Trust · The full storyNo black boxes

Methodology.

Every number you see in ClinicalRate carries its receipts. This is the full provenance chain from raw placement event to the percentile rendered in your browser.

Book a data demo →Read the chain

9 classesSources

Every row, every stepAudit

24 hoursRefresh

Empirical · winsorizedStatistical method

The chain By the numbers Pitfalls we refuse What you can verify vs Survey methodology Use cases What's included FAQ

By the numbers

How defensibility shows up at each step.

9Source classesPublic feeds, partner contributions, VMS ingests, public-domain.

0.92Confidence floorRecords below threshold route to human review, never silently bucketed.

1.5×IQRWinsorizationCrisis-rate outliers visible at p90 without bending the median.

n≥30Distribution floorBelow floor, individual percentiles suppress and "thin market" surfaces.

The chain

From placement event to percentile, end to end.

01SOURCE

Permitted feeds, partner contributions, VMS ingests.9 source classes · per-record provenance

Nine source classes — public job-board feeds where re-use is permitted, contributing staffing agencies under data-share agreements, direct ingest from supported VMS platforms, and the curated public domain (BLS, state filings, public RFPs).

02INGEST

Schema reconciliation.Idempotent · de-duped · hashed

Each source lands in a typed staging table with source ID, ingestion timestamp, and content hash. Duplicate detection runs before normalization — a single placement contributed by both an agency and a VMS is reconciled, not double-counted.

03NORMALIZE

Resolve to canonical role.428 roles · ≥0.92 threshold

Every record passes through the Taxonomy Mapper before it enters a distribution. Records that don't resolve at confidence ≥ 0.92 are routed to a human review queue, not silently bucketed.

04STATISTICS

Empirical percentiles, winsorized.p10 → p90 · 1.5×IQR · n≥30

Linear-interpolated empirical percentiles on the trailing 30-day window, winsorized at 1.5×IQR to suppress single-record outliers. Never a mean masquerading as a median. Thin markets (n < 30) suppress individual percentiles and surface a "thin" indicator instead.

05SERVE

Provenance in every response.UI · API · CSV · all surfaces

Every UI cell and every API response carries as_of, sample_size, and the source-class distribution. Defensible by design — the buyer can always ask "where did this number come from" and you can answer in one click.

What we don't do

Statistical pitfalls the platform refuses.

PitfallMean dressed as median

Survey reports sometimes publish a "median" derived from a parametric fit. ClinicalRate uses empirical percentiles on the raw data — interpolation only, never assumption.

PitfallMixed-cohort distributions

"RN" is not a role. The platform refuses to render a distribution that mixes ICU and Med/Surg, or travel and per diem, into one chart. Every distribution is a single canonical cohort.

PitfallStale data dressed as fresh

Every cell carries its as_of timestamp. If a source feed lags, the affected distributions are flagged in the UI rather than served as current.

PitfallOutliers dragging the median

Crisis-rate placements ($300/hr Travel RN bookings during 2020) are real, but rare. Winsorization at 1.5×IQR keeps them visible at the 90th percentile without bending the 50th.

What you can verify

Open audits, on request.

Audit surface	What you see	Tier
Sample size on every cell	n events behind every percentile, in UI and API.	All tiers
As-of timestamp on every cell	Most recent ingest contributing to the cell.	All tiers
Source-class distribution	What share of the cell came from each of the nine source classes.	All tiers
Per-record provenance trace	Down to the individual record's source, ingest timestamp, and normalization confidence.	Enterprise
Annual methodology audit by external firm	Report shared under NDA.	Enterprise

Comparison

ClinicalRate methodology vs. survey methodology.

The decisions that distinguish a defensible distribution from a polished number.

Capability	ClinicalRateEmpirical · winsorized · per-cell provenance	Conventional surveySelf-reported · annualized
Source	Reconciled placement events from 9 source classes	Self-reported employer/agency questionnaire
Refresh	Every 24 hours	Annual
Statistical method	Empirical percentiles, linear interpolation	Mean, parametric quartile fit
Cohort handling	Travel / per-diem / local separated	Often blended into one role row
Outlier handling	Winsorized at 1.5×IQR · visible at p90	Top-coded or removed silently
Per-cell provenance	Sample size + as-of + source mix	Methodology appendix
Thin-market handling	Suppress percentile, surface indicator	Render anyway, no warning

Enterprise tier · per-record provenance trace

Karen ParkHead of Internal Audit · Hospital system, 9 facilities

“The audit was three meetings, not three months. Every cell carried its receipts and our auditor signed off without a methodology dispute.”

Where it shows up

Three teams who need methodology to be public.

Agencies

Pricing leadership

When the rate-card meeting becomes a methodology meeting, the number defends itself.

Per-cellSample size + as-of Hospitals & facilities

Internal audit & finance

Every benchmark drillable to the source-class mix. Auditor sign-off in meetings, not months.

5 yrTrace retention (Ent.)AI systems

Explainability

A pricing agent can cite the n, the as-of, and the source mix in the same response. No "where did this come from" tickets.

Per-rowProvenance envelope

What's included

Methodology access by tier.

ClinicalRate · All plans

From $1,000/ month

Provenance is a platform default, not an upgrade. Sample size and as-of on every cell, on every plan. Per-record trace and external audit on Enterprise.

Sample size on every cell (UI + API + CSV)
As-of timestamp on every cell
"Thin market" indicator below n=30
Source-class distribution on every plan
Per-record provenance trace (Enterprise)
Annual external methodology audit (Enterprise)
Versioned canonical taxonomy with changelog
No PHI, no patient data, ever

Book a data demo →See full pricing

FAQ

Questions buyers ask about methodology.

What's actually in the 'placement event' that drives a percentile?

A placement event is a contracted role × geography × rate × contract-class instance. We dedupe across source classes (a placement contributed by both an agency and a VMS lands once), retain only fields needed for the empirical distribution, and never ingest PHI or candidate-identifying fields.

How do you treat crisis rates (e.g. 2020-style spikes)?

They are real — and they belong at the 90th percentile, not the median. Winsorization at 1.5×IQR keeps them visible in the upper tail without dragging the central tendency. We never silently delete data points.

What changes when the canonical taxonomy is updated?

New roles are additive. When an existing canonical role is split or merged, we publish a versioned changelog with effective date, surface a banner on affected views, and keep the prior version queryable for 12 months for trend continuity.

Are there things ClinicalRate refuses to publish?

Yes. We do not render distributions on cohorts below n=30; we do not publish individual employer/agency rates; we do not surface PHI under any contract. Every refusal is encoded in the platform, not a policy.

Other features

The rest of the surface.

Get started

Numbers with receipts.

Book a demo. Bring the question your CFO asked you that you couldn't answer — we'll show you where the answer lives in the chain.

Book a data demo →