Methodology

CargoNet coverage and reporting

CargoNet receives incident information from sources including law enforcement, insurers, transportation companies, manufacturers, retailers, and theft victims. Reporting is non-exhaustive, coverage may change over time, and figures may be revised as delayed reports arrive. The data should not be interpreted as a census or nationally representative estimate of every cargo-theft incident.

Supply-chain crime events and confirmed cargo thefts are separate concepts. The site uses the labels “CargoNet-recorded supply-chain crime events,” “Confirmed cargo thefts reported to CargoNet,” “Estimated losses from confirmed cargo thefts,” and “Average value per confirmed cargo theft.”

Experimental inputs & weights

The quarantined indicator used three CargoNet-derived components with fixed weights:

Experimental formula

For each displayed quarter, the experimental indicator was calculated in four steps:

1. Anchor to a baseline. Each component is compared against a fixed 2024 baseline (quarterly averages, listed below). The baseline is the comparison point.
2. Normalize to a 0–100 scale. Every component is rescaled so that a value equal to its baseline maps to 50:normalized = clamp((value / baseline) × 50, 0, 100)
3. Take the weighted sum. The normalized components are combined using the weights above.
4. Round and band. The weighted sum is rounded to a whole number, then mapped to a severity band.

Baselines

The baselines are 2024 quarterly averages. If a component matches its baseline, it normalizes to 50.

^† CargoNet does not publish 2024 quarterly losses, so the loss baseline uses the FY2023 quarterly average ($331.9M ÷ 4). The events and avg-value baselines use FY2024 figures (3,607 events ÷ 4; $202,364 average value per theft).

Experimental severity bands

The quarantined rounded score mapped to one of four unvalidated bands. Ranges are inclusive.

• Low0–35
• Elevated36–60
• High61–80
• Severe81–100

Caveats & limitations

Read the Index with these limits in mind:

• A baseline-equal quarter reads as Elevated. Because each component at baseline normalizes to 50, a quarter that exactly matches the baseline sums to roughly 50. That falls inside the Elevated band, not at a neutral midpoint. This is intentional: 2024 was a record-high cargo-theft year, so matching it is still elevated.
• The top end is compressed. The linear normalization clamps each component at 100 once it reaches twice its baseline. A merely bad quarter and a record-shattering quarter can therefore both register at the ceiling, so the Index understates how extreme the worst quarters truly are.
• Sources publish on different cadences. CargoNet/Verisk publish on rolling quarterly and annual schedules. Some displayed loss figures are publisher estimates, and some per-theft averages are derived from figures within the same reporting period. No quarterly observation is substituted, interpolated, or inferred from a different reporting period.
• It is a young series. The Index currently spans only a handful of published quarters, one of them a reporting gap. Quarter-to-quarter moves are noisy, and the bands are best read as broad signals rather than precise rankings. Confidence grows as more quarters publish.
• One loss figure is a lower bound.Q2 2025 losses were reported as “more than $128M”; that lower bound is scored and plotted at face value, so if anything it understates that quarter.
• QoQ requires adjacent calendar quarters. Because the displayed series lacks a standalone Q4 2025 observation, the calculated Q1 2026 QoQ delta is null. A comparison with Q3 2025 is labeled “vs Q3 2025,” never QoQ. CargoNet separately reports a 12.2% event-count decline from Q4 2025; that fact is directly attributed and is not reconstructed from this site’s series.

Update cadence & reporting gaps

The site follows CargoNet/Verisk's quarterly and annual release cadence. Annual releases remain annual observations and are never treated as quarters. Formula, weight, baseline, or band changes to the experimental indicator add a dated changelog entry.

• Q4 2025: No standalone CargoNet release. Shown as a gap, not derived.

Freight-volume normalization

Freight-volume normalization is not currently published because the available CargoNet numerator covers the United States and Canada, while the BTS Freight Transportation Services Index covers the United States only. Reintroduction will require geographically matched data and methodological review.

Changelog

Every change to the methodology is recorded here, newest first, so the Index is a fixed reference rather than a moving target.

1. v1.32026-07-18

   Aligned publication language and removed unsupported context claims.

   • Clarified that current quarterly outputs keep incompatible reporting systems separate and do not publish a composite score.
   • Clarified that publisher-estimated losses and same-period derived averages are never substituted across reporting periods.
   • Removed the unsupported flagged-entity and dated annual-loss context claims and their non-direct source links.
2. v1.22026-07-16

   Corrected geographic scope and quarantined unsupported derived indicators.

   • Corrected CargoNet geography throughout the site from U.S.-only to United States and Canada.
   • Added a visible disclosure that CargoNet is a non-exhaustive incident-reporting database, not a census or nationally representative estimate.
   • Revised the frequency/severity interpretation: reported frequency remains a problem alongside increasing loss severity.
   • Removed the CargoNet ÷ BTS freight-volume normalization because its numerator and denominator cover different geographies.
   • Restricted QoQ labels to adjacent calendar quarters, left Q4 2025 null, and labeled Q1 2026 comparisons with Q3 2025 explicitly.
   • Removed the experimental composite score, gauge, bands, delta, and score citation from public outputs pending methodological review.
3. v1.12026-06-15

   Re-weighted the Index for real (cited) data and switched to quarterly cadence.

   • Cadence: quarterly (CargoNet/Verisk publish quarterly + annual). Q4 2025 has no standalone release and is shown as a gap, never derived.
   • Components re-weighted to what is solidly sourced: incident volume 40%, financial loss 40%, severity (avg value-per-theft) 20%.
   • Dropped strategic-fraud share and flagged-entity growth from the computed score (no clean time series / single proprietary source); both now shown as attributed context.
   • Baseline anchored to 2024 (events) and avg value/theft (FY2024); the loss baseline uses the FY2023 quarterly average because CargoNet does not publish 2024 quarterly losses.
4. v1.02026-06-01

   Initial Freight Fraud Index methodology published.

   • Four weighted components: incident volume (35%), financial loss (35%), strategic-fraud share (20%), flagged-entity growth (10%).
   • Baseline period = 2024 monthly average; each component normalized so baseline maps to 50.
   • Severity bands: Low 0–35, Elevated 36–60, High 61–80, Severe 81–100.