veripath/cold-audit-protocol.md

# DOP Cold Audit Protocol — Live Evidence Standard
**Purpose:** Produce verifiable, first-hand visibility evidence for a target business to replace Tier-3 competitor self-reports in GTM materials.

## Signals measured (each scored 0–5, human-verified)
1. **GBP completeness** — categories, hours, phone, attributes, posts, Q&A, review count/rating. (observed on Google Business Profile)
2. **Foursquare claim status** — is the listing claimed and correct? (ChatGPT draws ~70% of local data from Foursquare)
3. **Website schema** — LocalBusiness / Service JSON-LD present and correct (service types, area, hours).
4. **AI-answer presence** — does ChatGPT / Perplexity name the business for 3 high-intent queries?
5. **Citation consistency** — NAP (name/address/phone) across GBP, Foursquare, Yelp, Facebook, BBB, industry directories.

## Evidence tiering (honesty standard)
- **Tier-1** — observed directly on a business-owned asset (GBP, the business's own website schema). Highest trust.
- **Tier-2** — observed via an independent third party but verifiable (Foursquare claim status, AI-answer screenshot, directory NAP). Medium trust.
- **Tier-3** — vendor or business self-reported dashboard numbers. **Excluded from DOP proof.** Used only as wedge framing in outreach.

## Method per signal
- GBP: browser inspect the live profile; record completeness gaps.
- Foursquare: query the business; note claimed/unclaimed + wrong fields.
- Schema: `web_extract` the site; check for JSON-LD `LocalBusiness`/`Service`.
- AI-answer: run 3 high-intent queries (e.g. "emergency plumber open now near <city>"); capture whether the business appears.
- NAP: cross-check 5+ directories for consistency.

## Output per target
One markdown file: target name, metro, queries tested, observed results, gaps, per-signal score, tier breakdown, evidence references (URLs/screenshots). No self-reported numbers presented as fact.

## Commit path
`research/cold-audits/<vertical>-<business-slug>.md` — authored as `tonyjbala`.