The State of US Financial Service Providers 2026
ServeAssess examined identity, source depth, category coverage and selected outcome studies without turning unlike measures into one league table.
Official identity was often available. Decision ready detail was not.
Three quarters of the catalogue cleared a strong identity and evidence threshold. Only a small fraction cleared even a moderate completeness threshold. That gap is the main finding of this report.
Sources per provider
The average describes linked source depth. It does not mean four independent confirmations existed for every claim.
Services per provider
Category assignments overlap, so service totals cannot be added to estimate unique companies.
Official site matched
A matched site helped with product detail but did not replace regulator evidence.
Geography repair queue
This measured an unresolved data field, not a financial service desert.
More evidence did not automatically mean a more assessable provider.
The catalogue linked 54,049 sources to 13,574 provider records. That is about four sources per provider, but averages hide uneven coverage. A provider could have extensive official identity and financial records while publishing little about pricing, eligibility, ownership or ordinary customer outcomes.
ServeAssess therefore separates verification strength from research completeness. Missing customer or pricing evidence lowers completeness. It does not become a negative quality judgment.
A provider count is not the same as customer access.
The snapshot contained 38,010 service assignments across 20 categories. One legal entity may appear in banking, lending, payments and another service, while one consumer brand may rely on several legal entities.
State access needs a similarly careful definition. Headquarters, licensing, stated service area, physical branches and national online availability answer different questions. The 47.5 percent geography repair backlog could not support a state access ranking, so this edition does not publish one.
The best known institution is not automatically the best documented or the best fit.
JPMorgan Chase led a 69 bank group study by deposits at $2.79 trillion. KeyBank recorded 8.4 complaints per $1 billion of deposits among the defined large branch bank peer group, compared with 11.9 for PNC. PNC presented a stronger rounded documentation case across the fields used in that study.
Each result has a different denominator. Deposits measure scale. The adjusted complaint rate uses deposits as an imperfect exposure proxy. Documentation measures what could be verified. None of the three determines the right account for an individual customer.
Review collection method changed the meaning of a star average.
In a 16 profile pilot, seven profiles in the mostly unsolicited group averaged 1.33 out of 5. Six lending profiles with visible invitation programmes averaged 4.62. The observed difference was 3.29 points across 297,519 displayed reviews.
The groups were small and nonrandom. The result does not show that invitations caused higher ratings. It shows why collection method, product, timing and profile identity must be controlled before public ratings are compared.
Growth was uneven across the credit union sample.
About 68 percent of the largest credit unions in the sample added members, compared with about 32 percent of the smallest group. The result records adoption during the measured period. It does not establish customer satisfaction or explain why any institution gained or lost members.
A stronger future edition can compare growth within consistent asset bands while preserving mergers, charter changes and denominator dates.
A publication gate for research claims
The research centre will publish a ranking only when the identity, variable, denominator and observation period are defensible. Other ideas remain labelled as pilots or future studies.
Official records, dated sources and exact entity matching.
Source depth, field completeness and category overlap.
Collection method needs a larger controlled sample.
Search, advertising and news observations need a fixed period.
Parent and subsidiary links need broader verification.
Headquarters, licence and service availability must stay separate.
How to read this edition
- Fixed snapshot
Catalogue figures use the 22 July 2026 export. Later enrichment does not rewrite these results.
- Overlapping categories
Service assignments are not unique providers and cannot be added as a market total.
- Separate axes
Verification, completeness, reputation, visibility and outcomes are never blended into a hidden winner score.
- Explicit proxies
Deposits are labelled as a size proxy when customer or account counts are unavailable.
- Missing stays missing
An unavailable result is not converted into a zero or a negative provider judgment.
Open the underlying records
Every result on this page belongs to the dated sample above. Later catalogue updates do not silently change its denominator.