Original research · Evidence transparency · 2026
How well documented are Minnesota agent profiles?
A reproducible audit of exact sources, state-license identifiers, public review aggregates and evidence confidence across 182 curated profiles.
182
profiles audited
The full published directory cohort
28.6%
with exact sources
171 dated records in total
31.3%
license ID coverage
A stored number, not a status guarantee
51.1%
review coverage
A recorded public aggregate
Published · Evidence measured as of · Dataset version 2026-08-25
Analysis: Minnesota Top Agents Editorial Team
Executive finding
Transparency is measurable—and still incomplete
Only 52 of 182 profiles currently contain at least one exact field-level source record. The other 130 profiles may contain public facts, but the supporting URLs have not yet been stored claim by claim in the dataset.
License and review documentation are more common, but neither is complete. 57 profiles publish a recorded Minnesota license number and 93 include a public review aggregate. Missing values are treated as unknown—not as evidence that a license, review history or underlying source does not exist.
The evidence-confidence result makes that gap visible: 130 profiles currently fall in the low documentation tier, while 24 meet every high-confidence prerequisite. These labels evaluate this publisher's documentation, never the agent's professional quality.
Claim-level evidence map
What the stored citations actually support
A profile can have multiple sources and still leave important claims uncited. These five categories classify the support text attached to every exact source record. Only 38 profiles currently have citations across all five areas.
Category mapping is deterministic and published with the evidence ledger. One source can support more than one category, so source record counts overlap.
Identity & credentials
Professional identity, brokerage, office, license or team roster.
122 supporting source records
Public reviews
A dated public rating, review count or review-platform snapshot.
79 supporting source records
Transaction activity
Sales, listings, transactions, volume, price range or closing activity.
112 supporting source records
Market & specialty focus
Service areas, local markets, property types or client specialties.
115 supporting source records
Experience & recognition
Experience, tenure, designations, awards or professional recognition.
82 supporting source records
Source independence
Who published the evidence?
Official and independent sources generally provide stronger corroboration. First-party sources remain useful for identity, service and attributed claims when labeled as such.
Observation freshness
How recently were sources checked?
Freshness is calculated from each source's observation date, not from a generic page publication date. Dynamic facts should still be rechecked before an agent is hired.
Coverage audit
Four distinct documentation signals
Each percentage uses all 182 profiles as its denominator. Signals overlap, so they should not be added together.
28.6%
52 of 182
At least one exact source
Dated field-level record with a URL and publisher.
31.3%
57 of 182
License number recorded
Minnesota license identifier stored; verify current status with Commerce.
51.1%
93 of 182
Review aggregate recorded
Public rating/count pair observed during review.
13.2%
24 of 182
High confidence
All v1.0 high-confidence minimums met.
License × reviews
Coverage is not the same thing as overlap
A profile is most directly checkable when both a license identifier and a dated public review aggregate are present. This table shows where those two fields intersect; it does not evaluate the license status or review sentiment.
Verify a specific agent| Recorded fields | Profiles | Share |
|---|---|---|
| License + review | 34 | 18.7% |
| License only | 23 | 12.6% |
| Review only | 59 | 32.4% |
| Neither recorded | 66 | 36.3% |
Evidence confidence
A confidence score for the documentation—not the agent
Version 1.0 allocates up to 100 points across source traceability, independence, breadth, core records and freshness. The cohort median is 20/100; the mean is 36/100.
A high result also requires at least two exact sources, an independent or official source, coverage of four required evidence categories, a recorded license number and a public review aggregate. The numeric rating itself does not affect confidence.
Inspect the complete confidence model →High documentation confidence
Meets the model's minimums for exact sourcing, independence, breadth, license and review documentation.
Medium documentation confidence
Some meaningful documentation is present, but one or more high-confidence minimums are missing.
Low documentation confidence
The stored evidence is not yet complete enough for medium or high documentation confidence.
Source density
Most profiles still lack claim-level citations
An exact source record stores a URL, publisher, observation date, verification type and the fields it supports. Counting those records exposes documentation depth without judging what the evidence says about the agent.
Reproducible method
How every number was calculated
The study uses only fields published in the downloadable dataset. No missing value was inferred, and no profile was removed because its documentation was incomplete.
01
Population
All 182 rows in dataset version 2026-08-25; one row per published profile.
02
Exact-source coverage
Rows where evidence_source_count is greater than zero. That field counts unique canonical URLs per profile; the normalized ledger has one row per evidence/claim record, so a URL may recur under distinct verification classes.
03
License and reviews
Nonblank state_license_number and public_review_count fields, counted separately and as intersections.
04
Confidence tiers
Published v1.0 levels: high at 80–100, medium at 45–79 and low below 45, subject to the high-tier prerequisites.
05
Percentages
Category count ÷ 182 × 100, rounded to one decimal. Counts remain the primary result.
06
Category, source and freshness audit
Each ledger row is mapped to five published evidence categories, one source type and a freshness bucket from its observation date.
07
Revision behavior
The study regenerates from the current public rows. Its as-of date and confidence calculations advance with dataset version 2026-08-25.
Model basis
v1.0: source traceability, source independence, evidence breadth, license/review documentation and evidence-review recency; independent of editorial score.
Interpretation guardrails
What this study cannot establish
- It is not a census of Minnesota licensees. Results describe this curated directory cohort only and cannot be generalized statewide.
- It does not measure agent quality, ethics, transaction volume, closing outcomes, client fit or current license status.
- A blank field means the item was not stored in this dataset edition; it does not prove the item is absent in the world.
- Exact-source coverage measures traceability, not whether a claim is favorable, complete or permanently current.
- Public review aggregates are point-in-time observations. They can change, differ by platform and may represent an individual or a team.
- The publisher created both the profiles and this audit. Reproducible downloads reduce—but do not eliminate—that conflict, so readers should inspect the data directly.
Why publish the gap?
The audit is a roadmap, not a victory lap
The most useful next step is not to hide incomplete profiles. It is to add official identifiers, dated third-party evidence and claim-level citations until more consumers can distinguish a documented fact from an editorial assertion.
- 1. Capture identity first. Add Minnesota license identifiers, then attach official Commerce records where available.
- 2. Cite the material claims. Store dated URLs for reviews, production context, specialties and service areas.
- 3. Recheck the evidence. Refresh changing facts and preserve visible observation dates.
Suggested citation
Minnesota Top Agents. “Minnesota Agent Evidence Transparency Study 2026.” Published 2026-07-20; analysis of 182 published profiles, dataset version 2026-08-25.
Questions about the study
Does low evidence confidence mean an agent is low quality?
No. Evidence confidence measures how completely this site has documented a profile with exact, dated public sources. It does not measure agent skill, ethics, production or client fit. A low-confidence profile may simply be under-documented.
Does a blank license number mean an agent is unlicensed?
No. It means Minnesota Top Agents has not yet stored a Minnesota license number in that profile. A recorded number is an identifier, not a current-status guarantee; consumers should confirm it through the Minnesota Department of Commerce before hiring an agent.
Is this a census of every Minnesota real-estate agent?
No. The study describes the 182 profiles in Minnesota Top Agents' curated editorial dataset. It cannot estimate documentation coverage for all Minnesota licensees or for the industry statewide.
Can the calculations be reproduced?
Yes. Download the profile table and normalized evidence ledger, use the dataset version shown in this report, and calculate the published profile-, category-, source-type- and freshness-level counts. Percentages are rounded to one decimal place and every exact citation retains its URL and observation date.