Snapshot taken 22 July 2026

The state of US financial provider evidence

A July 2026 catalogue audit measures what ServeAssess had verified, what remained incomplete and why missing fields describe the research backlog rather than the market.

ServeAssess study13,574Dated sample
13,574provider records
54,049linked evidence sources
20service categories in the audited snapshot
500words in this edition
The evidence gap

Identity was much easier to verify than provider detail.

The two bars use different, published thresholds and should not be read as provider quality scores.

Verification score of at least 7575.6%Legal identity and official evidence were often available.
Research completeness of at least 403.2%Pricing, customer fit and comparable outcomes were usually incomplete.
54,049linked evidence sources
65.0%records with a matched official website
2.8category assignments per provider
47.5%records in the geography repair backlog
The geography result measures a data backlog, not market access.
Reading rule

Every percentage and rate belongs to the sample, threshold and period shown beside it. A descriptive relationship is not presented as causation.

01
Evidence led analysis

A verified identity is the beginning of the record

The audit found that 75.6 percent of records reached a verification score of at least 75. Only 3.2 percent reached a research completeness score of at least 40 under the model used for the snapshot.

Those measures answer different questions. A regulator record can establish that an entity exists or reports financial data, while price, customer eligibility and operational outcomes may remain unknown.

02
Evidence led analysis

Category counts overlap

The catalogue contained 38,010 category assignments across 13,574 providers, an average of 2.8 assignments per record. Category totals cannot be added to estimate unique providers because one legal entity may offer several services.

The overlap also creates an entity resolution problem. A brand may use a bank partner for one product and a separate lender or administrator for another.

03
Evidence led analysis

Website matching improved usable detail

The audited export matched an official website for 65.0 percent of records. Records with a matched site tended to contain more customer facing detail, although the association does not prove that a provider was more transparent or performed better.

Provider pages remain lower trust evidence for claims that require regulator or government confirmation. They are still useful for current product terms, eligibility and access channels when the observation date is retained.

04
Evidence led analysis

Missing geography described our backlog

The snapshot placed 47.5 percent of records in a headquarters state repair backlog. That figure measured unresolved catalogue fields and should not be interpreted as a claim about where United States providers are located.

A state page is therefore a research lens. It does not guarantee product availability or permission for a resident or business in that state.

05
Evidence led analysis

What changed after the audit

Rating Model v3 separated verification strength from research completeness and added entity roles, product permissions and peer group outcome dimensions. Missing outcome data now produces an insufficient data result instead of borrowing meaning from source volume.

The snapshot remains a dated record of the earlier catalogue. Later provider and category totals should not be substituted into its historical denominators.

06
Evidence led analysis

How the audit counted evidence

The export counted public provider records once by organization and retained overlapping category assignments separately. Linked sources were counted at the profile level, so one official record could support several claim level facts without becoming several independent sources.

Verification and completeness bands were descriptive thresholds from the assessment model used on the snapshot date. They were not estimates of market quality, customer satisfaction or regulatory risk.

07
Evidence led analysis

Reproducibility and later updates

The report linked its public export and stated the observation date so the aggregate figures could be reconstructed. A later export can differ because records are added, merged, archived or enriched after the snapshot.

Future audits should publish a new dated edition rather than overwrite this denominator. That keeps improvements in collection visible without changing the historical result.

Source trail

Open the underlying records

Every result on this page belongs to the dated sample above. Later catalogue updates do not silently change its denominator.