Key findings
- Addresses are not automatically people, and transfers are not automatically commercial payments.
- Keep observation dates, release dates, access dates and revisions separate.
- Label sourced facts, calculations, interpretations and hypothetical scenarios consistently.
More numbers do not automatically mean stronger evidence
Crypto has abundant transaction records and dashboards. Data volume is different from economic clarity: an address can represent many people, one person can use many addresses and the same asset can move repeatedly inside a complex transaction. Translating those records into users, revenue or payments requires interpretation.
This series aims to attach appropriate evidence to each claim, not maximize the number of statistics. Distinguishing what a source establishes from what an analyst infers lets readers identify what they accept and what they need to check.
| Type | Example | Required disclosure |
|---|---|---|
| Fact | An institution published a document on a date | Original source, date and scope |
| Calculation | Official comparison values rebased to 100 | Inputs, formula and rounding |
| Interpretation | Budget should target a specific bottleneck | Logic, conditions and counterargument |
| Assumption | Hypothetical costs define a break-even point | Explicit hypothetical label, not a forecast |
CBK editorial principles for this series; not independent certification.
Between on-chain transfers and payments
Visa Onchain Analytics explains adjustments intended to reduce effects such as bots, internal transfers and bridge activity. Address labels are estimates, and adjusted volume can still contain several economic uses. Calling the entire adjusted amount commercial payments goes too far. [1]
Read provider filters and retain units before and after adjustment. A changed filter can produce a change unrelated to market activity. A continuous monthly chart does not establish that its definition stayed constant.
Three dates and a version
H1 2025 statistics published in September 2025 have an observation period and a later availability date. Research prepared in 2026 should state both. Backdating the article would place information into an earlier decision when it was not yet available. [2]
All 23 papers were first published on 18 October 2026 as retrospective or structural analysis. The January 2025 onward labels organize topics; they are not historical publication dates. This full English edition was updated on 19 October 2026. Dynamic sources retain their recorded access dates, and current terms are not asserted as past observations.
| Field | Record |
|---|---|
| Claim | What needs verification |
| Original | Publisher, title and direct URL |
| Observation | Period or timestamp measured |
| Publication | Date information became public |
| Access | Date reviewed by the analyst |
| Transformation | Formula, filter, unit and rounding |
| Limits | Sample, duplication, estimation and omissions |
Not every original source supplies every field. Missing fields should remain visible.
Platform metrics have definitions too
Discord visits and communication, and Google Analytics campaign tags, describe different observations. Copying a platform’s label into a report does not ensure readers interpret it correctly. Explain definitions briefly and avoid adding incomparable platform totals. [3] [4]
A campaign click identifies a path, not necessarily a causal effect. The user might have acted through another route anyway. Causal language needs a comparison design and its limitations. An attractive result does not improve the grade of the evidence.
How counterarguments improve conclusions
Stating when a thesis may fail clarifies where it applies. A fee can matter for repeated use while conversion dominates a one-off transfer. This is more useful than an unconditional claim that fees are unimportant.
Uncertainty does not require vague writing. State confirmed facts directly, make calculations reproducible and attach conditions to interpretations. Research quality depends on whether readers can reconstruct the judgment, not the confidence of the tone.
When sources disagree
Compare period, population, unit and aggregation method before choosing between an official statistic and a dashboard. Different measures can both be correct. If they measure the same thing and conflict, inspect revisions and source explanations.
Do not resolve unexplained disagreement through an arbitrary average. Explain the discrepancy, limit the use or withhold the calculation. State what further evidence could change the conclusion rather than filling every gap with precise assumptions.
| Conflict | First check | Avoid |
|---|---|---|
| Units | People, addresses, accounts; currencies | Adding similarly named values |
| Dates | Observation, publication and updates | Applying current data to the past |
| Definitions | Filters and inclusion rules | Choosing the convenient figure |
| Revisions | Version and correction history | Calling a revision a market change |
CBK analysis. This is an analytical framework, not observed customer performance or a market benchmark.
Maintain a research library that can be corrected
Laws, products and protocol documents change. Record correction dates and reasons instead of silently replacing material. Distinguish numeric revisions, link moves and eligibility changes because their implications differ.
The series’ tables and charts are attributed recalculations or original analytical diagrams, not screenshots presented as original findings. Covers use original brand graphics rather than another research firm’s logo or cover. Visual quality and evidence transparency must improve together.
Scope & limitations
These are the series’ editorial and measurement standards, not a claim that every original source is error-free or independently peer reviewed. Unverified matters remain unverified; no private client outcomes were invented.
CryptoBridge Korea analysis based on public materials. Source facts are distinguished from interpretation and hypothetical calculations. References to organizations or projects do not imply their sponsorship or review of this paper.
Sources & scope
- Visa Onchain Analytics, Stablecoin Transactions methodology ↗Adjusted volume is not automatically commercial-payment volume; address labels include estimates.
- FSC / KoFIU, H1 2025 Virtual Asset Service Provider Survey ↗H1 2025 and end-June balances. Comparative values are taken from the same release.
- Discord, Understanding Server Insights ↗Distinguishes visits, communication and retention. Generic examples are not crypto benchmarks.
- Google Analytics, URL builders: Collect campaign data ↗Campaign source tagging, distinct from causal-effect estimation.