report from AnnualCreditReport.com and list each card’s reported balance and limit.
2. Compute every card’s ratio and the overall ratio. Mark the single highest card.
3. If any card is above roughly half its limit, plan to pay it down first, timed to post a few days before that card’s statement closing date.
4. Leave paid-off cards open so their limits keep counting toward your total.
5. Re-check after the next statements report to confirm both numbers moved.

The utilization explainer covers the basic formula if you want the full background, and once both ratios are in decent shape, the open question is how low to go, which we take up in is there a best utilization percentage?.