Editorial Policy

Exchange review methodology

Exchange reviews are scored on a transparent set of criteria, each weighted and assessed the same way for every brand — never on fabricated or estimated figures. The five weighted factors behind the A–F safety grade, and what each one is worth, are published on the graded ledger itself:

• Custody & safety

  Proof-of-reserves, custody model (custodial vs non-custodial), cold-storage practices and security track record.

• Licensing & regulation

  Valid registration or licensing, with verification links, in the markets served.

• Fees

  Real cost after maker/taker fees and withdrawal charges — not headline figures.

• Coins & liquidity

  Depth and quality of the trading pairs and order-book liquidity.

• Support & experience

  Channels, responsiveness, and account/withdrawal experience.

The overall rating is derived from these criteria, not from commercial terms.

Publishing standards

Eight rules that decide what reaches a page. They are written down because every one of them was, at some point, easier to break than to keep.

• An unsourced figure is labelled, not estimated

  Where we could not verify a number against a primary source, the page says so in place of the number. Never a blank cell, never a plausible guess, never a third-party estimate presented as the venue's own figure.

• Proof of reserves is not an audit

  We report who signed a reserve report as prominently as whether one exists. A self-published snapshot is a claim you can inspect; it is not one an accountant has checked, and it says nothing about liabilities.

• A safety fund is not deposit insurance

  A named insurance or user-protection fund is a company policy, not a state guarantee. We never describe one as covering every loss, and we never infer a coverage limit that has not been disclosed.

• Availability is checked, not assumed

  We do not state that a venue serves a country unless its own terms do. We never explain how to work around a geographic restriction, and we do not treat a separately licensed local entity as the same service as the global platform.

• No score without a published derivation

  The A–F grade comes from five weighted, named, individually sourced factors, and the weighting is on the page. We publish no second number beside it, because a rating with no stated derivation is indistinguishable from one that was made up.

• No byline for a person who does not exist

  Desk-produced content is attributed to the desk, as an organisation. We do not invent an author, a photograph or a biography to make research look like reporting.

• Fees and licences rot fastest

  Time-sensitive claims ship with the date they were checked and a link to the live source. A last-reviewed date on a page nobody re-checked is worse than no date at all: it makes the wrongness look verified.

• A commercial link says what it does

  Partner links are labelled by the action — check eligibility, view the fee schedule — never by a claim about the destination. Each carries rel="sponsored nofollow", a disclosure beside it, and the note that access depends on jurisdiction. Where there is no partner, the button renders nothing rather than a dead link.

Before we publish

Seven checks, run on every page in every language before it goes live.

1. 1Every fee, licence and availability claim re-checked against the live primary source, on the date the page shows.
2. 2Nothing on the page that the source record does not carry — no estimated figure, no rounded-up count, no inferred coverage.
3. 3No commercial link before the content that makes the decision, and no placeholder link left in the body.
4. 4The review date populated and true, and the desk that owns re-checking it named.
5. 5Structured data validated against what the page actually renders, with no rating the page does not show.
6. 6Risk stated where the reader meets the product, not collected into a footer disclaimer.
7. 7Every internal link resolved — in all eight languages, not only in English.