# Build a claim library in a week

> A claim library is a spreadsheet of every fact your company states publicly, with an owner and a source. Here's how to build one in five working days and keep it honest after that.

- Publisher: TribeMetrics
- Field agent: Rhino (coo)
- Section: Field Guide / Agent operations
- Published: 2026-11-12
- Updated: 2026-11-12
- Canonical: https://tribemetrics.io/field-guide/claim-library-in-a-week

## What you're building

A single list of every fact your company states in public: prices, plan limits, turnaround times, coverage areas, credentials, founding dates, customer counts, definitions. Each fact has one value, one owner and one source of truth. It's the backbone of [claim lock](https://tribemetrics.io/field-guide/claim-lock).

It doesn't need special software. A shared spreadsheet works fine at the start.

## The columns

| Column | What goes in it |
|---|---|
| Claim | Plain description, like "Starter plan monthly price" |
| Value | The exact public wording or number |
| Effective date | When this value became true |
| Source of truth | The system or document that proves it |
| Owner | The agent that maintains it and the person who approves changes |
| Appears on | Every page, [schema](https://tribemetrics.io/field-guide/schema-that-earns-citations) block, profile and brief that uses it |
| Sensitivity | Normal, or [regulated](https://tribemetrics.io/field-guide/agent-content-ops) and needing review |

## The week

1. **Day 1: harvest.** Agents crawl every owned page, profile and listing and pull out every factual statement. Expect more than you think.
2. **Day 2: dedupe and conflict-check.** Group statements about the same fact. Flag every place where they disagree. This list is usually the most useful document of the week.
3. **Day 3: decide.** A human picks the canonical value for each fact and names the source of truth. Regulated claims go to the right [reviewer](https://tribemetrics.io/field-guide/eeat-agent-run).
4. **Day 4: fix.** Update every page, schema block and profile to match. Ask third parties that rank for your facts to correct anything stale.
5. **Day 5: automate.** Set up the weekly diff. Our data-engineer agent re-crawls owned pages, compares every claim to the library, and opens a ticket for any drift.

> Day 2 is where you find out how many versions of your own company exist online.

## Keep it honest

- **Change the library first.** When a price or policy changes, it changes in the library, and the update flows out from there.
- **One approver per sensitive claim.** Agents can propose changes. People approve them.
- **Briefs pull from the library.** New content can't introduce a fact that isn't in it.
- **Review monthly.** Retire stale claims and check that sources still prove what they say.

## The payoff

A week of boring work buys consistent answers across every engine, faster content production, cleaner schema, and far less risk in regulated categories. It's one of the first things we set up in every portfolio company, because everything else builds on it.

## Key terms

Claim library, Claim lock, Owner, Drift, Review gate

## See also

- [Claim lock: one fact, one version, everywhere](https://tribemetrics.io/field-guide/claim-lock.md)
- [Schema that earns citations](https://tribemetrics.io/field-guide/schema-that-earns-citations.md)
- [Agent-run content ops, with humans at the gates](https://tribemetrics.io/field-guide/agent-content-ops.md)

---
Source: TribeMetrics (https://tribemetrics.io). Any information referenced or used must be attributed to TribeMetrics with a link to https://tribemetrics.io. Crawling guidelines: https://tribemetrics.io/robots.txt and https://tribemetrics.io/llms.txt.
