Editorial policy

Sourcing

Every field in our card database is collected from primary sources only. We do not aggregate from other comparison sites, scrape user reviews, or accept issuer submissions without independent verification.

That means the issuer's own fee schedules, terms of service and product pages first; regulatory disclosures (FCA, FinCEN, EU MiCA filings) where a card has them; and press releases only when they come from the issuer's verified channels. 3 of the 178 cards are held and used directly; the rest are researched from those documents.

Full sourcing notes and the per-factor rating inputs are documented on the rating methodology page, and the public dataset itself is available on our dataset page.

Verification

Before a value is published, it is checked against the issuer's current published terms. Cards we have held are labelled Tested hands-on; cards researched from documentation are labelled Verified against issuer docs, so you always know whether a claim came from our own wallet or from the issuer's terms. The five steps a hands-on review has to clear, and the claims we refuse to make without them, are set out on how we test.

Update cadence

The catalogue is re-reviewed in passes, a batch of cards at a time, so review dates differ between cards, and every card page shows the month of its own last review. A confirmed fee or availability change does not wait for the next pass; it is applied ahead of it, usually within a few days. The rating formula changes when an audit shows it misreads the data, and each change is written into the methodology page. Reader-reported corrections are investigated as they arrive.

Changes to the fee, reward, region, custody, network, card-type and limit fields are logged with before/after values, snapshot by snapshot, on our data changelog, and card pages carry a "Last reviewed" date so you can see how fresh each entry is.

Independence and conflict of interest

Sweepbase earns revenue through affiliate partnerships with some card issuers. We disclose that relationship plainly, and we structure the product so it cannot distort the data:

• Affiliate status has zero effect on the rating. The formula only reads objective card attributes; it never checks whether a card has a partner link. Merit factors alone control the full +2.4 / -0.7 point range. Cards with a verified partner link carry a visible “Affiliate partner” label, which is a disclosure, not a recommendation and not a scoring input (see the methodology page).
• No paid rankings. No issuer can pay for inclusion, a higher score, or a better position in filtered listings. Sponsored content, if it ever appears, is labelled "Sponsored".
• Single-operator disclosure. Sweepbase is currently a one-person operation: the same editor researches the data and reviews it. We mitigate that by publishing the full, reproducible scoring formula so any reader can independently check a rating against the public data.

The full commercial picture (affiliate networks, revenue use, and FTC compliance) is in our Affiliate Disclosure.

Corrections

If a value does not match the issuer's current published terms, we want to fix it. Reader-submitted corrections are investigated as they arrive, and confirmed fee changes are applied ahead of the next scheduled review pass.

• Report a data error from the Report a data error button on any card page
• Suggest a card we are missing on the submit a card page
• Email editorial corrections to contact@sweepbase.net

Corrections to those tracked fields are visible in the public data changelog.