--- title: "Source tiers explained: roles, not credibility grades | Len P. van der Hof" description: A source tier records how a document entered your research, not how credible it is. One vendor case study, three claims, and a test for your own sources. image: "https://lenvanderhof.com/media/generated/blog-hero-source-tier-is-a-role-not-a-grade.bdab3ed73b99.wide.webp" --- [Venture Building](https://lenvanderhof.com/en/blog/category/venture-building/) Research guide # A source tier is a role, not a grade The case study stays the same. Its support changes when the sentence in your memo changes. Len P. van der HofPublished 29 October 202610 min read Sorting a source does not weigh its support for a claim. Direct answer A source tier records how a source entered an inquiry, using one of four labels: Primary, Secondary, Advocacy or Grey. It is a role, not a truth ranking or a fixed evidence weight. In TRACE, the method in The Verified Research Loop, you keep one source ID per source body and judge each claim-source relation separately. A vendor case study can directly show what the vendor said, give limited support for a customer result it reports, and fail to support a forecast for your own team. The document is unchanged; the claim and what it requires have changed. ## Key takeaways - Primary, Secondary, Advocacy and Grey are source roles. Tier 1 is not a credibility grade, and a tier is never a verdict. - Check the body and version you actually obtained before judging what it says about a claim. - Keep one source ID for one body; write a separate relation for each proposition it is used to support. - For each relation, record its support ceiling, prohibited inference and missing check. **A source tier records how a source entered your inquiry; it does not grade every claim you might draw from it.** A vendor case study can directly show what the vendor said, give limited support for a customer result it reports, and fail to support a forecast for your own team. The document has not changed. The proposition has. Treat “Tier 1” or “peer-reviewed” as a grade and you let a label decide what a source proves. TRACE sets the weight per claim instead. This is the Assess step of [TRACE](https://lenvanderhof.com/frameworks/trace/), the research method in [The Verified Research Loop](https://lenvanderhof.com/books/the-verified-research-loop/) (introduced in [What is TRACE?](https://lenvanderhof.com/en/blog/what-is-trace/)). Retrieve records which substantive body and version you obtained: the document itself, not a search result about it. Assess asks what that body can bear for one precise proposition. Cross-check then gives each claim-source pair its own relation. A search snippet or a tier number cannot do any of those jobs for you. ## What does a source tier describe? TRACE keeps four tier labels as a short way to say how a body entered the inquiry: how it stands relative to an event, measurement, argument or interested party. Each label tells you which question to ask next, not what the answer must be. LabelRole in the inquiryA question it promptsTier 1: PrimaryClose to an event, measurement, transaction, rule or original analysis: a system log, dataset, contract, filing or study report.Is this the right event, version, measure or authority for my claim?Tier 2: SecondaryReports on, interprets, compares or synthesizes other bodies: a review, investigative article or analyst report.Which bodies and methods sit underneath the account?Tier 3: AdvocacyProduced to advance an interested party’s product, policy, position or reputation. It may also contain original data.What was selected, omitted or framed for that purpose?Tier 4: GreyRelevant material outside the other three roles: a working paper, technical report, forum post or unattributed slide, for example.Can I inspect its body, provenance and limits? The numbers do not run from best to worst. A primary record can be false, partial or stale. An advocacy body can copy a public dataset with care. A strong synthesis can explain mixed results better than one original paper. A grey lead can point you to the controlling authority. Chapter 5’s composite worked record even tags a commercial market report as Secondary while noting that its publisher sells advisory services. Neither role settles whether its Europe-wide estimate can carry a claim about demand in one Netherlands B2B segment. The appraisal does: the scope is too broad, key method inputs are undisclosed, and the report cannot determine that claim. Record the visible role or roles without awarding a general trust score. Then test the body against the sentence you want to write. ## Why does the proposition change the weight? Because a document can answer only the questions its contents, method and scope allow. A firm’s annual filing directly shows what the firm reported, if you have the right filing and table. It may tell you little about whether the firm’s outlook will come true, or whether the reported number covers everything you care about. The body stays the same. The burden of the proposition changes. Chapter 5 separates three judgments that are easy to compress into one label: 1. **Source role:** How was this body positioned in relation to the event, data, argument or campaign? 2. **Body appraisal:** Is the correct, sufficiently complete and usable version available, and which production controls or limitations can you inspect? 3. **Claim-relative appraisal:** For this exact proposition, how direct and applicable is the body? Where do its method, scope, uncertainty, dependence or currentness limit use? Only the first judgment fits in a tier label. A label can route your attention; it cannot set a fixed weight for the whole body. A working link does not answer the second question. Before any claim is judged, Chapter 5 gives the body a preliminary state: `Assessable` when its identity, relevant content, version and permitted use are enough to begin, or `Pending appraisal` when the body is present but a method, authority, language, conflict or currentness check remains. The other states are `Insufficient body`, `Restricted`, `Out of scope` and `Superseded or withdrawn`. “Peer-reviewed” or “official” cannot answer the third question. Keeping the three judgments apart stops a source label from quietly becoming a claim verdict. ## Worked example: one vendor case study, three claims **Hypothetical example.** A software vendor publishes a case study saying that a featured retailer’s median first-response time fell from six hours before a six-week pilot to two hours during it. Your team is considering the tool for its own support queue. The vendor, retailer, figures and pilot are invented. No customer records were inspected. Give the case study one source ID, `S-CASE-01-v1`, and record its title, publisher, version and exact paragraph. It is sales material, so its role in the registry is **Tier 3: Advocacy**. Keep that source record stable while you test three propositions: Proposition and relationWhat this body showsSupport ceiling and next checkC1: The vendor stated that the retailer’s median first-response time fell from six to two hours. R-CASE-01-v1: Supports.The vendor made that statement, if the page is authentic and the version and paragraph are preserved.It does not establish that the retailer measured the change correctly. Recheck for a correction or a changed page.C2: The retailer’s median first-response time fell from six to two hours across the stated periods. R-CASE-02-v1: Supports, limited to the vendor’s account; Pending appraisal.The case study reports this result. In this example it supplies no customer record.Check the retailer’s ticket data, its definition of first response, the before and during windows, exclusions and other changes during the pilot. Until then, do not call the outcome verified.C3: Our team’s median first-response time will be lower during a six-week rollout than in the six weeks before it, if we buy the tool. R-CASE-03-v1: Contextualizes; it does not support the forecast.The case suggests questions about setup and measurement for a local test.The retailer’s result, even if verified, cannot predict this team’s queue, staffing or outcome. The C1 relation is direct for a narrow reporting act. C2 asks whether the reported change happened as described. C3 adds a transfer problem and a prediction problem. One body, three claims, three different limits. Do not relabel `S-CASE-01-v1` as Tier 1 because it directly shows what the vendor said. Chapter 5’s rule is plain: do not rename a source when its role changes with the claim; add a relation record. Note the direct role **for C1** in that relation, and leave the body’s advocacy role as it is. If the vendor revises the page, record a new source version and recheck each relation. [Case study vs evidence](https://lenvanderhof.com/en/blog/case-study-vs-evidence/) examines what one success can tell you. The question here is narrower: what a tier label cannot decide for three uses of one source. ## What goes in a claim-relative appraisal? The book’s appraisal record has eight parts, filled in for one body against one proposition. It is not a score. 1. **Body fidelity:** Do you have the right body, with enough of it to check the claim, rather than a snippet or a generated summary? Can another reader reopen the exact passage in the same version? 2. **Method or authority:** For an empirical source, how were the people, cases, measures and comparisons chosen? For a rule, which instrument, issuer, version and date in force? For an operating record, how was the event logged? 3. **Scope and directness:** Do the firm, group, place, period, comparison and outcome match the sentence you intend to write? If you must transfer the finding, state the bridge and how it might fail. 4. **Result and uncertainty:** What was actually reported? Keep units, base counts, ranges, nulls and material qualifiers, not only the attractive headline. 5. **Incentives and selective reporting:** Who paid for, wrote, chose or benefits from the account? Which cases or outcomes might be missing? An interest tells you where to look; it does not prove the content false. 6. **Provenance and dependence:** What does the body share upstream with other records: data, study, announcement, method or organization? Different publishers can still trace back to one origin. 7. **Integrity and currentness:** Has the body been corrected, withdrawn, superseded or overtaken for this use? What event should trigger a recheck? 8. **Limits on use:** State the narrowest claim this relation can support and one inference it cannot carry. For C2, a complete URL and a polished quotation cannot supply the retailer’s measurement method, so the appraisal stays pending. For C3, even full access to the retailer’s records would leave the transfer to your team open. The two gaps call for different checks. ## Does “official” or “peer-reviewed” settle the claim? No. Chapter 5 puts it as a warning: “A tier is never a verdict.” Neither “Tier 1” nor “official,” “peer-reviewed” or “primary” makes a source true, sufficient, independent, current or decision-ready. Those labels name features worth recording. An official filing may be the record of what an organization submitted. Peer review is one step in how a paper gets made; it does not grant access to the raw data or catch every error. Neither label shows that a number was measured well, that no correction followed, or that the studied group matches your decision. Apply the proposition test. “The regulator issued this text” needs the issuer and the operative version. “This text permits our proposed data flow” also needs scope, the date in force and a qualified reading of how it applies to the actual flow. “This finding applies to our customers” needs the study’s group, measure and a defensible transfer. Where interpretation needs a specialist, leave the relation open for that review. A label routes inspection. It cannot repair a missing body, a mismatched group or an old version, and neither can a total score: a sum can hide a fatal flaw. ## What happens after appraising one source? Assess leaves you with the source’s role and a bounded appraisal for each material claim. Cross-check then builds a claim capsule: the proposition with its entity, period, measure, result, uncertainty and intended use still attached. It keeps one source per relation row. If another body bears on C2, give it its own relation and compare what each body actually measured. Do not merge two IDs into one cell called “support.” That comparison has its own trap. [Corroboration is not a source count](https://lenvanderhof.com/en/blog/corroboration-is-not-a-count/) asks whether several apparently agreeing sources come from distinct evidence-generating paths. The question here comes one step earlier: **what can this one body support for this one claim?** Get that wrong and every later comparison inherits the error. Neither an appraisal record nor a relationship summary authorizes action by itself. For a decision-critical claim that is unresolved or in material conflict, the book’s default disposition is `HOLD`. The decision owner may then record another disposition, such as `Specialist review`, `Pilot` or `Accepted risk`, but an unqualified `Proceed` waits until the evidence state changes. Risk acceptance changes the action record; it does not strengthen the evidence. ## Try this today Take one source behind a sentence in a memo, proposal or research note you are working on now. Open the substantive body. If you have only a snippet or summary, record that limit and pick a body you can inspect for this exercise. Write down its title, version and exact locator. Name its tier as a role, without assigning a weight. Copy the claim from your text, then write two *other* propositions someone might cite the same body for. Make one row per claim-source relation: ``` Source ID / version: Tier and visible role: Body and exact locator: Claim 1: Shows: Cannot establish: Claim 2: Shows: Cannot establish: Claim 3: Shows: Cannot establish: Missing check for each claim: Owner / next review: ``` For each row, state what the body directly shows, what you would have to assume to go further, and the inference it cannot carry. If a method, the relevant group or the current version is missing, mark the appraisal `Pending appraisal` and name the next check. You are done when another reader can explain from your record why the three relations differ. The exercise tests the clarity of your record. It does not show that the source is true or that a decision is ready. Cite this:A source tier is a role, not a grade.Len P. van der Hof. [https://lenvanderhof.com/en/blog/source-tier-is-a-role-not-a-grade/](https://lenvanderhof.com/en/blog/source-tier-is-a-role-not-a-grade/) · Published 29 October 2026. ## Sources 1. [The Verified Research Loop](https://lenvanderhof.com/books/the-verified-research-loop/) 2. [TRACE framework](https://lenvanderhof.com/frameworks/trace/) ## Further reading - [What is TRACE?](https://lenvanderhof.com/en/blog/what-is-trace/) - [Corroboration is not a source count](https://lenvanderhof.com/en/blog/corroboration-is-not-a-count/) - [Case study vs evidence](https://lenvanderhof.com/en/blog/case-study-vs-evidence/) - [TRACE framework](https://lenvanderhof.com/frameworks/trace/) - [The Verified Research Loop](https://lenvanderhof.com/books/the-verified-research-loop/) About the author ## [Len P. van der Hof](https://lenvanderhof.com/en/authors/len-p-van-der-hof/) Entrepreneur, AI Innovator and Venture Builder Len P. van der Hof builds practical AI systems, digital ventures and evidence-informed tools for founders. ```json { "@context": "https://schema.org", "@graph": [ { "@type": "Person", "@id": "https://lenvanderhof.com/#person", "name": "Len P. van der Hof", "alternateName": [ "Len van der Hof", "L.P. van der Hof", "Leendert Pieter van der Hof" ], "honorificSuffix": "MSc", "url": "https://lenvanderhof.com/", "image": [ "https://lenvanderhof.com/photos/len-portrait-1.jpg", "https://lenvanderhof.com/photos/len-portrait-2.jpg", "https://lenvanderhof.com/photos/len-portrait-3.jpg", "https://lenvanderhof.com/photos/len-portrait-4.jpg", "https://lenvanderhof.com/photos/len-speaking.jpg", "https://lenvanderhof.com/photos/len-hero.jpg" ], "jobTitle": "Entrepreneur, AI Innovator and Venture Builder", "description": "Len P. van der Hof, MSc, is a Dutch entrepreneur and AI innovator in Zwijndrecht. He builds ReasonKit, MindSesh, Undominated.ai, books under his name, the fiction imprint LPH98.lifestyle, and technology ventures through LPH98.ventures. Eleven titles in Systems for the Strategic Self are available now, in English and Dutch.", "address": { "@type": "PostalAddress", "addressLocality": "Zwijndrecht", "addressCountry": "NL" }, "alumniOf": { "@type": "CollegeOrUniversity", "name": "Rotterdam School of Management, Erasmus University" }, "knowsAbout": [ "Artificial intelligence", "AI agents", "Agentic AI systems", "LLM routing", "SEO", "Generative engine optimization", "Answer engine optimization", "Venture building", "Founder performance", "Founder psychology", "Evidence-based decision-making" ], "sameAs": [ "https://www.linkedin.com/in/lenvanderhof/", "https://x.com/LenvanderHof", "https://www.youtube.com/channel/UCTG20buKqYYbitqqf7l3zJA", "https://www.instagram.com/Lenvanderhof/", "https://www.threads.com/@lenvanderhof", "https://github.com/Lenvanderhof", "https://huggingface.co/LPH98", "https://www.npmjs.com/~lenvanderhof", "https://www.goodreads.com/author/show/70983905.Len_P_van_der_Hof", "https://www.amazon.com/author/lenvanderhof", "https://www.bol.com/nl/nl/b/len-p-van-der-hof-msc/609879394/", "https://bsky.app/profile/lenvanderhof.com", "https://mastodon.social/@Lenvanderhof", "https://crates.io/users/Lenvanderhof", "https://cursor.com/@Lenvanderhof", "https://medium.com/@Lenvanderhof", "https://gitlab.com/Lenvanderhof", "https://hub.docker.com/u/lenvanderhof/", "https://dev.to/lenvanderhof", "https://www.facebook.com/Lenvanderhof", "https://soundcloud.com/Lenvanderhof" ], "affiliation": [ { "@id": "https://lenvanderhof.com/#publisher" }, { "@id": "https://lenvanderhof.com/#mindsesh" }, { "@id": "https://lenvanderhof.com/#lifestyle" } ] }, { "@type": "WebSite", "@id": "https://lenvanderhof.com/#website", "url": "https://lenvanderhof.com/", "name": "Len P. van der Hof", "description": "Len P. van der Hof, MSc, is a Dutch entrepreneur and AI innovator in Zwijndrecht. He builds ReasonKit, MindSesh, Undominated.ai, books under his name, the fiction imprint LPH98.lifestyle, and technology ventures through LPH98.ventures. Eleven titles in Systems for the Strategic Self are available now, in English and Dutch.", "inLanguage": [ "en", "nl" ], "publisher": { "@id": "https://lenvanderhof.com/#person" } }, { "@type": "Organization", "@id": "https://lenvanderhof.com/#publisher", "name": "LPH98.ventures", "url": "https://lph98.ventures", "founder": { "@id": "https://lenvanderhof.com/#person" } }, { "@type": "Organization", "@id": "https://lenvanderhof.com/#mindsesh", "name": "MindSesh", "url": "https://mindsesh.net", "founder": { "@id": "https://lenvanderhof.com/#person" } }, { "@type": "SoftwareApplication", "@id": "https://lenvanderhof.com/#reasonkit", "name": "ReasonKit", "url": "https://reasonkit.sh", "creator": { "@id": "https://lenvanderhof.com/#person" } }, { "@type": "SoftwareApplication", "@id": "https://lenvanderhof.com/#undominated", "name": "Undominated.ai", "url": "https://undominated.ai", "creator": { "@id": "https://lenvanderhof.com/#person" } }, { "@type": "Organization", "@id": "https://lenvanderhof.com/#lifestyle", "name": "LPH98.lifestyle", "url": "https://lph98.lifestyle", "founder": { "@id": "https://lenvanderhof.com/#person" } }, { "@type": "ImageObject", "@id": "https://lenvanderhof.com/en/blog/source-tier-is-a-role-not-a-grade/#primaryimage", "url": "https://lenvanderhof.com/media/generated/blog-hero-source-tier-is-a-role-not-a-grade.bdab3ed73b99.wide.webp", "contentUrl": "https://lenvanderhof.com/media/generated/blog-hero-source-tier-is-a-role-not-a-grade.bdab3ed73b99.wide.webp", "representativeOfPage": true }, { "@type": "BreadcrumbList", "@id": "https://lenvanderhof.com/en/blog/source-tier-is-a-role-not-a-grade/#breadcrumb", "itemListElement": [ { "@type": "ListItem", "position": 1, "name": "Home", "item": "https://lenvanderhof.com/" }, { "@type": "ListItem", "position": 2, "name": "Blog", "item": "https://lenvanderhof.com/en/blog/" }, { "@type": "ListItem", "position": 3, "name": "Venture Building", "item": "https://lenvanderhof.com/en/blog/category/venture-building/" }, { "@type": "ListItem", "position": 4, "name": "A source tier is a role, not a grade", "item": "https://lenvanderhof.com/en/blog/source-tier-is-a-role-not-a-grade/" } ] }, { "@type": "WebPage", "@id": "https://lenvanderhof.com/en/blog/source-tier-is-a-role-not-a-grade/#webpage", "url": "https://lenvanderhof.com/en/blog/source-tier-is-a-role-not-a-grade/", "name": "A source tier is a role, not a grade", "description": "A source tier records how a document entered your research, not how credible it is. One vendor case study, three claims, and a test for your own sources.", "isPartOf": { "@id": "https://lenvanderhof.com/#website" }, "primaryImageOfPage": { "@id": "https://lenvanderhof.com/en/blog/source-tier-is-a-role-not-a-grade/#primaryimage" }, "breadcrumb": { "@id": "https://lenvanderhof.com/en/blog/source-tier-is-a-role-not-a-grade/#breadcrumb" }, "inLanguage": "en-GB" }, { "@type": "BlogPosting", "@id": "https://lenvanderhof.com/en/blog/source-tier-is-a-role-not-a-grade/#article", "mainEntityOfPage": { "@id": "https://lenvanderhof.com/en/blog/source-tier-is-a-role-not-a-grade/#webpage" }, "headline": "A source tier is a role, not a grade", "description": "A source tier records how a document entered your research, not how credible it is. One vendor case study, three claims, and a test for your own sources.", "datePublished": "2026-10-29T20:36:00.000Z", "author": { "@id": "https://lenvanderhof.com/#person" }, "publisher": { "@id": "https://lenvanderhof.com/#person" }, "image": [ "https://lenvanderhof.com/media/generated/blog-hero-source-tier-is-a-role-not-a-grade.bdab3ed73b99.square.webp", "https://lenvanderhof.com/media/generated/blog-hero-source-tier-is-a-role-not-a-grade.bdab3ed73b99.landscape.webp", "https://lenvanderhof.com/media/generated/blog-hero-source-tier-is-a-role-not-a-grade.bdab3ed73b99.wide.webp" ], "articleSection": "Venture Building", "inLanguage": "en-GB" } ] } ```