Editorial Policy

? Last updated: May 10, 2026 Next scheduled review: August 2026

This page is the source of truth for every editorial standard on DeFi Explained - defi-explained.dev. It covers how we find facts, how we avoid conflicts of interest, how we handle corrections, and where AI fits into our drafting workflow. If you are a reader checking our work, a protocol team evaluating our coverage, or a partner asking about sponsored content, everything you need to know is here.

1. Our Mission and Editorial Stance

DeFi Explained exists to make DeFi mechanics legible to people who have general financial literacy but no background in smart-contract development. That audience is served by three commitments:

• Accuracy over speed. We would rather publish a thorough explainer two weeks late than a confident-sounding error on day one. When a major protocol change breaks our existing content, we prioritize a correction banner over a full rewrite because readers deserve to know the old content is wrong immediately.
• Mechanistic clarity. We explain how protocols work, not just what they do. A description of Aave's interest rate model includes the kink formula. A description of Uniswap V3 includes the constant product invariant. Numbers without mechanism leave readers unable to evaluate edge cases.
• Consistent risk disclosure. Every protocol has failure modes. We describe them fully and in the same section as the description of the protocol - not in a footnote or a separate risk page that most readers won't find. If a protocol has a history of governance attacks or oracle failures, we say so when we explain the protocol, not just in a list of warnings.

2. Sourcing Standards

We use a four-tier source hierarchy. Higher tiers always override lower tiers.

For smart-contract behavior, we reference the verified source code with a specific contract address and function name. For numeric claims, we state the query method and (where practical) the block number or timestamp of the data. For descriptions of governance decisions, we link to the on-chain proposal or governance forum post.

3. Fact-Check Process

Every published piece goes through the following steps before it appears on the site. No step is skippable.

3a. Draft against primary sources

The writer produces a first draft referencing tier 1 and 2 sources. The draft includes citations in brackets - not a bibliography footnote, but an inline citation with a contract address, function name, or document section. This makes review auditable.

3b. Technical review

A second person with smart-contract reading experience checks every technical claim against the deployed source code or official documentation. This is where confident-sounding errors get caught: a parameter named differently than expected, an interest rate formula that doesn't match the docs, a TVL number that was stale when the article was drafted.

3c. Numeric verification

APYs, collateral factors, fee percentages, token supply figures, and any other number that can be queried on-chain are verified against a live query or a named Dune dashboard as close to publish time as possible. Numbers from documentation are checked against the current version - not a cached version from a previous draft session.

3d. Calculator validation

Every interactive calculator, simulator, or slider on the site is validated against a known test case with a known expected output before it is embedded in a published page. This is run by the technical reviewer, not the writer.

3e. Editorial copy-edit

The final draft is reviewed for clarity, plain-language accuracy, and consistency with our editorial stance. The editor checks that risk disclosures are co-located with protocol descriptions and not deferred to a footnote.

3f. Publish with last-updated signal

Every published page carries a visible "Last updated" date. Major revisions that change a factual conclusion are noted in our internal changelog, which is the source for the correction notes that appear on-page for significant errors.

3g. Content Review Workflow

4. Update Cadence

Content freshness is governed by three tiers:

• Tier 1 - Breaking updates (within 48 hours): Major governance votes that change economic parameters (borrowing rates, collateral factors, fee switches), emergency upgrades, or audit completions on featured protocols. A correction banner is added to the relevant page immediately.
• Tier 2 - Rolling review (90-day cycle): The top 50 pages by organic search traffic are reviewed on a rolling schedule. This catches parameter drift, new feature launches, and emerging failure modes that weren't known when the content was written.
• Tier 3 - Annual review (12-month cycle): All other pages are reviewed at least once per year. Pages that haven't been updated in more than 12 months carry a "Last reviewed" banner if the content is still materially accurate, or an explicit "Needs update" notice if we know the content is stale.

5. Conflicts of Interest and Affiliate Disclosure

DeFi Explained is funded by two revenue streams: display advertising served through Google AdSense, and affiliate commissions from DeFi protocols that appear in our comparison tools and protocol directories. Both are disclosed on every page where they are present.

Affiliate links

Some pages contain links to DeFi protocols, wallets, or exchanges. When a reader clicks one of these links and creates an account or deposits funds, we receive a commission. We maintain this revenue stream because it lets us keep the site free without paywalls. Our rule is absolute: affiliate relationships do not change editorial conclusions. If a protocol that pays us a commission has a known failure mode, we describe it the same way we describe the failure modes of protocols that don't pay us. A page with affiliate links carries a disclosure banner at the top and again at the point of the first affiliate link. The disclosure does not appear in search engine metadata.

Sponsored content

We do not publish content written at the direction of a protocol or its marketing agent. If a protocol team asks us to write or update a page, we treat the request as a tip about a topic that warrants coverage - not as an assignment. The protocol gets no advance review of the content, no right to request changes before publication, and no guarantee of a favorable conclusion. If the protocol's mechanics don't support a positive description, the page will be neutral or negative. Sponsored content is labeled with a "Sponsored" badge in the page header and in the structured data.

AI writing tools

We use AI writing tools (including GPT-4 and Claude) as drafting aids. A writer might ask an AI to summarize a 40-page audit report into key findings, or to translate a mathematical specification into a plain-language paragraph. The AI output is always a starting point - never a final text. Every AI-assisted draft goes through the full fact-check process above. The AI is explicitly prohibited from generating numeric data, fee percentages, APY figures, or any claim that requires primary-source verification. AI use is noted in our internal workflow documentation but is not separately disclosed to readers because we hold AI-assisted content to the same standard as content written without AI assistance - the review standard is identical.

6. Corrections Policy

We distinguish between two correction tiers:

• Minor corrections (typos, broken links, formatting errors, small numerical rounding differences) are fixed silently. The last-updated date is bumped. No correction note appears on-page.
• Substantive corrections (wrong APY, incorrect parameter description, flawed mechanical explanation, outdated risk disclosure) are corrected with a visible correction note in a highlighted box at the top of the affected section. The note states what was incorrect, when it was corrected, and what the correct value or description is. The page's last-updated date is also bumped. If the error is significant enough to change a reader's financial decision, we consider whether a public correction note on social media is warranted.

Reader corrections are welcome at our contact page or [email protected]. We aim to acknowledge every substantive correction within 48 hours and to publish a corrected version within seven days.

7. Author Expertise

Content on DeFi Explained is produced by the editorial team and reviewed by the senior editor. The team reads protocol specifications, on-chain data, and audit reports as part of its normal workflow. Individual pages do not carry bylines because content goes through team review before publication - the byline would imply individual authorship for work that is genuinely collaborative.

Senior editors have backgrounds in smart-contract development, financial engineering, or protocol research. The editorial team's work history includes: quantitative DeFi research at a crypto fund, Solidity audit contributions to Trail of Bits and OpenZeppelin, and governance participation in Aave, Compound, and Uniswap DAOs. This background is mentioned to establish credibility, not to create individual accountability - the team as a whole is responsible for every page.

Frequently asked questions

How often is DeFi Explained content updated?

High-traffic pages (top 50 by organic search) are reviewed on a rolling 90-day cycle. Feature pages covering active protocols are updated within 48 hours of a major governance vote, audit completion, or parameter change. Calculator inputs (APYs, fees, collateral factors) are verified at publish time and checked quarterly. All pages carry a last-updated timestamp so you can see exactly when the content was last reviewed.

What are the primary sources DeFi Explained uses?

We work from four tiers of sources in priority order: (1) Deployed smart-contract source code on block explorers, especially event signatures and storage variable names; (2) Official protocol documentation and governance forum posts; (3) Published third-party audits from recognized firms (OpenZeppelin, Trail of Bits, Consensys Diligence, Certora); (4) On-chain data from Dune Analytics, Etherscan, or direct RPC queries. Secondary sources like news articles and analyst reports are used only to corroborate or to point us at primary material.

How does DeFi Explained handle corrections?

Substantive corrections are handled in two tiers. Minor errors (typos, broken links, outdated fee percentages) are fixed silently and bump the last-updated date. Significant errors - wrong APYs, incorrect parameter descriptions, flawed mechanical explanations - are corrected with a visible correction note at the top of the affected section stating what was wrong and when it was fixed. We never quietly correct numbers after publication. Corrections are tracked in our internal changelog visible to our editorial team.

When is sponsored or affiliate content disclosed?

Any page that contains affiliate links to DeFi protocols carries a visible disclosure banner at the top of the content and again at the point of the link. Sponsored content - content written at the request of a protocol or its agent - is held to the same editorial standards as all other content: the sponsor may not review or approve the piece before publication, and every factual claim must still be verified against primary sources. A 'Sponsored' label appears in the page metadata and in the article header. No affiliate relationship changes our description of a protocol's risks or failure modes.

How does AI factor into DeFi Explained's editorial process?

AI writing tools are used as a drafting aid - to generate a first pass at an explainer, restructure unclear explanations, or translate a technical specification into plain language. AI-generated drafts are clearly marked in our internal workflow and are never published without human review. The review step is non-negotiable: every published paragraph has been read and edited by a human editor who checks that the AI's output accurately represents the primary sources. AI is never used to generate numeric data, APY figures, or any claim that requires on-chain or document verification.