Blog

How to work out whether a card's annual fee pays for itself

A seven-line worksheet comparing a fee card with a no-fee card using only rewards and credits you would really use, with three labeled examples.

An annual fee is a price, and a price can be worth paying or not. The mistake most people make is to judge it against the card's marketing, which adds up every credit and perk at full face value, instead of against their own year. The only question that matters is whether this card leaves you better off than the card you would otherwise carry, after you count what you would actually earn, actually use and actually avoid paying for elsewhere.

This is a worksheet. It takes about twenty minutes the first time and five after that, and it works for any fee card. All the card details in the examples are invented and labeled as illustrations. They do not describe any real card's current terms, which are on the card's profile page and in the issuer's published terms. Two different questions are often tangled together here. This post is about the recurring yearly sum. Whether a first-year welcome bonus justifies the application is a separate sum, covered in Is a welcome bonus worth it? Work out what the spend really costs.

The sum in one line

Net yearly value = (rewards on the fee card, minus rewards on your no-fee alternative) + credits you would really use + benefits you would otherwise pay for, minus the annual fee.

If the result is above zero, the fee pays for itself on your numbers. If it is below zero, it does not, and the size of the gap tells you how far off it is. The seven lines below break that sentence into pieces you can fill in.

The worksheet

Use the last twelve months of statements as your base, because that is spending you can see.

Line What to enter
1 Annual fee you would pay from year two on (the first year is often different)
2 Your yearly spend in each rewards category the fee card pays extra on, and the rest
3 Rewards the fee card would have earned on that spending, in dollars
4 Rewards your best no-fee alternative would have earned on the same total spending
5 Credits you would really use, each at the amount you would actually spend
6 Benefits you would otherwise pay for, each at what you would pay
7 Net: (line 3 minus line 4) plus line 5 plus line 6 minus line 1

One refinement turns the net into a decision. Break-even credits needed = line 1 minus (line 3 minus line 4) minus line 6. That is the amount of credits you would have to use for the card to break even. If it is more than the credits you could honestly use, the card does not pay for itself on your numbers, regardless of its headline.

Line 1: the fee

Use the fee that applies in year two and beyond, since the first-year charge is sometimes waived or discounted. Also check whether the fee is likely to rise: card fees have been changed by notice, and a rise changes the sum. My card's annual fee just went up. What are my options? covers what to do when it happens.

Lines 2 to 4: what the card earns against the alternative

This is where people compare the wrong things. The honest comparison is not "rewards earned versus nothing," because if you did not have this card you would have another one. It is the fee card against the best no-fee card you would realistically carry, with the same spending run through both.

Start with your spending. Add up the last twelve months by the categories the fee card pays extra on, for example dining and travel, and put the remainder in "everything else." Be careful with category definitions, since a card's bonus categories depend on how a merchant is classified, not on how you think of the purchase. A grocery run, for example, can miss a grocery bonus if the merchant is coded differently. Then check each bonus category for caps and exclusions in the card's terms.

If the points are not cash, you have to choose what a point is worth, and this choice can swing the whole result. Cash back and statement credits are typically worth 1 cent per point. Anything more needs transfers or travel bookings you actually make. Published estimates of valuations differ by publisher and method; The Points Guy's mid-2026 estimates for some bank transferable points currencies were around 2 cents, while other sources gave wider ranges, and these are opinion-based estimates, not guaranteed values. how to value a point, and why published valuations disagree explains why, and a cents per point figure should be the lowest one you have honestly achieved, not the highest one you have read about. The cash back comparison calculator is a quick way to run line 4 against several no-fee options.

Lines 5 and 6: credits and benefits, valued honestly

This is the line that decides most cases, and the most overstated. Take each credit and ask three questions.

  1. Would I spend this money anyway? A $10 monthly statement credit for a streaming service is worth $10 only if you already pay for that service. If you would sign up just to claim it, you are spending money to get money back.
  2. Will I claim all of it? A monthly credit that you use in nine months out of twelve is worth nine times the monthly amount, not twelve.
  3. Does it have strings? Enrollment, a specific merchant, a booking channel, a time window. If you cannot see yourself meeting the condition, the credit is worth nothing to you.

Write down only the amount you would really use. how to count statement credits against an annual fee honestly is a step-by-step guide to this valuation, and

Benefits you would otherwise pay for go on line 6, valued at what you would pay. A trusted-traveler fee is a clean example. Global Entry costs $120 for five years according to U.S. Customs and Border Protection, so a card that reimburses it is worth about $24 a year, if you would apply anyway. Rental car damage coverage is another, but only if you rent cars, decline the rental company's waiver and the coverage fits the way you travel. Each card's coverage, caps and exclusions are in its Guide to Benefits, and they change. Do not count a benefit you have not read the terms of.

Worked example 1: a dining-and-travel card

Say Rosa spends $22,000 a year on cards. Of that, $7,000 is on dining and travel and $15,000 is on everything else. She is looking at an invented Card F: a $250 annual fee, 3 points per dollar on dining and travel, 1 point per dollar on everything else and $200 a year in credits. Her no-fee alternative earns a flat 2% cash back.

Line 3 (Card F). Points: 3 × $7,000 = 21,000, plus 1 × $15,000 = 15,000, for 36,000 points. At 1 cent each that is $360.

Line 4 (alternative). 2% of $22,000 = $440.

Line 5. Of the $200 in credits, Rosa honestly uses $120.

Line 6. Nothing she would otherwise pay for.

Line 7. ($360 − $440) + $120 + $0 − $250 = −$80 + $120 − $250 = −$210.

Card F loses by $210 on Rosa's numbers. The points valuation matters, so here is the sensitivity:

Value per point Card F rewards Net
1.0 cent $360 −$210
1.5 cents $540 −$30
2.0 cents $720 +$150

Each net is rewards minus $440, plus $120, minus $250: at 1.5 cents, $540 − $440 + $120 − $250 = −$30. Rosa would need to redeem for roughly 1.6 cents a point, a figure that needs a travel redemption she actually uses, before the card breaks even. The break-even credits needed are $250 − ($360 − $440) − $0 = $330, more than the $200 the card lists, so at one cent a point no amount of credit use would save it.

A different question is how much more dining and travel spending Rosa would need. With a total spend of $22,000 and a value of 1 cent per point, Card F earns $220 + 2% of her dining and travel spend D ($220 comes from the 1 point per dollar on all $22,000, and the extra 2 points per dollar on D adds 2% of D). Setting the net to zero: $220 + 0.02D − $440 + $120 − $250 = 0, so D = $17,500. She would need to put about 80 percent of her spending into dining and travel.

Worked example 2: a flat-rate card with a small fee

Say Card G charges a $95 annual fee and pays a flat 2.5%, and the no-fee alternative pays a flat 2%. With no credits and no other benefits, the only thing separating them is the 0.5 percentage point gap, so the break-even spend is $95 ÷ 0.005 = $19,000 a year.

  • At $12,000 of spending: Card G earns $300, the alternative $240. Net = $60 − $95 = −$35.
  • At $30,000: $750 against $600. Net = $150 − $95 = +$55.

The break-even formula for any flat-rate pair is fee divided by the rate gap. It also shows how thin the advantage is. A household at $30,000 comes out ahead by $55 for the effort of managing a second card.

Worked example 3: a card kept for its perks

Say Dev's invented Card H has a $200 fee and earns the same on Dev's spending as the no-fee card, so line 3 minus line 4 is zero. It reimburses a Global Entry application, which at $120 over five years is worth $24 a year to Dev, who plans to apply. It also has a $10 monthly credit for a streaming service Dev already pays for, which is $120 a year. Dev rents cars twice a year, and the card's coverage, once Dev has read the terms, would replace the waiver Dev usually buys at an illustrative $90 a year.

Net = $0 + ($24 + $120) + $90 − $200 = +$34. A small win, with three conditions: Dev actually applies for Global Entry, claims the streaming credit and declines the rental company's waiver. If any one fails, the sum drops, and with all three missed it is −$200. The example shows how a perks-driven card depends on habits, not on the card.

Gathering the numbers without losing an evening

The worksheet is only as good as the spending data, and most people get stuck on line 2. A short routine makes it manageable.

  1. Download twelve months of transactions. Most issuers let you export a statement history as a spreadsheet. If you spread spending across several cards, download all of them and combine the files.
  2. Sort by merchant and tag the categories. You do not need to be perfect. Tag the big categories the fee card rewards (for example dining, airfare and hotels) and total each one. Everything left over is "everything else."
  3. Remove what would not move. If some spending sits on a card for a reason, such as a business reimbursement or a promotional rate on a particular purchase, leave it out of both columns.
  4. Check the total against your bank statements. If the card totals are lower than you expected, you may be paying some bills by other means. Count only spending you could realistically move to the fee card.
  5. Round the way that hurts the fee card. If you are not sure whether $300 of a spending line is in a bonus category, put it in the plain column. A card that clears the bar only when every estimate goes its way is not clearing it.

The last rule is a habit worth keeping. Optimism in a rewards sum is cheap, and the fee is a certain cost, so errors should land on the side that protects you.

Reading the result

A net of a few dollars either way is a tie, and a tie is a reason to prefer the simpler arrangement, because the fee card asks more of you. A net of $100 or more in the card's favor, with credits you have already planned to use, is a case you can reasonably act on. A net below zero by more than your most optimistic adjustment is a case for the keep-downgrade-or-close decision. Margins matter because the inputs move: your spending changes, the issuer changes the fee or the credits, and your valuation of points changes if your travel plans do.

Notice what the sum does not ask. It does not ask whether the card is good. It asks whether you are better off with it than without it, given what you do. Two people with the same card can reach opposite answers, which is why published rankings and "worth it" headlines are weak guides, and a worksheet filled in with your statements is a stronger one.

What the worksheet leaves out

  • Interest. If you carry a balance, the interest is much larger than any of these figures, and a rewards sum is the wrong place to start. Interest at rates near the Federal Reserve's 22.36 percent average for accounts assessed interest in the third quarter of 2026 outweighs rewards of a few percent.
  • Welcome bonuses. They belong in a one-time, first-year sum, not in the recurring yearly sum.
  • Credit-score effects. A new account and a keep-or-close decision move your reported limits and average account age, and those effects are in the decision post, not here.
  • Your time. Tracking credits has a cost. If a plan requires a monthly reminder, count that.

When the answer is "not quite"

A negative net does not mean the card is a mistake, only that it is not paying for itself in cash terms. You have several options: keep it anyway for a reason the sum does not capture, ask the issuer about a no-fee version or change the card to a no-fee product, or close it. I have an annual fee card I barely use. Keep, downgrade or close it? sets out the criteria. If you are comparing a fee card with a no-fee card for the first time, Annual fee or no annual fee? Three households run the numbers works the same sum across three different households. You can also use the annual fee break-even calculator to run your own numbers in a form.

Redo the sum each year, and again whenever the fee, the credits or your spending change. Prices and benefits are changed by issuers with notice, and your life changes too.

Key takeaways

  • Compare the fee card with the no-fee card you would otherwise carry, not with nothing; the net is the rewards gap, plus credits and perks you would really use, minus the fee.
  • Value credits at what you would actually spend and claim, not at face value; a credit you would not use is worth zero.
  • Points are worth 1 cent unless you can name a better redemption you will really book. In the invented Rosa example, the same card moved from a $210 loss to a $150 gain between 1 and 2 cents.
  • For two flat-rate cards, break-even spend is the fee divided by the rate gap: $95 ÷ 0.005 = $19,000.
  • Redo the sum every year and whenever the fee changes, and keep interest out of the picture by paying the statement balance in full.

Terms used in this article

Related tools

Latest on this topic