How to choose a credit card that pays you the most on your real spending

October 29, 2026 · 12 min read

Most advice on how to choose a credit card ends with "look at your spending habits." Then it hands you a list of cards. That skips the hard part. Your spending isn't a habit. It's a record. You have twelve months of it sitting in your statements right now.

This guide shows you how to use that record. You'll learn how to turn past statements into category totals, how to run any card's rules against them, and how to compare one, two or three cards after fees. By the end, you'll know what a card would have paid you last year, not what an ad says it could pay you.

Start with a reality check before you compare rewards

Rewards only make sense if you pay your balance in full. If you pay in full by the due date, you generally avoid interest on purchases during the grace period, per the CFPB. If you carry a balance, interest will almost always cost more than any card earns.

The numbers on this are blunt. The CFPB's 2024 report on credit card rewards found that cardholders who revolve debt pay 94 percent of total interest and fees but get less than 30 percent of rewards. A Federal Reserve working paper estimated that rewards cards shift about $15 billion a year between cardholders, with people who run larger unpaid balances effectively paying for other people's rewards.

So the first rule is simple. If you carry a balance month to month, choose a card for its interest rate, not its rewards. Everything below assumes you pay in full.

Two more checks before you go further:

• Know your credit. You can check your reports from all three bureaus once a week for free at AnnualCreditReport.com, as the FTC explains. Premium cards usually expect good to excellent credit.
• Know the cost of applying. A new application usually means a hard inquiry. The CFPB notes that most scoring models look at how recently and how often you apply. That's a reason to test a card on paper first, so you only apply for the one that wins.

Why your spending matters more than the card's headline rate

A card ad shows its best number. "6% back" or "4x points" is the rate in one category, often up to a limit. What you earn depends on how much of your spending lands in that category, and how much lands everywhere else at 1x.

Here is an example. Say you spend $2,600 a month, which is $31,200 a year:

Now compare two made-up cards.

• Card A: 2% back on everything. No annual fee.
• Card B: 6% back at grocery stores on up to $6,000 a year, then 1%. 3% on gas. 1% on everything else. $95 annual fee.

Card B looks better on groceries. Run the numbers:

• Card A: $31,200 × 2% = $624
• Card B: $360 (first $6,000 of groceries at 6%) + $12 (the other $1,200 at 1%) + $54 (gas at 3%) + $222 (the other $22,200 at 1%) = $648, minus the $95 fee = $553

The flat card wins by $71, even though Card B pays triple on groceries. Most of this person's spending lands outside Card B's bonus categories, and the cap cuts the grocery bonus short.

The card with the best headline rate is often not the card that pays you the most. Only your own totals can tell you which one does.

Use effective rewards rate to compare credit cards

Divide what a card pays you in a year, after the fee, by what you spent. That's your effective rewards rate.

• Card A: $624 ÷ $31,200 = 2.00%
• Card B: $553 ÷ $31,200 = 1.77%

This one number lets you compare a cash back card, a points card and an airline card side by side. It also shows you right away when a fancy card is losing to a plain one.

How to check a card against your own past statements, step by step

This is the part most guides leave out. You can do it by hand or with a tool. The steps are the same.

Step 1: Download 12 months of statements

Pull a full year from every card you use, and from your debit account if you put real spending there. Most issuer websites let you download activity as a CSV file, a PDF statement, or both. CSV is easier to work with.

Why twelve months and not three? Spending has seasons. Holiday shopping, summer travel, back-to-school and once-a-year bills like insurance all show up in different months. A short window can make a travel card look useless or a gift-heavy month look normal.

Step 2: Clean out what doesn't count

Remove anything that isn't a purchase:

• Payments to the card, transfers and balance transfers
• Interest, late fees and other fees
• Refunds, by netting them against the original purchase
• Cash advances and cash-like transactions

Experian's list of transactions that don't earn rewards is a good checklist. It includes cash advances, balance transfers, fees, interest and, on some cards, gift cards and taxes.

Step 3: Sort purchases the way the issuer will

This step trips up most people. Issuers don't care what you bought. They care what kind of store sold it. Each merchant carries a four-digit merchant category code, and Experian explains that these codes help decide which purchases earn rewards. It also notes that warehouse clubs and big discount stores often don't code as grocery stores.

So when you sort your purchases:

• Put superstore and warehouse club spending in its own bucket, not in groceries
• Keep gas bought at a warehouse club separate from gas station spending
• Note brand spending on its own, such as a specific airline or hotel chain, since co-branded cards pay extra only at that brand
• Split travel into airfare, hotels and other travel, because cards treat them differently

The more closely your categories match how cards define them, the more accurate your test will be.

Step 4: Total each category by quarter, not just by year

Many caps reset every quarter. If you only have yearly totals, you can't tell whether you'd hit a quarterly cap in December but not in March. Keep a total for each quarter in each category.

Step 5: Apply each card's real rules

Read the card's terms on the issuer's own site, or look up the card's agreement in the CFPB's credit card agreement database. For each card, note:

• The rate in each category
• Any cap, and whether it's yearly, quarterly or combined across categories
• Whether you have to activate or pick a category
• What counts and what's excluded
• The annual fee

Then multiply. Your category totals times each rate, cut off at the caps, with everything over the cap at the base rate.

Step 6: Turn points into dollars, and be honest about it

Cash back is already dollars. Points and miles are not. Their value depends on how you redeem them. A point might be worth one cent as a statement credit and more if you transfer it to an airline for the right flight.

Use the value you would actually get. If you're not sure, run the test twice: once at the cash value, once at a higher travel value. If the card only wins at the higher value, ask yourself if you'll really book those flights. Remember too that programs can change what points are worth. The CFPB's report lists devaluation as one of the main problems consumers raise about rewards.

Step 7: Subtract the fee, then add only the credits you'd use anyway

Annual fees come off the top. Then look at statement credits, like a monthly dining credit or a yearly travel credit. A credit is worth its full value only if it pays for something you'd buy anyway. A $10 monthly credit at a delivery app you never use is worth $0 to you.

A fair way to do it: count each credit at the amount your past statements show you would have used. If you spent $6 a month at that merchant, the credit is worth about $72 a year to you, not $120.

Is a credit card annual fee worth it? Let the math decide

An annual fee is worth paying when the extra rewards and the credits you really use beat the best no-fee option by more than the fee. That's it.

Go back to the example. Add Card C: 4 points per dollar on dining and groceries, 3 on travel, 1 on everything else, a $250 fee and a $10 monthly dining credit.

• Points earned: 28,800 (groceries) + 14,400 (dining) + 3,600 (travel) + 19,200 (everything else) = 66,000 points
• At 1 cent each: $660 − $250 fee + $120 credit = $530, or 1.70%
• At 1.5 cents each: $990 − $250 + $120 = $860, or 2.76%

At cash value, Card C loses to the plain 2% card. At travel value, it wins. Same spending, same card, opposite answers. The only thing that changed was how you'd use the points. That's why a single "best card" list can't answer this for you.

Year one vs. year two: how much a sign-up bonus is really worth

A sign-up bonus can make almost any card look great in year one. Keep two numbers for every card you test:

• Year-one value: ongoing rewards, minus the fee, plus credits you'd use, plus the bonus
• Ongoing value: the same thing without the bonus

Choose based on ongoing value, unless you plan to downgrade or cancel after a year. A large bonus can hide a card that pays you less every year after.

Also check that your normal spending meets the bonus requirement in time. The CFPB found that in 2022, about one in four consumers eligible for a promotional bonus didn't meet the requirements to earn it. The same report notes that gift cards usually don't count toward the spending minimum, and returns can reduce your total. Your past statements show whether you'd hit the minimum without buying things you don't need.

Rotating categories and spending caps: where estimates go wrong

Caps and rotating categories are where back-of-the-envelope math breaks down. Two real examples, both as of October 2026:

• The Chase Freedom Flex pays 5% on up to $1,500 in combined purchases in rotating categories each quarter you activate, then 1%. That caps what you earn at 5% to $75 a quarter. It's only worth that if the quarter's categories match what you buy, and only if you remember to activate.
• The Bank of America Customized Cash Rewards card pays a higher rate in a category you choose and 2% at grocery stores and wholesale clubs, on the first $2,500 in combined purchases each quarter, then 1%. If you spend $1,800 a quarter on groceries and $900 on your chosen category, you've passed the combined cap.

Compare those with a flat card like the Citi Double Cash, which as of October 2026 earns 1% when you buy and another 1% as you pay, on every purchase, with no annual fee. There's nothing to track, and nothing to forget.

Your quarterly totals from Step 4 tell you which kind of card fits. If your spending is spread out, a flat card often wins. If it's heavy in one or two categories that stay under the caps, a category card can win.

How to pick the right credit card combination: one, two or three cards

You don't have to pick a single card. Often the best result is a flat card for everything plus one card that's strong where you spend the most.

Back to the example. Use Card B for groceries up to its cap and for gas, and Card A for everything else:

• Card B: $360 (groceries to cap) + $54 (gas) = $414
• Card A: $24 (the extra $1,200 of groceries at 2%) + $444 (the other $22,200 at 2%) = $468
• Total: $882, minus Card B's $95 fee = $787, or 2.52%

That's $163 a year more than Card A alone, and $234 more than Card B alone. Card B lost on its own but earns its place as a partner.

How to test combinations:

1. For each category, find which card pays the most on it.
2. Assign each category to that card, watching caps. Spill anything over a cap to the next-best card.
3. Add up the rewards, then subtract every fee in the set.
4. Add a card only if it raises the total by more than its fee and is worth the extra thing to manage.

A third card rarely adds much unless you have a big category the first two don't cover, like heavy travel with one airline. Each extra card is another due date to hit.

Mistakes to avoid when choosing a credit card

• Testing on one month. Use twelve, or at least six.
• Counting superstore runs as groceries. Check how your stores code.
• Counting every credit at face value. Count only what you'd use.
• Valuing points at a number you'll never get. Test at cash value too.
• Picking on year-one value. Bonuses end. Fees don't.
• Forgetting activation. A missed quarter on a rotating card earns the base rate.
• Applying for several cards at once to compare. Compare on paper, then apply for the winner.

Frequently asked questions

How many credit cards should I have for rewards?

Most people get most of the value from one or two. Test your best single card, then your best pair. If the third card adds less than it costs in fees and effort, stop at two.

Which credit card should I get if I don't want to track categories?

A flat-rate card with no annual fee. In the example above, it beat a card with a 6% grocery rate. Run your own numbers, but simple often wins.

What if my spending changes next year?

Adjust last year's totals for what you know will change, then rerun the test. Recheck once a year, since card terms change too.

Conclusion

Learning how to choose a credit card comes down to one shift. Stop asking which card is best. Ask which card would have paid you the most on what you actually spent.

Pay in full first. Pull a year of statements, sort them the way issuers code purchases, and apply each card's real rates, caps and fees. Value points honestly, count only the credits you'd use, and test pairs as well as single cards. Then you'll know, with your own numbers, whether a new card is worth applying for. If none beats what you have, that's a useful answer too. It saves you a hard inquiry.

Next step: see which cards would have paid you the most

You can do all of this by hand. Or you can let PointMaxx do the math on your real statements.

Add your credit card statements as CSV or PDF files. PointMaxx sorts each purchase into a spending category, then works out which one, two or three cards would have paid you the most in rewards after annual fees. It compares 53 US cards whose terms were researched in October 2026. It counts caps, rotating and choose-your-category cards, brand bonuses such as extra points at one airline, and statement credits. You can count the credits you'd really use or leave them out. It also shows year-one value from sign-up bonuses.

When it's done, you can see every card ranked on your spending, a swipe guide for which card to use on each kind of purchase, and your spending by category.

It's free. Everything runs in your browser. Your statements are never uploaded or stored, and there are no accounts. You can read how that works on the privacy page, and find answers to common questions in the FAQ.

PointMaxx gives you an estimate based on your past statements and each card's published terms. It isn't financial advice, and it doesn't check your approval odds. Use it to narrow the field, then read the issuer's terms before you apply.

Add your statementsSee which cards would have paid you the most last year.