CRI rating methodology — how we score crypto cards

Version 1.0 · July 2026

This document describes the transparent methodology we use to score crypto cards for our catalog. The methodology produces a single aggregated score (Card Rating Index, CRI) on a 0–100 scale, used for: the homepage ranking, catalog sorting, ranking within themed collections, and the score shown on each card's page.

The goal is a score that is objective, reproducible, and understandable to the user (important for trust and E-E-A-T in SEO). Comparable approaches: RIS-score (bestcryptocards.info), sentiment score (cryptocards.so).

1. Principles

Transparency. Every criterion, weight, and scoring method is public (the "Methodology" page).
Reproducibility. Two evaluators working from the same data get the same score — scoring is formulaic, not "by eye."
Independence from monetization. A card's sponsorship status does not affect the CRI. Sponsored placements are marked with a separate "Sponsored" badge and pinned visually, but their score is calculated under the same rules.
Currency. Data is re-evaluated at least once a quarter, and whenever an issuer makes a material pricing change (tracked via updated_at).
Context-awareness. Besides the overall CRI, profile weights apply to themed collections (e.g., for arbitrage, the "availability/BINs" weight is higher — see Section 6).

2. Score structure: 6 categories

The overall score is made up of six weighted categories. Each category is scored 0–100, then multiplied by its weight.

#	Category	Weight (overall)	What it measures
A	Cost & fees	25%	Total cost of owning and using the card
B	Availability & onboarding	20%	Geography, KYC level, ease and speed of issuance
C	Functionality	20%	Assets, networks, payment methods, limits, app
D	Rewards	15%	Cashback, bonuses, referral program, staking
E	Trust & reliability	15%	License, age, reputation, security, custody model
F	Transparency & support	5%	Clarity of fees, quality of support, documentation
	Total	100%

Formula: CRI = 0.25·A + 0.20·B + 0.20·C + 0.15·D + 0.15·E + 0.05·F

3. Sub-criteria and scoring

A. Cost & fees (25%)

This measures "cheapness": the lower the fees, the higher the score. Each sub-criterion is normalized (see Section 4) and inverted (a lower fee means a higher score).

Sub-criterion	Share of category	Scoring logic
Conversion fee / spread	30%	0% → 100 points; ≥3% → 0 points
Card issuance cost	15%	Free → 100; expensive → lower
Subscription / maintenance fee	15%	$0/month → 100; high monthly fee → lower
Top-up fee	20%	0% → 100; ≥5% → 0
ATM withdrawal fee	10%	A free monthly allowance → higher
FX / foreign transaction fee	10%	0% → 100; high → lower

Example reference points for conversion: 0% = 100; 0.5% = 85; 0.9% = 70; 1.5% = 50; 2% = 35; ≥3% = 0 (linear-threshold scale, see 4.2).

B. Availability & onboarding (20%)

Sub-criterion	Share	Scoring logic
KYC level	35%	No KYC → 100; minimal (ID, ~minutes) → 75; full KYC+PoA → 40
Geographic availability	30%	Number of supported countries; global → 100
Issuance speed	20%	Instant virtual → 100; physical 7–14 days → lower
Onboarding simplicity	15%	Number of steps, availability of an app/bot

Note: "No KYC" gives a high score in the availability category, but does not exempt a card from risk assessment in category E (Trust). The balance is built into the weights.

C. Functionality (20%)

Sub-criterion	Share	Scoring logic
Supported assets	20%	Count and popularity (BTC/ETH/USDT/USDC/SOL…)
Supported networks	15%	Ethereum, Polygon, Solana, Tron, BSC, Arbitrum…
Apple Pay / Google Pay	15%	Both → 100; one → 60; neither → 0
Card type	10%	Virtual + physical → 100
Spending/withdrawal limits	20%	Higher and more flexible limits → higher
App / dashboard / API	20%	Mobile app, convenient dashboard, API (for teams)

D. Rewards (15%)

Sub-criterion	Share	Scoring logic
Cashback	50%	0% → 0; ≥5% → 100 (adjusted for how realistic the conditions are)
Bonuses / promos / partner discounts	20%	Presence and value
Referral program	15%	Presence and generosity of the %
Yield / staking on balance	15%	APY on balance, staking rewards

Cashback is scored with an adjustment for conditions: if a high % is only available with large token staking requirements, or is paid in the issuer's volatile token, a discount factor (0.6–0.9) is applied.

E. Trust & reliability (15%)

Sub-criterion	Share	Scoring logic
License / regulation	30%	Presence of a license (EMI, VASP), jurisdiction
Age and track record	20%	Years on the market without major incidents
Reputation / reviews	25%	Aggregated reviews, mentions, complaints (ban frequency for arbitrage use cases)
Security	15%	2FA, 3DS, funds insurance, audits
Custody model	10%	Non-custodial (key control) → higher for privacy-focused profiles

F. Transparency & support (5%)

Sub-criterion	Share	Scoring logic
Fee transparency	40%	A full, clear fee schedule with no hidden charges
Support quality	35%	24/7, channels (chat/Telegram), response speed
Documentation / FAQ	25%	Availability of guides, clarity of terms

4. Normalization and formulas

4.1. General category formula

Each category is a weighted sum of its sub-criteria:

Score(category) = Σ (sub_criterion_score × sub_criterion_share)

where sub_criterion_score ∈ [0, 100].

Final index:

CRI = Σ (Score(category) × category_weight)      // result 0–100

4.2. Normalizing numeric metrics (fees, cashback, limits)

Numeric parameters use linear-threshold normalization between a "best" and "worst" reference point:

For metrics where "lower is better" (fees): score = clamp( 100 × (max − value) / (max − min), 0, 100 )
For metrics where "higher is better" (cashback, limits, number of assets): score = clamp( 100 × (value − min) / (max − min), 0, 100 )

The min/max thresholds are set on the "Methodology" page and reviewed quarterly (so the scale stays aligned with the market). Example for conversion: min = 0%, max = 3%.

4.3. Scoring boolean / categorical parameters

Boolean (Apple Pay, 3DS, IBAN): a fixed score for presence/absence (e.g., 100 / 0, or 100 / 60 / 0 for "both/one/neither").
Categorical (KYC, license, card type): a fixed point scale by tier (see the tables in Section 3).

5. Scale and interpretation

CRI	Label	Interpretation
90–100	⭐ Best choice	Top-tier across most criteria
80–89	Excellent	Very strong card with minor drawbacks
70–79	Good	A solid option with some trade-offs
55–69	Fair	Usable, but with notable limitations
40–54	Below average	Suitable only for narrow use cases
< 40	Weak	Significant drawbacks / risks

On a card, the score can also be shown as 0–10 (CRI/10) — a single format will be chosen at rollout. Recommendation: show the number + label + a short "why this score."

6. Profile (contextual) weights for collections

For themed collections, it makes sense to redistribute weights, since "the best card for privacy" is not the same as "the best card for arbitrage." The overall CRI stays the baseline, and within a collection a profile is applied:

Collection profile	A Cost	B Availability	C Functionality	D Rewards	E Trust	F Transparency
General (default)	25%	20%	20%	15%	15%	5%
No-KYC / privacy	20%	30%	20%	5%	20%	5%
Lowest fees	40%	15%	15%	10%	15%	5%
For arbitrage / advertising	20%	30%	25%	10%	10%	5%
Maximum cashback	20%	10%	15%	35%	15%	5%
Non-custodial / self-custody	20%	15%	20%	10%	30%	5%

For the arbitrage profile, category B is dominated by BIN geography and stable compatibility with ad accounts, while category C is dominated by team functionality, bulk issuance, API access, and USDT top-ups.

7. Data sources and verification

Primary sources: issuers' official websites, pricing pages, documentation, apps.
Secondary sources: aggregated reviews, communities (Telegram/Reddit), industry media — used for the reputation category.
Editorial testing: where possible, actual sign-up/use (especially for arbitrage cards: whether BINs "stick").
Verification mark: every card has an updated_at timestamp; critical fields (fees, KYC, geography) are checked quarterly.
Conflict-of-interest policy: sponsorship does not affect the score; it is disclosed via a disclaimer.

8. Managing the methodology in the system

Category and sub-criterion weights, as well as normalization thresholds, are stored in configuration (admin panel) rather than hard-coded — so the editorial team can adjust them without a release.
The CRI is recalculated automatically whenever a card's data changes.
Score change history is retained (for audit purposes and for a "score updated" indicator).
Profile weights are attached to the Collection entity.

9. Worked example (illustrative)

Hypothetical card "X": 0.9% conversion fee, free issuance, $0 subscription, 1% top-up fee, ATM with a free allowance, 0% FX; minimal KYC (~minutes), available in 120 countries, instant virtual + physical; BTC/ETH/USDT/USDC, 3 networks, Apple+Google Pay, high limits, has an app and API; 2% cashback, has a referral program; licensed EMI, 4 years on the market, good reputation, 2FA/3DS; transparent pricing, 24/7 support.

Approximate category scores: A ≈ 82, B ≈ 80, C ≈ 88, D ≈ 55, E ≈ 78, F ≈ 90.

CRI = 0.25·82 + 0.20·80 + 0.20·88 + 0.15·55 + 0.15·78 + 0.05·90
    = 20.5 + 16.0 + 17.6 + 8.25 + 11.7 + 4.5
    ≈ 78.6  →  "Good", ~7.9/10

10. Disclaimer

This rating is an editorial judgment based on publicly available data and our own testing. It is not financial advice and does not guarantee the terms of any specific issuer. Users should independently verify current fees and terms on the official card website before use.

Methodology for comparing and rating crypto cards