Individual Utilization: 96% Cost 22 Points

The alert said twenty-two points gone. The dashboard directly beneath it said my overall credit utilization was nine percent.

The two facts did not fit. Nine percent was well below the familiar thirty percent figure. I had not missed a payment, applied for credit, or opened or closed an account. Every summary metric said I was doing this right, yet my score still dropped twenty-two points.

So I stopped reading the summary and opened the underlying report, card by card, the way you would open a patient rather than a chart. What I found was not a mystery once it was laid out: my overall ratio and my worst individual ratio were telling completely different stories, and the model appeared to be reading the one my dashboard had averaged into invisibility.

One caution before the numbers. This is a single file over two ordinary months, not a controlled experiment. I can show you what changed and what followed it. I cannot prove causation from one report, and neither can anyone else.

On the morning of the drop, my report showed four revolving accounts. A primary card with an eleven thousand dollar limit and a nine hundred dollar balance. A second card with a seventy-five hundred dollar limit and a five hundred and twenty dollar balance. A third card with a two thousand dollar limit and a zero balance. And a store card from a furniture purchase two years earlier, with a five hundred dollar limit and a four hundred and eighty dollar balance.

Add the limits and you get twenty-one thousand dollars of available revolving credit. Add the balances and you get nineteen hundred dollars of debt. Divide one by the other and you get nine point zero five percent, the number on my dashboard, the number I had been quietly congratulating myself about.

Two ratios, same file, same day: nine percent and ninety-six percent. Both true. Only one of them was in front of me, and it was the wrong one.

The reason the aggregate hid it is pure arithmetic, and once you see the mechanism you cannot unsee it.

The store card’s five hundred dollar limit was only two point four percent of my total available credit, so it barely affected the denominator. But its balance still mattered: if the card had reported zero while remaining on my report, overall utilization would have been six point seven six percent rather than nine point zero five. The card therefore accounted for about two point three points of my overall ratio. That looks small in the aggregate, but the same balance was ninety-six percent of that card’s own five hundred dollar limit.

That is the whole trick. The aggregate does what aggregates are built to do: it dilutes a severe local condition into a mild global number. A card at ninety-six percent and a card at zero percent sitting side by side produce a blended figure that describes neither of them. My dashboard was not lying to me. It was answering a different question than the one the model was asking, and I had never noticed the difference because for two years the two numbers had happened to agree.

My first fix was the obvious one, and it barely worked.

I had about five hundred dollars available, so I did what feels intuitive: I applied it to the largest balance on the file, the nine hundred dollars on my primary card, taking that card down to four hundred.

Run the arithmetic on what that accomplished. My total balances fell from nineteen hundred to fourteen hundred. My overall utilization dropped from nine point zero five percent to six point six seven percent. On paper that is real progress, more than a quarter of my revolving debt gone in a single payment.

The next score refresh moved three points.

Three. For five hundred dollars. The reason is plain in hindsight: I had moved a ratio that was already comfortable and left the damaged one completely untouched. The primary card went from eight point two percent to three point six percent, a healthy number becoming a slightly healthier one. Meanwhile the store card was still at four hundred and eighty on five hundred, still ninety-six percent, still reporting that way on its own statement date. I had spent my only available money optimizing the number that was not the problem.

The second fix came the following month, was slightly smaller, and did nearly everything. I put four hundred and eighty dollars against the store card and took it to zero.

In aggregate terms this looked minor. My overall utilization moved from six point six seven percent to four point three eight percent, about two and a third points on a ratio that was already comfortable. But the individual ratio on that account went from ninety-six percent to zero, and my highest individual utilization on any card dropped to six point nine percent. The worst single number on my report stopped being a ninety-six percent account.

The next refresh returned twenty-two points.

Two payments of almost identical size with very different results. The variable was not the amount. Four hundred and eighty dollars applied to a five hundred dollar limit clears an entire account. Back at the start, that same four hundred and eighty aimed at my eleven thousand dollar card would have moved it about four percentage points and still left a balance, because that card was carrying nine hundred at the time. When your dollars are limited, the limit you apply them to decides what they are worth.

What the models read explains why this is even possible. In the models whose documentation is public, utilization is not scored as a single blended figure. FICO's own education materials describe amounts owed through several separate considerations, including the proportion of revolving lines currently in use and the amounts owed on specific accounts. VantageScore's 4.0 user guide likewise describes attributes drawn from individual trades rather than the portfolio total alone. The attributes differ by model and version and their weights are unpublished, but the common thread is that a single account's ratio is visible to the model on its own terms.

A high individual ratio can matter on its own, not only through an average. The key factors included with certain credit score disclosures under the Fair Credit Reporting Act use language such as high balance-to-limit ratios on revolving accounts and amounts owed on specific accounts. Those disclosures arise in specified situations, including adverse action and risk-based pricing notices; they do not accompany every score a lender sees. ECOA separately governs the reasons a creditor must provide when denying credit. When a factor statement does appear, it can identify an account-level problem even if the overall ratio looks excellent.

So the ordering rule I use now inverts the instinct most people have. When money is limited and several cards carry balances, I rank them by individual utilization percentage rather than by balance size, and pay from the top of that list down.

On my file that meant a store card holding four hundred and eighty dollars outranked a primary card holding nine hundred, because clearing the first removed a near-maxed account from the report while paying the second moved a number that was already comfortable. I want to be honest that this is a heuristic drawn from one file and from how the models describe their own inputs, not a published scoring rule with a documented cutoff. If a card carries a high interest rate, the cheapest path for your wallet may differ from the tidiest path for your report, and that trade-off is yours to weigh.

The honest summary is that "which matters more" is the wrong question. Both readings exist in the models I can find documentation for, they are computed separately, and they fail in different ways. The aggregate describes your overall exposure and is the number most people track. The individual ratio is the one that can produce a sudden, confusing point drop on a file that looks healthy from ten thousand feet.

The change was a habit, not a technique: before choosing where a payment goes, I now check the percentage on each of my four cards. That takes about ninety seconds. My twenty-two-point result came from one file over two months, so it is correlation rather than proof. The useful takeaway is the order of operations.

1. What is the difference between individual and overall credit utilization?

Overall utilization is your total revolving balances divided by your total revolving credit limits across every card. Individual utilization is a single card's balance divided by that same card's limit. They are calculated separately and can differ dramatically. My file read 9.05% overall while one card sat at 96%.

2. Which matters more, per-card utilization or overall utilization?

Both are read, and they fail differently. The aggregate ratio describes your overall exposure, while a near-maxed account can register as a negative in its own right. That is why a point drop can appear on a file whose overall ratio looks excellent.

3. Why did paying down my largest balance barely move my score?

If your largest balance sits on a high-limit card, paying it lowers an aggregate ratio that was probably already comfortable while leaving a near-maxed small-limit card untouched. In my file, $500 against a $900 balance on an $11,000 limit returned three points.

4. Which card should I pay down first?

As a heuristic, rank your cards by individual utilization percentage rather than balance size and clear the highest-percentage account first. Weigh interest rate separately, because the cheapest payoff order for your wallet can differ from the tidiest order for your report.

5. Does a small credit limit hurt my utilization?

A small limit is easy to max out and contributes almost nothing to your total available credit. In my file a $500 limit was 2.4% of a $21,000 pool, so a $480 balance on it barely moved the aggregate while reporting 96% on its own.

6. When should I pay my card so utilization reports correctly?

Issuers typically report the balance shown on your statement rather than the figure on your due date, but schedules vary and not every creditor reports to all three bureaus. Paying before the statement closes, with time to post, is what changes the reported figure on the accounts where that pattern holds.

7. How many points can clearing a near-maxed card return?

It depends on your file, your limits, your balances, and your history. Mine was 22 points on one report over two months, which is a correlation rather than a controlled result. Anyone who hands you an exact number in advance is guessing.