There is a specific version of the points-and-miles hobby that goes like this. You read a review of a card with an eighty-thousand-point welcome bonus. You do the math — the bonus alone, if valued at the middle of any honest valuation chart, is worth about twelve hundred dollars. The annual fee is under three hundred. Even after the fee, that's a nine-hundred-dollar arbitrage on a piece of plastic. You apply. Six weeks later a card arrives, you hit the spend threshold, the bonus posts, and you have made — on the raw math — several hundred dollars an hour on the twenty minutes it took to apply. This is the version of the story every review is built to tell you.
The version that doesn't get told is what happens in month thirteen. In month thirteen the bonus is gone. What remains is the card's ongoing earn rate, its ongoing credits, and its ongoing fee. If the card was built for the bonus and the ongoing earn rate is mediocre, month thirteen is when you notice that the two hundred and ninety-five dollars is now protecting a card that earns you less on the spend you actually do than the card underneath it does. You keep it for the perks. You keep it because closing it will thump your average account age. You keep it because the fee already hit and the year is running. You are, technically, still ahead. You are also, technically, holding a card you would not have applied for if you'd been shown the year-two ledger instead of the year-one one.
The paradox is that the bonus is real. It is not a scam. It is not a trick. It is the biggest single event in the first year of holding almost any premium card, and the math of the year with the bonus is genuinely favorable. The paradox is that the bonus, being that big, quietly overrides the criteria that should decide whether the card is worth holding for the many years after year one. You end up choosing cards for the shape of a one-time event and living with them as a recurring product.
The thing that eventually helped me — and I am, by any honest measure, still recovering from this — is a rule I stole from someone smarter than me. One card at a time, held for at least a year, evaluated on year two's math, not year one's. Not one card ever, at a time, in your wallet — a normal wallet holds three or four for good reasons. One new card at a time. Applied for on the basis of the ongoing product, not the bonus. Given a year to prove out. Kept if the ongoing math still works, downgraded or closed if it doesn't. This is a slow way to accumulate points. It is also the way that produces a wallet you'd hold in a year with no bonuses on any card at all.
The bonus, in this framing, becomes a subsidy on a decision you would have made anyway. If the card is the right card for your ongoing spend, the bonus makes year one especially good. If the card is not the right card for your ongoing spend, the bonus makes year one look good and years two through five look like a slow tax on a decision you took in month zero when you were being paid to take it.
The bonus is the biggest single event in the first year. The card is the compound event of every year after it.
The tell for whether you're in the bonus-shopper's paradox is small and consistent. It's the moment, six months in, when you find yourself running a purchase through the card whose bonus you're closest to hitting, rather than the card whose ongoing earn rate on that purchase is highest. It's the moment you route a work reimbursement through the wrong card because a spend threshold is looming. It's the moment you notice that most of your points balance is technically in transferable currencies but effectively in one currency because the last three cards you got were all in the same family. The bonus is doing its work. The card is not.
Two other things, honestly, help. The first is to price the bonus at what you'll actually redeem it for, not at what a valuation chart says it's worth. Sixty thousand points is worth about six hundred dollars if you'll cash them out, about a thousand dollars if you'll transfer them to the right partner on the right route, and about zero if they expire in a program you don't understand well enough to use. The bonus is not the number on the marketing page. It is what you will actually redeem it for, and — this is the harder part — when.
The second is to be honest about your calendar. A hundred-thousand-point bonus that requires eight thousand dollars of spend in three months is not the same math for someone who runs sixteen thousand a quarter through cards as it is for someone who runs five. If the threshold requires you to time major purchases into a window, or float a payment you'll carry, or run a friend's expense through your card to hit a number, the effective cost of the bonus is not zero — it's some fraction of the interest on the float, or the friction of the coordination, or the tax on being the person who volunteers at group dinners. That fraction is often small. It is never zero. Adding it in reveals the real yield on the twenty minutes.
None of this is an argument against welcome bonuses. It is an argument for putting them back in their box — the biggest single event in year one, weighted correctly against the compound event of every year after. A card that would only be worth holding for the bonus is a card that becomes bad the day the bonus posts. A card that would be worth holding without the bonus is a card the bonus just made better. Applying for the first one is a common mistake. Applying for the second one is what the wallet you'll be happy with in three years is made of.
The trip you'll take with the bonus, in the end, is a small one relative to the trips you'll take with the ongoing earn on a card you actually use. The bonus books one good week. The ongoing earn books the rest of the decade. Pick for the decade.