How we research this stuff

Every rate, fee, APR, price and eligibility rule on this site is read from a published source, recorded in one reviewed file with a link and a date, and rendered on the page with that link and date attached. The arithmetic that turns those figures into a verdict is ours, and we publish it in full so you can re-run it with your own numbers. When a figure can only be confirmed against a second-hand source, the page says so on the line where the number appears. When it cannot be confirmed at all, we remove it rather than guess. Our build fails if any figure has gone 120 days without a check.

We work down a hierarchy and always take the highest tier that is actually reachable.

At the time of writing this site publishes 41 sourced figures drawn from 6 published sources. Each one carries its link and its verification date next to the number, not in a footnote.

This is the part most methodology pages leave out, so here it is plainly. Amazon’s product and help pages return server errors to automated requests, and Synchrony’s site did not respond to ours at all. Chase’s card pages and Amazon’s newsroom did respond, which is why the Prime Visa, the Amazon Visa and every Prime price on this site are sourced to the issuer or to Amazon directly.

The Amazon Store Card figures are the ones that suffered, and we handle that in three ways:

• Every affected figure is labeled “secondary source” on the line where it appears. Not on a disclaimer page — on the number.
• We only publish what more than one independent publisher describes consistently. A single second-hand figure is not enough to print.
• We tell you to check the issuer before acting. On financing terms in particular, the terms shown at Amazon checkout on the day are what govern your account, not our page.

The alternative — presenting a second-hand number as if we had read it from the bank — is the thing this whole method exists to avoid.

Every “is it worth it” page on this site publishes a calculation rather than a verdict. The pattern is always the same:

1. The inputs are sourced figures. A reward rate, a membership fee. Each one carries its link and date.
2. The comparison is stated explicitly. Not “is this card good” but “is this card better than this specific alternative”. On the Prime Visa the alternative is the no-membership Amazon Visa, because that is the honest counterfactual.
3. The arithmetic is shown line by line. Fee divided by rate difference equals the spending threshold. Three lines you can check.
4. The threshold is stated as a sentence. The point at which the answer flips from yes to no.
5. The simplifications are named. Where a calculation ignores something — the value of Prime’s non-delivery benefits, say — we say which thing and why.

The numbers are sourced; the arithmetic is ours. That distinction is the whole product. If you disagree with an input, substitute yours and the method still works — which is why we publish the working rather than only the answer.

One place we deliberately do not supply a number: the Prime delivery break-even needs a per-order delivery cost, and that varies by item, speed and order value. We ask you to supply it rather than publishing a figure we cannot source.

This is enforced rather than promised. A script runs before every build and fails the build if any figure’s verification date is more than 120 days old. A stale rate on a financial page cannot ship from this repository without someone re-checking it first.

Stated as a list, because a methodology page that only describes strengths is marketing.

• We have not tested any of these products. No card has been carried, no membership has been tried by us for review purposes, and there is no test lab. Reading published terms is reporting; describing what a card felt like to use would be fabrication, and we do not cross that line.
• Nobody here holds a financial credential. Stephen V. writes this site as an enthusiast. There is no CFP, no analyst, no editorial board, and no “financially reviewed by”. We do not have those things and will not imply them.
• We cannot see or affect your account. Not your credit file, not your application, not your rewards balance, not a charge you are disputing. We cannot expedite a decision or influence one.
• We do not publish approval odds. No issuer publishes a score cutoff for these cards. Any specific number is someone’s estimate, and presenting one as a rule would be inventing certainty.
• This is general information, not personalized advice. Your circumstances are not in our arithmetic because we do not know them. See the terms page for the full disclaimer.

Amazon pays us a flat fee when a reader signs up for a program — a card or a Prime membership — through one of our links. That is a bounty rather than a commission: we are not paid per purchase and we earn nothing on anything you buy afterwards.

The reason we can state that without flinching is on the pages themselves. Every decision page carries a “Skip this if…” section at the same visual weight as the upside. One of the two cards we cover gets a recommendation against it for most readers. Trust pages like this one carry no link at all. If commission were driving the content, none of that would survive a first draft. The full disclosure.

Email info@pointsandprime.com with the page and the figure. Corrections are the most useful thing anyone sends us.

What happens next:

1. We check the claim against its source.
2. If it is wrong we correct it, update the page’s modified date and re-stamp its verification date.
3. Where the error was material — a wrong rate, fee or eligibility rule — we add a dated correction note at the foot of the page rather than editing silently.

Quietly fixing a wrong number on a financial page is the fastest way to lose the only thing this site has. Our full editorial policy sets out the rest.