Every year, millions of cardholders face the same question when an annual fee posts to their statement: keep it or kill it? And every year, most of them get the answer wrong — not because they're bad at math, but because they're running the wrong equation entirely.
Some people look at a $325 annual fee on their Amex Gold and immediately flinch. Others shrug off an $895 Platinum fee because they used a Centurion Lounge twice. Both approaches are broken. The flincher ignores the value the card generates. The shrug-it-off person is counting sunk benefits that have already been consumed.
I've spent years refining a four-step framework that strips the emotion out of annual fee decisions and replaces it with arithmetic. It works for any card, at any fee level, regardless of issuer. Once you internalize it, you'll never agonize over an annual fee again.
Step 1: Calculate Credits You Actually Use
Every premium card advertises a wall of credits. The Amex Platinum lists over $1,629 in theoretical credits. The Chase Sapphire Reserve advertises around $1,400. These numbers are marketing — they represent the maximum possible value if you contort your life to use every single one.
You shouldn't do that. Instead, go through last year's statements and identify which credits you naturally used without changing your behavior. The operative word is "naturally." A credit you have to force yourself to use — buying Lululemon gear you wouldn't otherwise buy, or dining at restaurants you don't enjoy just because they're on Resy — isn't really worth full value. Discount forced credits by 50% or more.
For the Amex Gold ($325 annual fee):
- $120 dining credits ($10/mo at Grubhub, etc.) — Do you order Grubhub at least once a month? If yes, full $120.
- $100 Resy dining credit — Do you eat at Resy restaurants anyway? Full $100. If you have to search for Resy restaurants specifically, maybe $50.
- $84 Dunkin' credit — Do you buy Dunkin' coffee regularly? If so, $84. If not, $0.
A realistic total for someone who orders delivery regularly and dines out at sit-down restaurants: $220–$304. Not all $304 for most people, but definitely more than $0.
Step 2: Calculate Your Earn Differential
This is where most analyses fall apart. People calculate what they earn with their premium card, then stop. That's meaningless without a baseline. The right question: how much more do I earn compared to a free card?
The baseline is a no-annual-fee card earning 2% cash back (something like the Citi Double Cash or Wells Fargo Active Cash). Every dollar you earn above that 2% baseline is your earn differential — the incremental value the premium card generates.
Amex Gold example:
Say you spend $800/month on groceries ($9,600/year) and $600/month on dining ($7,200/year). The Amex Gold earns 4x MR on groceries and 4x on dining.
| Category | Annual Spend | Amex Gold (4x MR) | Free Card (2%) | Differential |
|---|---|---|---|---|
| Groceries | $9,600 | 38,400 MR | $192 cash | 38,400 MR - $192 |
| Dining | $7,200 | 28,800 MR | $144 cash | 28,800 MR - $144 |
| Total | $16,800 | 67,200 MR | $336 cash | 67,200 MR - $336 |
At a conservative 1.8 cents per MR point (achievable through Hyatt, ANA, or Flying Blue transfers), those 67,200 MR are worth $1,209.60. Subtract the $336 you'd have gotten from a free card, and your earn differential is $873.60.
If you value MR at only 1 cent each (cash redemptions), those same 67,200 MR are worth $672 — an earn differential of just $336. The point valuation assumption matters enormously.
Step 3: Value Your Perks
After credits and earn differential, you need to assign realistic values to soft perks. These are harder to quantify because they depend on how often you use them and what you'd pay otherwise.
Common perks and reasonable values:
- Airport lounge access: $30–$50 per visit (what you'd spend on airport food and drinks)
- Travel insurance (trip delay, baggage): Expected value of $50–$150/year for frequent travelers, based on claim frequency and average payout
- Hotel status: Harder to value — maybe $0 if you rarely stay at that chain, or $200+ if you get meaningful upgrades regularly
- Purchase protection and extended warranty: $25–$75/year in expected value for most people
Be honest here. If you flew twice last year and used a lounge once, that's $40 in lounge value — not the $500 some people claim by pricing it as "Priority Pass membership."
Step 4: The Final Equation
Now add it up:
Net Value = Credits Used + Earn Differential + Perk Value - Annual Fee
If the number is positive, the card is earning its keep. If it's negative, you should cancel or product-change.
Worked Example: Amex Gold
Using the numbers from above for a moderate spender who uses most credits:
- Credits used: $250 (realistic, not max)
- Earn differential: $873.60 (at 1.8 CPP valuation)
- Perk value: $75 (travel insurance, purchase protection)
- Annual fee: $325
Net value: $250 + $873.60 + $75 - $325 = +$873.60
That's a strong keep. Even at a conservative 1.2 CPP valuation, the earn differential alone ($470) plus credits ($250) exceeds the $325 fee. The Amex Gold is genuinely hard to lose money on if you spend meaningfully in groceries and dining.
Worked Example: Chase Sapphire Reserve
Different story. The CSR costs $795 and comes with hefty credits — $300 travel, $500 The Edit hotel credit, $300 dining, $300 StubHub. Theoretical total: ~$1,400.
But let's be realistic. Most people will use the $300 travel credit (it applies broadly). The $500 Edit hotel credit requires booking through Chase's luxury hotel program — maybe you use $300 of it if you take one nice trip. The $300 dining credit requires spending through specific channels. StubHub? Maybe if you go to concerts.
Realistic credits for a moderately active traveler: $600–$800.
Earn-wise, the CSR earns 3x on dining and travel. Against a 2% baseline, that's 1x incremental per dollar on those categories. At $20,000/year in dining + travel spend: 20,000 incremental UR at 1.7 CPP = $340 earn differential.
- Credits used: $700
- Earn differential: $340
- Perk value: $250 (Priority Pass lounges + travel insurance for frequent traveler)
- Annual fee: $795
Net value: $700 + $340 + $250 - $795 = +$495
Positive, but not by a huge margin. And if you only use $500 in credits (which plenty of people do), you're down to +$295. If you also value UR at only 1.25 CPP (portal value), the earn differential drops to $200, and suddenly you're at +$155. That's cutting it close for a card that requires $795 upfront.
Why Most People Get This Wrong
Three common mistakes I see constantly:
Mistake 1: Counting theoretical credits as real value. The Amex Platinum's "$1,629 in credits" is not $1,629 in value to you unless you would have spent that money in those exact places anyway. Most people get $400–$700 in actual, behavior-unchanged value from those credits.
Mistake 2: Ignoring the baseline. Earning 4x MR on groceries isn't 4x value — it's 4x MR minus whatever a free card would have earned. If you skip the baseline, you'll overvalue every premium card by 30–50%.
Mistake 3: Using last year's value to justify next year's fee. The annual fee is a forward-looking commitment. That incredible Hyatt redemption you got last January? Already consumed. The question is whether next year's spending patterns and travel plans will generate enough value. This mistake is so common it gets its own article.
Run the four steps honestly, with your real numbers, and the answer will be obvious. No agonizing required.