Last spring I pulled my FICO 8 score from MyFICO and stared at a 684. Not catastrophic, not good. My Equifax report showed a maxed-out $4,200 card, a 30-day late payment from a car loan in 2024, and an average account age of 4 years 2 months. I gave myself one quarter — March 3 through June 1, 2026 — to see how far I could push it.

Final score on June 1: 747. A 63-point jump. Not the 87-point story I wrote about for a single 30-day sprint, but a more durable move.

This is the actual 90-day plan, including the parts that did nothing. If you want the slower, 12-month version, I wrote that up in my Fair to Excellent 12-month breakdown. Here I’m talking only about the compressed timeline.

What the 90-day window can and cannot fix

A credit score is built from five inputs, and they don’t all move at the same speed. Here’s the reality check I wish someone had given me before I started.

The two levers that matter for a 90-day improvement are utilization and inquiries. Everything else is either slow, small, or already fixed. If you’re carrying a 30-day late from last month, no amount of strategy erases it in a quarter — only the passage of time does.

When I tested this, I deliberately avoided a credit repair service. The FTC has repeatedly flagged that these companies can’t do anything you can’t do yourself for free, and they often charge $79–$149/month. I spent $0 on the strategy and $11 on credit report pulls. That was it.

Days 1–14: The diagnostic phase (and the boring part everyone skips)

You cannot manage what you haven’t read. I downloaded all three bureau reports at AnnualCreditReport.com — Equifax, Experian, TransUnion — and printed them. Paper. Highlighters. Yes, really.

What I found in my own report

• A duplicate medical collection for $127 that had been paid in 2023 but was still reporting on Experian.
• An old address tied to a former roommate’s defaulted utility account (wrong person, but linked to my file).
• A $0-balance store card I’d forgotten about, opened in 2019, still helping my average age.

The collection was the single biggest find. I disputed it via Experian’s online portal on March 5, 2026, and it came off on March 19 — 14 days later. My Experian FICO 8 went from 691 to 703 the day the score refreshed.

Open a free account with all three bureaus

Don’t just pull the reports — create accounts so you can lock your files and see soft-inquiry score updates. Equifax offers a free VantageScore 3.0; Experian gives you FICO 8 free through its app. VantageScore and FICO are different models, so don’t compare them directly. Pick FICO as your north star and ignore VantageScore fluctuations unless you’re specifically applying for a lender that uses it (some auto lenders do).

The dispute process, condensed

1. Pull reports at AnnualCreditReport.com (free weekly, by law).
2. Circle anything inaccurate: wrong balances, accounts not yours, duplicate collections, late payments that weren’t actually late.
3. Dispute ONLINE at each bureau (Experian, Equifax, TransUnion all have portals). Do not call — paper trails are better.
4. Bureaus must investigate within 30 days (45 if you dispute from a free annual report per FCRA §611).
5. Take screenshots of everything. I kept a running Google Doc.

I disputed three items total. One was removed, one was corrected (a balance dropped from $412 to $0 after the creditor updated), and one was verified as accurate — a 60-day late from 2023 that was, in fact, my fault. That one cost me nothing further, but it also didn’t disappear. Learn to tell the difference between “inaccurate” and “unflattering.” Only the first is disputable.

Days 15–45: Crush utilization — the highest-leverage move

Credit utilization accounts for 30% of your FICO score, and it recalculates every time a creditor reports (usually once a month, on your statement date). This means you can improve your score in as little as 30 days by changing what gets reported.

My starting utilization was 61%. Here’s what happened when I ran it down:

The jump from 61% to 32% gave me 15 points. Going from 32% to 9% gave me 16 more. The final 9% → 7% gave 13. Notice the pattern: the biggest returns came from crossing below thresholds, not from the raw percentage. FICO scorecards have utilization brackets, and 30%, 10%, and 1% appear to be meaningful cutoffs.

The statement date trick nobody tells you

Your balance is reported on your statement closing date, not your payment due date. If you pay your card down before the statement closes, the lower balance is what gets reported. I set calendar reminders for three days before each statement close and made a manual payment. That single habit was worth roughly 25 points over the quarter.

Code version of the same logic, if you use a budgeting tool with an API:

Pseudocode: pull statement dates, schedule a payment reminder

curl -X POST https://api.yourbank.com/v1/reminders
-H “Authorization: Bearer $TOKEN”
-d ‘{ “account”: “visa-4242”, “trigger”: “3_days_before_statement_close”, “action”: “pay_down_to_under_9_percent” }’

You don’t need an API. A repeating phone reminder does the same job. I used a paper calendar because it was on my desk.

Should you ask for a credit limit increase?

Yes — but time it right. I called Chase and asked for a limit increase on my $4,200 card on April 22, 2026. They raised it to $7,500 (a soft pull, no hard inquiry, confirmed on the call). That single change dropped my utilization from 32% to 18% overnight without me paying another dollar.

Caveats: a hard inquiry hurts for up to 12 months. Only ask when (a) you’ve had the card 12+ months, (b) you have no recent late payments, and (c) you’ve already paid down the balance so you’re not just asking for more rope to swing on. Also — limit increases can trigger a financial review on some issuers if your income doesn’t support the new limit. If you’re asked for income documentation and can’t provide it, some banks will reduce your limit instead. That happened to a friend at Capital One in 2025.

If you’re carrying balances you can’t pay down this fast, jump to my credit card debt payoff guide before you mess with limits. Utilization improvement without a debt plan is theater.

Days 46–75: Stop the leaks and add a controlled new account

By mid-April I was at 718 and plateauing. Two things got me past it.

Freeze new inquiries (except one)

Every hard inquiry costs you roughly 5–10 points and stays on your report for 24 months (though it only affects your score for 12). I had four inquiries from 2025: two from a mortgage pre-approval, one from a car lease, and one from a store card I opened for the 15% signup discount. Stupid move. That $80 in savings cost me an easy 5 points.

For the 90 days I applied for nothing. Not a card, not a limit increase on a second bank, nothing. The inquiries aged without me adding to the pile.

Optional: an installment loan to fix credit mix

Credit mix is 10% of your score, and if you have only revolving accounts (cards), adding an installment loan can add a few points. I did not do this — I already had a car loan reporting, so my mix was fine. But if your report shows 100% credit cards, an installment loan from a credit union (some call these “credit builder loans”) can nudge you 5–10 points within 60 days.

Warning, and I want to be blunt: this is the section where people do the most damage to themselves. Do not take out a $5,000 personal loan at 24% just to “diversify.” A credit builder loan is a deposit-secured product — you pay $50/month, the bank holds the money, and you get it back at the end. That’s fine. A real loan with real interest is not worth 8 points.

Autopay everything (this is the real payer-historian move)

I set autopay on every account — even the two I pay manually — with the minimum autopay enabled as a safety net. This is table stakes, but 30% of the reason my score didn’t drop during the quarter. If you want an even more aggressive system, I wrote up 14 automation rules in my finance automation guide that covered my bills, savings transfers, and these autopay setups.

Days 76–90: Let the statements report and don’t blow it

This is the boring stretch. Nothing to dispute, nothing new to open.

What I did:

• Kept utilization below 9% on every card (both individual and aggregate).
• Let one card report $0 for two consecutive cycles (a “zero-heavy” pattern, which some scoring models slightly prefer over tiny balances).
• Didn’t dispute anything further, because nothing new surfaced.
• Pulled my score twice — once on May 15 and once on June 1 — to confirm the trend, not to obsess.

That’s the whole plan. No consultants, no credit repair letters, no subscription services.

What didn’t work (so you don’t waste your time)

I tried four things that produced zero measurable improvement. You should skip all of them.

1. Paying for a credit monitoring service beyond the basics. I trialed Identity Guard for one month ($25). It showed me the same data Experian gives away free. Cancelled after 28 days.
2. Disputing accurate late payments. I tried one anyway. It was verified as accurate within 22 days, and there’s some evidence a “frivolous” dispute can flag your file for a period. Not worth it.
3. Closing a card I wasn’t using. I almost closed a 2019 store card with a $0 balance. Doing so would have cut my average account age and dropped my score an estimated 8–12 points. I left it open.
4. Asking for a limit increase on a card I’d only held for 8 months. I tried this on my Capital One card and was declined. No score change, but a wasted inquiry risk. I should have waited until the 12-month mark.

The pattern: anything that costs money, or that asks the credit system to make an exception for you, generally doesn’t work in 90 days. Anything involving your own behavior — paying down balances, avoiding applications, pulling the calendar forward — does.

Where a 90-day plan stops and a 12-month plan begins

The reason I could hit 63 points in a quarter is that my starting point was messy, not catastrophic. I had high utilization and a small collection. If your report has:

• a recent bankruptcy (Chapter 7 stays 10 years; Chapter 13, 7 years),
• multiple 30/60/90-day latenesses in the last 24 months, or
• a home foreclosure or short sale within 3 years,

then 90 days won’t move much. It can’t. The scoring formula has a baked-in recovery time, and no strategy compresses it. In that case, the honest plan is the 12-month rebuild, not a sprint.

One more thing: if you’re using a credit score to qualify for a mortgage, know that mortgage lenders typically pull a tri-merge report and use the middle of your three FICO scores. My 747 in June was the Experian score. My TransUnion score was 731, and Equifax 739. Lenders use the median — so the number to track is your middle score, not your best one. If you want to read more on the underlying factors, I laid out the category breakdown in Understanding Your Credit Score with a slightly different angle on which factor to attack first.

Where the 90-day plan fits in your larger financial picture

A higher score is a means, not an end. Mine unlocked two things this summer: a 3.4% lower APR on a used-car refinance (saving roughly $890 over the life of the loan), and a $7,500 limit increase I didn’t need but that permanently improved my utilization baseline.

But a 63-point jump in 90 days doesn’t fix a $15,000 debt problem, doesn’t build an emergency fund, and doesn’t change your savings rate. Those things need their own plans — I use a sinking fund alongside my emergency fund to separate “car repair in October” from “I lost my job.” If you want to see how credit, savings, and debt strategy fit together, the personal budget I landed on is the framework I built all of this on top of.

Last honest caveat: your score can drop 20–40 points for reasons entirely outside your control — a creditor reporting a closed account as a $0 balance, a balance transfer landing mid-cycle, or a bureau’s model update. In the 90 days I tracked, my score swung -9 points one week for no reason I could identify, then recovered the next. Don’t panic-trade your credit based on a single week’s reading. Look at the trend line.

If you want to run the same 90-day experiment, start with the three reports, then do the utilization math. The rest is calendar discipline and patience.