← Back to Product × B2 Methodology

B2 Classification: Evidence & Review

B2 classification is neither a first-match waterfall nor a point score. Evidence is gathered at the exact-order level, then applied in a documented hierarchy: direct entity proof first, hard order/product evidence next, and exact checkout evidence within defined conditions. Address intelligence exposes context or conflict; human review records the final, auditable classification.

Open the HandiTreads B2 evidence-review dashboard →

Documented evidence precedence Government proof overrides B2B Hard order signals carry real weight Address intelligence prompts review No numeric point scores Human decisions are preserved
B2 classification evidence and review model Direct buyer and order evidence is applied in a documented hierarchy: direct entity and government proof first, hard product and transaction evidence next, then exact checkout evidence within its conditions. It runs alongside residential and commercial address intelligence from USPS, Smarty, Google Maps, and approved search research. The evidence is compared, triaged, and resolved by documented human review before an auditable B2B, B2C, B2G, or Unknown result is retained. 1. COLLECT INDEPENDENT EVIDENCE 2. COMPARE, TRIAGE, AND REVIEW 3. PRESERVE THE AUDITED RESULT Parallel inputs, then a documented decision Strong evidence has priority; address research adds context or conflict. Neither weak signals nor address type alone can set the result. OPEN BOTH STREAMS COMPARE AT EXACT-ORDER GRAIN RESOLVE WITH HUMAN JUDGMENT RETAIN RESULT + BASIS STARTING RECORD Invoice plus exact linked order Keep invoice, checkout, buyer, and ship-to context together. EVIDENCE PRECEDENCE · HIGHEST TO LOWEST 1. DIRECT ENTITY / GOVERNMENT PROOF → B2G or B2B Exact reviewed buyer, official .gov identity, or conservative government-name review. 2. HARD ORDER / PRODUCT SIGNAL → B2B MH, CVR/Yard, hose-pipe-cable, or >$1K non-slip aluminum on one SO. 3. EXACT CHECKOUT EVIDENCE → B2G / B2B / B2C WITHIN ITS RULE ADDRESS + EXTERNAL INTELLIGENCE Context, corroboration, and conflict detection USPS DPV / Smarty RDI plus Google Maps and approved public search surface residential, commercial, public-facility, or inconclusive clues. It can strengthen or challenge the provisional result. It never outranks direct proof or automatically sets a label by itself. EVIDENCE SYNTHESIS + TRIAGE Apply precedence, then compare the full record Stronger evidence wins. Signals never average into a numeric score. Conflict, missing identity, or weak research moves the order to review. B2G overrides B2B. Consumer checkout alone cannot downgrade hard B2B evidence. MANUAL REVIEW + INTERVENTION Resolve the buyer, not just the delivery address Reviewer checks property type, business footprint, public operator, and the exact purchasing entity. They can confirm, refine, or override. They may retain Unknown with a recorded rationale and evidence source. FINAL, AUDITED CLASSIFICATION B2B / B2C / B2G / Unknown The result remains traceable to the exact order and its evidence basis. AUDIT + CONTROLLED REUSE Keep the basis with the result Reuse only exact confirmed evidence.

Starting record

Keep the invoice, exact linked checkout, buyer, and ship-to context together.

Evidence hierarchy, not a numeric score

  1. 1. Direct entity / government proof
    Exact reviewed buyer, official .gov identity, or conservative government-name review. B2G overrides B2B.
  2. 2. Hard order / product evidence
    MH, CVR/Yard, hose-pipe-cable, or >$1K non-slip aluminum on one SO gives B2B.
  3. 3. Exact checkout evidence
    Sets B2G, B2B, or B2C only within its defined conditions.

Independent evidence streams

The hierarchy sets the provisional classification. Address context is examined alongside it to corroborate, surface conflict, or leave the order unresolved.

Direct buyer + order evidenceApply the documented precedence above. Exact reviewed domains may be reused; generic work domains may not.
Address + external intelligenceUSPS DPV / Smarty RDI, Google Maps, and approved public search surface residential, commercial, public-facility, or inconclusive clues. This never automatically sets the label.

Evidence synthesis + manual intervention

Stronger evidence wins. Signals never average into a score. A reviewer resolves disagreements or missing identity by checking the property, business footprint, public operator, and actual purchasing entity. They can confirm, refine, override, or retain Unknown with a documented rationale.

Audited classificationB2B, B2C, B2G, or Unknown. Unknown remains valid when evidence is insufficient.
Controlled reuseKeep the run, source, address finding, and reviewer decision. Reuse only exact confirmed evidence, never a broad generic domain rule.

The method is deliberately granular: direct government/entity proof carries the most weight; defined hard order/product evidence carries more weight than checkout; address intelligence provides context; and documented human review can resolve an exception. That is how the model distinguishes a business buyer shipping home, a consumer using a work email, a public facility, and a genuinely unresolved record.

Precedence is explicit

Government/entity proof leads; hard transaction/product evidence follows; exact checkout fills defined gaps. B2G overrides B2B, and a consumer checkout indicator alone cannot downgrade hard B2B evidence.

Address evidence is useful, not simplistic

A residential delivery does not automatically negate a business buyer, and a commercial delivery does not automatically prove one. RDI, Maps, and search evidence identify context and contradictions worth resolving.

Human judgment is part of the method

Review is not cleanup after the real process. It is the control that lets the classification reflect the exact buyer and purchasing circumstance instead of a blunt surrogate.

Auditability prevents rule creep

Preserve the source, provisional rule, address finding, and reviewer rationale. A confirmed exact organization may inform future reviews; generic business-email logic may not.

Version control: historical runs keep their original rule and evidence basis. Address intelligence and reviewer decisions enrich the current review layer without silently rewriting historical facts.