The Swoopr Source Ladder

Direct Answer

A financial claim is only as useful as the evidence behind the exact statement being made. The Swoopr Source Ladder is a six-level framework for matching a claim to the strongest appropriate source: controlling law or regulator, first-party filing or market record, official dataset, formal methodology or research, high-quality secondary analysis, and finally commentary or community discussion. The goal is not to force every sentence to cite a government website. The goal is to make sure the source actually has authority over the claim, was current for the period being discussed, and preserves enough context that another reader can independently check the conclusion.

For investment research, that distinction matters. A company filing can verify what a company reported. It cannot prove management's forecast will happen. An IRS page can verify a current retirement-plan limit. It cannot tell you whether maximizing that account is the right decision for your household. A Treasury yield table can verify an official published yield for a date. It does not prove a bond fund will earn that yield. Verification is not the same thing as prediction.

Key Takeaways

• Match the source to the claim. The strongest source is the one with authority over the statement, not simply the most famous website.
• Separate facts from interpretations. A filing can establish revenue, debt, dilution, and disclosed risks; your conclusion about whether those facts are attractive remains analysis.
• Record time explicitly. Publication date, effective date, fiscal period, data-as-of date, and retrieval date answer different questions.
• Use primary sources first for high-stakes facts. Taxes, retirement rules, securities regulations, filings, exchange procedures, and official economic data should normally trace to the body that publishes or governs them.
• Preserve disagreement instead of hiding it. When two credible sources use different definitions or methodologies, name the difference before choosing a number.
• Treat AI output as a research assistant, not evidence. An AI response may help locate a source or frame a question, but the underlying source is what supports the claim.
• Build a claim ledger. For consequential research, record the claim, source, date, scope, limitations, and what would make the claim stale.

Why a Source Hierarchy Matters More in Investing

In many subjects, an imprecise source is inconvenient. In investing, it can change the meaning of a decision.

Consider four superficially simple statements:

1. "The company grew revenue 12%."
2. "The 10-year Treasury yield was 4.2%."
3. "You can contribute this much to a 401(k)."
4. "This broker is registered."

Each statement belongs to a different evidence system.

For revenue, the most relevant source is normally the issuer's filed financial statements, preferably the actual Form 10-K or 10-Q and the period in which the number appears. The Securities and Exchange Commission explains that the Form 10-K contains a detailed picture of a public company's business, risks, and financial results. The SEC's EDGAR system is the source of record for those filings, and the SEC also exposes company submissions and XBRL data through data.sec.gov.

For Treasury yields, the official Treasury interest-rate statistics are a better source than a screenshot from a finance app. Treasury explains that its par yield curve is derived from indicative market quotations obtained around 3:30 p.m. each business day and documents changes in methodology and series availability.

For retirement limits, the IRS publishes current limits and historical tables. A 2026 figure copied into an article may be accurate today and stale next year; the IRS page is the continuing authority.

For a broker or adviser, registration and disclosure records live in systems such as FINRA BrokerCheck and the SEC's Investment Adviser Public Disclosure database. A company biography or social profile may be useful context, but it is not the registry.

The practical rule is simple: the source should own, publish, regulate, file, or directly measure the thing you are claiming whenever that is possible.

The Six Levels of the Swoopr Source Ladder

Level 1: Controlling Law, Regulator, or Rulemaking Authority

Use Level 1 when the claim is about what is legally required, permitted, prohibited, limited, insured, registered, or regulated.

Typical Level 1 sources include:

• Internal Revenue Service for federal tax and retirement-plan rules;
• Securities and Exchange Commission for federal securities rules and filings;
• Commodity Futures Trading Commission for federal derivatives regulation;
• Federal Reserve Board for regulations within its authority;
• Department of Labor for employee-benefit rules within its authority;
• federal statutes and regulations;
• official state regulators when the question is state-specific.

What Level 1 can establish

A Level 1 source can establish wording such as:

• a contribution limit;
• an effective date;
• a filing requirement;
• a settlement rule;
• a registration obligation;
• the scope of an exemption;
• the text of an enforcement action;
• the current version of a regulation.

What Level 1 cannot establish by itself

Regulatory authority does not make every publication from that authority an investment forecast. An SEC investor bulletin may explain a product's risks, but it does not prove the product is unsuitable for a particular person. An IRS contribution-limit page tells you the legal ceiling; it does not tell you how much you personally should save.

Swoopr rule

When a statement is fundamentally a "what does the rule say?" question, do not cite a marketing page, finance blog, or AI summary as the only source if the governing authority is publicly accessible.

Level 2: First-Party Filing, Issuer, Exchange, or Market Record

Use Level 2 when the claim concerns what a specific organization reported, filed, published, contracted, or officially documented.

Examples include:

• SEC-filed 10-K, 10-Q, 8-K, proxy statement, prospectus, registration statement;
• an ETF prospectus or shareholder report;
• an exchange rulebook;
• an issuer's audited financial statements;
• a protocol's official technical documentation;
• a fund's official holdings file;
• an index provider's methodology.

The important distinction: reported fact versus economic truth

A filing can verify that a company reported a number. It does not automatically prove that the number is economically comparable to another company's number.

For example, two companies can both publish "adjusted EBITDA" while excluding different costs. If the claim is "Company A reported adjusted EBITDA of X," the company's filing or earnings release may support it. If the claim is "Company A has higher operating profitability than Company B," the analyst must normalize definitions before the comparison is defensible.

The SEC's EDGAR APIs make this distinction especially important for automated research. The API can provide structured XBRL facts, but data users still need to understand period alignment, units, fiscal calendars, and the specific taxonomy concept that generated the fact.

Swoopr rule

Use the first-party record for what was disclosed, then explain any transformation separately.

Level 3: Official Statistical or Administrative Dataset

Level 3 is where a large share of macroeconomic, rate, labor, inflation, demographic, and government-finance research belongs.

Examples include:

• Federal Reserve Economic Data (FRED);
• Archival FRED (ALFRED) for historical vintages;
• Bureau of Labor Statistics;
• Bureau of Economic Analysis;
• U.S. Treasury;
• Census Bureau;
• FDIC;
• SEC structured datasets.

Why an official dataset can still be misunderstood

Official does not mean context-free.

A Treasury par yield is not the same thing as the yield on every Treasury security of that maturity. Treasury explains how it interpolates the curve and documents series breaks.

A CPI release can later coexist with revised seasonally adjusted series. BLS warns that archived CPI news releases may differ from later database values because subsequent releases can include revisions.

A FRED observation can represent the latest available value, while ALFRED preserves historical vintages. That difference is crucial when reconstructing what an investor could have known on a past date.

Swoopr rule

Every dataset used for a consequential claim should carry enough metadata to answer:

• What series?
• What unit?
• What frequency?
• What transformation?
• What date range?
• What vintage?
• When retrieved?
• Was the value revised later?

Level 4: Formal Methodology, Standards, and Peer-Reviewed or Institutional Research

Level 4 is appropriate when the claim is not merely "what number was published?" but "how should this concept be defined, measured, or interpreted?"

Examples include:

• accounting standards;
• index methodology documents;
• peer-reviewed research;
• recognized professional methodology;
• technical standards;
• institutional research with transparent methods.

Why methodology belongs above secondary explanation

A formula can look universal while hiding a choice.

Consider return on invested capital, free cash flow, tracking error, duration, or a factor exposure. Different providers can produce different values from the same company because they define inputs differently. When Swoopr publishes a tool or research note, the defensible approach is to publish the exact formula and assumptions rather than simply cite a provider's final number.

That principle is already part of Swoopr's Research Loop and Evidence & Validation model: the site documents formulas, separates reusable calculation logic, and tests those calculations against known examples.

Swoopr rule

When the conclusion depends on a methodological choice, cite or publish the methodology itself. Do not let a polished output hide an undefined formula.

Level 5: High-Quality Secondary Analysis

Secondary analysis is not inherently weak. It is often the best way to understand a complicated subject after the governing facts are established.

Useful Level 5 sources can include:

• respected financial publications;
• well-documented institutional explainers;
• academic summaries;
• high-quality research organizations;
• specialist commentary with transparent sourcing.

When Level 5 is valuable

Secondary analysis helps with:

• translating a dense rule into plain language;
• comparing interpretations;
• explaining industry history;
• identifying where experts disagree;
• providing context around a primary dataset.

When Level 5 is not enough

Do not use secondary analysis as the sole support for a current legal threshold, an issuer's reported financial result, a fund expense ratio, a regulatory filing, or a statistic that is directly available from the official publisher.

Swoopr rule

Secondary sources may explain the evidence. They should not silently replace accessible primary evidence for material claims.

Level 6: Commentary, Community Discussion, Social Posts, and AI Output

Level 6 is useful for discovering questions, misconceptions, terminology, and real user problems. It is a poor final authority for a material financial claim.

Examples include:

• forums;
• social posts;
• community threads;
• unsourced videos;
• generic blogs;
• AI-generated answers.

Why Level 6 still matters

A forum thread can reveal that many investors misunderstand wash sales across multiple accounts. A Reddit discussion can reveal confusion over ETF distributions. A social post can surface a new rumor that is worth checking.

The mistake is not reading those sources. The mistake is stopping there.

AI output belongs here unless it points to evidence

Google's 2026 guidance on generative AI search reinforces a useful editorial distinction: AI experiences rely on retrieval and grounding, and website owners should keep foundational SEO and people-first content as the base. Google also warns against fabricated shortcuts such as special AI markup or inauthentic mentions.

Swoopr's own Citation & Sources Policy similarly states that AI systems are not authoritative sources and that claims requiring evidence should be checked against primary material.

Swoopr rule

Use AI and community material to find the question. Use stronger evidence to answer it.

The Claim-to-Source Test

Before citing anything, ask five questions.

1. What exactly is the claim?

"Apple revenue was $X" is different from "Apple's revenue quality improved."

The first is a reported figure. The second is an analytical conclusion requiring definitions and supporting observations.

Rewrite vague claims until they are testable.

2. Who has authority over the fact?

If the question is about an IRA limit, the IRS has authority. If it is about a company's reported segment revenue, the company filing is the primary record. If it is about a Treasury curve, Treasury owns the official series methodology.

3. What period does the claim cover?

A source can be authoritative and still be temporally wrong.

Always distinguish:

• publication date: when the source was published;
• effective date: when a rule or policy begins to apply;
• measurement period: the quarter, year, or date being measured;
• data-as-of date: the latest observation included;
• retrieval date: when Swoopr accessed the source.

4. Can the source support the wording you used?

A source that says "may" does not support "will."

A source that says "reported" does not support "proved."

A source that covers federal law does not automatically support a state-law claim.

5. What could make the claim stale?

The answer might be:

• a new fiscal filing;
• a new rulemaking;
• the next annual IRS inflation adjustment;
• a data revision;
• a methodology change;
• a fund prospectus update;
• a corporate action.

That answer becomes the review trigger.

The Swoopr Claim Ledger

For high-impact research, Swoopr should store evidence at the claim level, not merely as a list of links at the bottom of a page.

A practical claim record looks like this:

This is intentionally more demanding than adding a References section. A References section proves that sources exist. A claim ledger documents how the evidence connects to the sentence.

Swoopr's existing Evidence & Validation page correctly notes that automation can verify that references exist and links resolve, but it cannot determine whether a source truly supports a particular sentence. The claim ledger creates the structure needed for that human judgment.

Worked Example 1: A Retirement Contribution Limit

Suppose a page says:

"The 401(k) elective deferral limit is $24,500 in 2026."

The correct evidence path is straightforward.

1. Claim type: current federal retirement-plan rule.
2. Source level: Level 1.
3. Authority: Internal Revenue Service administering the applicable tax-law framework.
4. Source: IRS 2026 retirement contribution limit material.
5. Effective period: tax year 2026.
6. Review trigger: annual IRS cost-of-living adjustment or legislative change.

Notice what the source does not establish: whether an individual should contribute the maximum, whether Roth or traditional treatment is preferable, or whether an employer's plan permits every contribution type.

Worked Example 2: A Company's Revenue Growth

Suppose a research note says:

"Company revenue increased 12% year over year."

The evidence path should record:

1. filing and form type;
2. reporting period;
3. exact revenue line item;
4. prior-period comparison;
5. whether the figure was restated;
6. calculation used.

If a company changes segment definitions or restates prior periods, the analyst must decide whether the comparison is still like-for-like.

The SEC's guidance on Form 10-K emphasizes that the filing includes the business description, material risks, and financial results. EDGAR allows users to retrieve the filing itself rather than depending on a copied value.

Worked Example 3: An Economic Data Point

Suppose a chart says:

"The 10-year Treasury par yield was X on a historical date."

The relevant source should identify:

• the official Treasury series;
• the date;
• the curve methodology;
• whether the series existed at that maturity on that date;
• any documented series break.

Treasury's own page documents that constant-maturity values are interpolated from the daily par curve and identifies methodology changes and historical gaps.

This is why the phrase "Treasury yield" is incomplete. Bills, notes, bonds, par yields, constant-maturity series, auction rates, and secondary-market yields are related but not interchangeable.

Worked Example 4: Is an Investment Professional Registered?

A marketing biography is not the right evidence.

For a brokerage professional, FINRA BrokerCheck is a primary public lookup system. For investment advisers, the SEC's IAPD database provides information on firms and representatives registered with the SEC or states.

A registration record still does not prove that a person is trustworthy, competent for every specialty, or free from every conflict. It verifies a narrower claim: registration and disclosed record information within the database's scope.

What to Do When Sources Disagree

Disagreement is information.

Two sources may conflict because:

• they cover different periods;
• one is revised and one is an original vintage;
• definitions differ;
• one includes a subset the other excludes;
• accounting classifications changed;
• one uses market value and the other book value;
• one source is stale;
• one source contains an error.

The correct response is not to average numbers that have different meanings.

Instead:

1. restate each definition;
2. identify the date and period;
3. identify the governing authority, if any;
4. reproduce the calculation where possible;
5. explain which value Swoopr uses and why;
6. preserve the alternative when it remains defensible.

That approach is slower than copying a number. It is also what makes research reusable.

How the Source Ladder Supports SEO, AEO, AIO, GEO, and LLM Visibility

The Source Ladder is primarily an editorial-quality system, but it also strengthens discoverability because it produces pages with clearer evidence, better entity relationships, and less commodity explanation.

Google's generative-AI search guidance says the same SEO fundamentals continue to matter, emphasizes unique and useful non-commodity content, and recommends a clear technical structure. It explicitly says special AI markup, artificial content chunking, and inauthentic mentions are not required.

For Swoopr, that means the visibility strategy should not be "write for bots." It should be:

• answer the actual question quickly;
• make claims traceable;
• show the original source;
• explain transformations;
• preserve dates and definitions;
• publish tools and methods that another writer can verify;
• create genuinely reference-worthy assets.

That is useful to a human reader first. It also gives search and answer systems cleaner evidence to work with.

The Five-Minute Source Check

Before trusting a financial claim, run this short process.

Step 1: Classify the claim

Is it a rule, a reported company fact, a market datum, a methodology, or an interpretation?

Step 2: Move up the ladder

Can you replace a secondary source with the authority, filing, or official dataset?

Step 3: Check the date

Is the source current for the period the sentence describes?

Step 4: Check the definition

Does the source define the metric the same way the sentence uses it?

Step 5: Record the limitation

What does the source not prove?

If those five answers are clear, the claim is usually ready to use.

Common Mistakes

Mistake: "Government source" automatically means "best source"

A government site may explain a product generally, while the issuer's prospectus contains the exact fee or structure being discussed. Match authority to claim.

Mistake: Citing a current value for a historical decision

Today's revised economic data can contain information that was not available at the time. Use historical vintages when the research question asks what someone knew then.

Mistake: Treating a formula as universal

State the exact formula and inputs. Two tools can use the same metric name and different definitions.

Mistake: Copying a source list without connecting claims

A long bibliography does not compensate for a sentence whose evidence is unclear.

Mistake: Using an AI answer as the final source

Use AI to accelerate discovery, organization, and comparison. Verify the material claim against the underlying evidence.

Mistake: Confusing registration with endorsement

A registry can confirm registration information. It does not certify future performance or remove the need for due diligence.

Frequently Asked Questions

Related Swoopr Resources

• Swoopr Research Workbench
• Research Methodology and Formula Definitions
• Evidence & Validation
• Citation & Sources Policy
• Research Notes Library
• SEC EDGAR Filing Search
• U.S. Investment Rule & Limit Tracker

References

• Google Search Central: Optimizing your website for generative AI features on Google Search
• SEC: How to Read a 10-K
• SEC: How Do I Use EDGAR?
• SEC: EDGAR Application Programming Interfaces (APIs)
• IRS: COLA increases for dollar limitations on benefits and contributions
• IRS: Retirement Topics: Contributions
• U.S. Treasury: Interest Rate Statistics
• U.S. Treasury: Daily Treasury Rates
• BLS: Consumer Price Index Archived News Releases
• Federal Reserve Bank of St. Louis: FRED API
• Federal Reserve Bank of St. Louis: ALFRED
• FINRA: BrokerCheck
• Investor.gov: Check Out Your Investment Professional
• SEC: Investment Adviser Public Disclosure
• Swoopr Investment: Citation & Sources Policy
• Swoopr Investment: Evidence & Validation