The Evidence Record

The boundary is not truth. Both kinds of claim may be entirely accurate. The distinction is accessibility: whether the supporting evidence exists in a form a stranger — or a machine — can locate and assess unaided.

The two kinds

Independently verifiable. A published methodology, a registration record, a number with a stated derivation, a public artifact. Checkable without contact.

Verifiable only through the company. Customer outcomes under confidentiality, results shared in a call, evidence held in materials that require a request. Frequently the stronger evidence in substance, and inaccessible in the circumstances where no conversation occurs.

Why the distinction has become material

Historically the second category was sufficient, because serious evaluations included a conversation in which proof was delivered.

Two changes reduced that sufficiency. Buyers complete more evaluation before contacting vendors, narrowing the field before any vendor is aware a cycle has begun. And machine evaluation performs part of that narrowing, working only from locatable material — a claim without accessible supporting evidence is held with lower confidence and repeated less readily.

The asymmetry

Claims that are easiest to publish independently — feature lists, integration counts, certifications, service levels — are typically the least differentiating, because competitors can publish equivalents.

The strongest differentiating claims are commonly the least accessible, held under confidentiality or delivered conversationally.

The consequence is that independent assessment tends to render a company legible in precisely the dimensions where it is ordinary.

Remediation

Publish method rather than result. Where a specific outcome cannot be shared, the approach that produced it usually can. A documented method is independently verifiable, differentiating, and creates no confidentiality exposure.

Attach derivation to every figure. A number without stated sample, method, date and scope is an assertion rather than evidence. A smaller figure with a published derivation is more useful than a larger one without.

Convert deliberately. Some conversation-dependent proof can be made independent through approval, anonymisation or publication of internal analysis.

Test

For each claim on a primary surface: what would a third party examine to verify this without making contact? Claims without an answer function only in the presence of a conversation.