How we rank

1. Your numbers come first

Every card is scored on the spending you entered, category by category, against that card's verified earn rates and annual fee. Statement credits and perks are listed in full for you to judge; they are not yet added into the score, and the report says so. The score is a dollar figure: what the card would return over a year on your spend, after its annual fee. Commission data is not an input. If a card that pays us nothing scores highest, it is the recommendation.

2. Each entity is ranked on its own

Your personal spending and each business are evaluated separately, each with its own fee and its own application. They are not pooled into one imaginary wallet. One thing is shared: the application clock. Personal and business applications all land on one person's credit file, so the application order treats them as one sequence with spacing between them.

3. Two finalists, and we don't crown one

The card that scores highest by the numbers is always listed first. If a different card serves your stated goals better — the programs that reach the trips you actually described — it is shown as a co-equal finalist when it is within a defined range of the numbers pick. When it isn't within range, the numbers pick is primary and the goals card is shown with the cost of choosing it. Either way, the preference is priced in dollars. We tell you what the trade-off costs; you decide.

4. Goals are mapped, not guessed

"Caribbean all-inclusive" or "Japan in business class" is matched to the programs that reach it through a table we maintain — not by a model recalling it from memory. When the honest answer is that points don't reach a goal well, the report says so.

5. Facts have an expiry date

Card fees, welcome offers, and earn rates are re-verified on a rolling cycle; credits and transfer partnerships carry the date they were last checked by hand. Nothing older than 120 days is presented as current, and each public card page shows its last verification date. Offers that fall outside that window are suppressed rather than shown stale.

6. The better offer wins, even when it isn't ours

We hold no affiliate links yet; every apply link on this site goes to the issuer. The rule for when we do is fixed now: if the public offer beats the one behind our link, we say so and link the public offer. Targeted or mailed offers you receive directly from an issuer can beat anything we can see; we say that too.

7. What the report will not do

• It will not soften a recommendation because the card is expensive or unfamiliar. If the math says a fee is worth paying, it shows the number and lets you decide.
• It will not claim a bonus is at a historic high or that an offer follows a cycle until we have enough observed history to say so.
• It will not assume a business application avoids an issuer's rules, or that every entity qualifies for every offer. Eligibility is evaluated per entity from widely reported issuer rules, which change without notice, and issuers make the final decision.

8. How the Perko Pro numbers are computed

The Pro panel in a report is arithmetic on observed data, never a model's estimate. Timed-transfer potential multiplies the points your plan projects for a currency by the transfer bonuses we have actually recorded for that currency and program, at their observed percentages, valued at the same cents-per-point the plan uses. It is shown as a range. How often a bonus recurs is stated only when the record supports it: with fewer than two observed cycles we show the percentages and the count and nothing more; with two or more cycles across a year we say how many times a year it has run; with three or more we give a rough interval. Every bonus in that record carries the date and source it was observed from, and the record is public. The credits line is the sum of the statement credits printed on the recommended cards. Neither number is added to the plan's headline figure.