Transparency
How we research, score and disclose
You should not have to trust a crypto site, including this one. This page sets out exactly what we weight when we score a platform, which sources we use for which kind of claim, how our Los Angeles coverage ratings are built, what we do not claim to know, and how we are paid.
- Scoring pillars
- 5
- Paid placements
- Zero
- Source categories
- 5
- Last review
- Sep 2026
Platform scoring
The five pillars, and their weights
Our exchange scores are our own weighted assessment, not an average of anyone else's. The weights below are fixed across every platform we review, and they are published so you can judge whether our conclusions actually follow from our stated criteria. If you think a score does not, we would genuinely like the email.
| Pillar | Weight | What we test | Why it is weighted this way |
|---|---|---|---|
| License transparency | 25% | Can we independently verify the regulatory standing — legal entity named, NMLS ID published, state license numbers listed? | A claim we cannot check is worth nothing. Publishing specifics invites scrutiny, which is why weaker operators avoid it. |
| Real all-in cost | 25% | Dollars in versus market value of coins out, measured across each interface the platform offers. | Headline fee schedules omit spreads. Measuring the round trip exposes what a fee page cannot hide. |
| Liquidity & execution | 20% | How does a market order of realistic size behave? How wide are spreads outside US hours? | A thin book turns a 0.2% fee into a 1.5% total cost through slippage. Fee alone is a misleading metric. |
| California funding rails | 15% | ACH from a California bank, cards that do not decline, wires with a workable cut-off, stablecoin deposits. | A cheap platform you cannot fund is not cheap. California card blocking is common enough to matter. |
| Withdrawal & custody freedom | 15% | Can you move assets to an address you control, how fast, at what cost, and with what holds? | A platform that makes leaving difficult has changed the nature of what you own. |
Scores are expressed out of 10 and represent editorial judgment across these pillars. They are not a measure of investment merit, safety or suitability for you, and they are not a prediction that any platform will continue to operate.
Evidence
Which sources we use for which claims
We try to be disciplined about matching claim types to source types. A statutory number comes from the statute, not from a summary. A license claim gets verified against a public register rather than repeated from a marketing page. An operator's footprint is attributed to the operator, because it is their figure and we cannot independently count machines.
| Source type | Examples | What we use it for |
|---|---|---|
| Statute and regulator material | California AB 39 and SB 401 text, DFPI guidance, FinCEN registration rules, IRS digital-asset guidance, FTB material | Anything we state as a legal requirement or a statutory number |
| Public registers | NMLS Consumer Access, FinCEN MSB registrant list, Investor.gov | Verifying every license claim we repeat |
| Government datasets | FBI IC3 reporting, FDIC household survey, Census QuickFacts | Fraud figures, banking access and demographic context |
| Operator disclosures | Location directories, fee schedules, legal and licensing pages | Footprints, terms and product detail — always attributed as the operator's own figure |
| Our own field work | Account opening and funding, quote-checking at machines, neighborhood visits | Cost measurement, coverage density ratings and local practical detail |
Los Angeles ratings
How our coverage density ratings work
The coverage ratings on our locations guide and in the finder on our Bitcoin ATM guide are editorial judgments, and it is important to say what they are not: they are not machine counts.
A raw count would be misleading in a metro shaped like this one. A district with ten kiosks and 90,000 residents is worse served than one with six and 20,000. So we rate machines relative to population and to the density of the cash-handling retail that hosts them, then adjust for two practical factors: whether two-way units exist in the area at all, and whether venues are the kind you can reach at normal hours.
Inputs are operator location directories — their own, not third-party aggregators, which lag by weeks — plus population and retail context, plus our own visits. Where we walked a corridor and checked quotes ourselves, we say so in the text rather than implying broader coverage than we actually have.
These ratings are stable at the level of "which districts are well served", and they should not be read as precise. Machine estates change constantly. Always confirm a specific address on the operator's own locator before traveling.
What goes into a rating
- Operator-published locations for that area
- Machines relative to resident population
- Density of cash-handling retail
- Presence of any two-way units
- Venue type and realistic access hours
- Our own site visits, where we made them
What a rating is not
- A machine count
- A measure of transaction volume
- A safety rating for a neighborhood
- A guarantee any specific machine exists today
Honesty
What we do not know
We think a methodology page that only lists strengths is not a methodology page. So, plainly:
- We have no transaction data. No volumes, no operator revenue, no machine-level utilization. Anything we say about the size of this market is inference from public figures, and we label it as such.
- We cannot count machines. Nobody outside the operators and the regulator can. Where we quote a footprint it is the operator's own published figure, attributed.
- Our cost measurements are samples, not averages. We measure real transactions on real accounts, but a handful of purchases is not a statistical study of a platform's pricing across all conditions.
- Kiosk pricing varies machine to machine. Where a fee, limit or feature depends on the individual unit, we say so instead of inventing a single figure. Any site that quotes one number for an operator's whole fleet is guessing.
- Things change faster than we can review. We date every page and state when it was last reviewed. Between reviews, terms move. Confirm with the provider.
- We are not licensed to advise. Not on investments, not on tax, not on law. We explain mechanics and cost; decisions about your money need a licensed professional when the amounts justify one.
Money
Affiliate disclosure, in full
Some outbound links on this site may earn us a commission when a reader opens an account, at no additional cost to the reader. That is how this research is funded.
What that does not buy. No company can pay for inclusion in our rankings, for a position within them, for a score, or for anything we write. We have never sold a placement, we do not run sponsored posts, and no operator has seen a page before publication. Our top-ranked platform is top-ranked because of the pillars above, and the reasoning is published in full so you can disagree with it specifically rather than in general.
How we manage the conflict. Three practices. We publish the scoring weights before the scores, so the criteria cannot be retrofitted to a conclusion. We publish the negatives — every platform review on this site has a "what to watch" section, including the one we rank first. And we regularly recommend routes that earn us nothing at all: ACH-funded order-book purchases, credit-union accounts, hardware wallets bought direct from manufacturers, and telling readers not to use kiosks when they have a bank account.
Every external link on this site is marked nofollow. That
includes government sources, operator sites and platform links alike. We do not sell
links, we do not exchange them, and we do not pass ranking signal to anyone.
We think this disclosure matters more in crypto than in most categories, because the affiliate incentives here are unusually large and the harm from bad advice is irreversible. If our content ever reads like it was written for the commission rather than the reader, that is a failure and we want to hear about it at contact@losangeles-crypto.com.
Maintenance
How often we review, and how corrections work
Review cycle
- Regulatory pages — reviewed on any statutory or DFPI change, and at least quarterly.
- Platform rankings — reviewed quarterly, and whenever a platform materially changes terms or availability.
- Kiosk and coverage pages — reviewed at least twice a year; operator footprints move constantly.
- Every page — carries a visible review date. Last full site review: September 2026.
Corrections
Email contact@losangeles-crypto.com with the page URL, the specific claim, and what it should be. Factual errors get fixed quickly and the change is noted. We would rather publish a correction than leave a wrong number in front of the next reader, and we would rather hear that our reasoning is wrong than that our facts are — those emails are the useful ones.
Contact usNext in this series
See the methodology applied
Best crypto exchanges, scored
The five pillars applied to ten platforms, with the negatives published.
Read the guide ATMsBitcoin ATM guide and finder
Ten operators compared, with per-machine variability stated rather than averaged away.
Read the guide LocationsATM locations by neighborhood
Where the coverage density ratings come from.
Read the guide DataLA crypto market report
Our structural read, with its own limits section.
Read the guide AboutAbout LA Crypto Desk
Why we built this and what we will never do.
Read the guide ContactGet in touch
Send corrections. They get answered first.
Read the guide