Live · as of 2026-08-23
30d $7.16M60d $17.28M90d $30.35M
Analysis · Dataset & Method

How the early warning is produced

The system runs on a bundled operational dataset covering the full network. Detection is deterministic: no black-box model sits between the operational record and the exposure number an executive is asked to act on.

Operating history
548 days
2025-02-22 → 2026-08-23
Records
399K
14 tables · 36.8 MB CSV
Network
9 sites
220 trucks · 240 drivers · 420 lanes
Detected slippages
60
$30.35M of 90-day exposure

Detection pipeline

Five deterministic stages from raw event to owned action

01

Windowed comparison

For each metric and each entity, a 21-day rolling window is compared with the preceding 60-day baseline. A deviation is only recorded when it clears the metric's noise floor, so ordinary variance is not reported as risk.

02

Trend projection

A least-squares slope over the recent window is extrapolated 30, 60 and 90 days forward. The projection is deterministic and reproducible — no model weights, no randomness, and the same input always yields the same forecast.

03

Financial translation

Each metric carries its own unit economics: fuel gallons and price, detention rate per hour, loaded driver hour, expedite premium, contractual penalty rate, or margin per load. The projected gap is multiplied through that formula to produce exposure in dollars.

04

Causal attribution

Candidate causes are tested with lagged correlation at the same site. The strongest leading indicator becomes the root cause, and the ordered sequence becomes a causal chain across modules.

05

Routing and action

The responsibility matrix maps every metric to exactly one accountable role, which receives the slippage together with the single intervention that addresses the identified cause.

Source tables

Bundled CSV extracts served with the application

TableRowsSizeContents
dim_locations.csv91 KBTerminals, distribution centers and warehouses in the network.
dim_customers.csv807 KBCustomer master with SLA targets, penalty terms and contracted revenue.
dim_trucks.csv22013 KBTractor and trailer master with domicile and preventive-maintenance interval.
dim_drivers.csv24013 KBDriver master with domicile, hire date and employment type.
dim_lanes.csv42019 KBOrigin/destination lanes with distance, target rate and mode.
dim_owners.csv121 KBAccountable owners and the responsibility matrix used to route slippages.
fact_loads.csv84,80114.6 MBLoad-level shipments with revenue, cost components, miles, transit and OTIF outcome.
fact_dock_events.csv111,97111.6 MBInbound and outbound dock events with appointment, wait, load and detention minutes.
fact_fuel.csv87,9224.6 MBDaily per-truck miles, gallons, MPG, idle hours and fuel cost.
fact_maintenance.csv3,515249 KBPreventive and unplanned work orders with downtime and repair cost.
fact_driver_hours.csv92,9384.6 MBDaily driver drive, on-duty, idle, detention and overtime hours.
fact_warehouse_ops.csv3,288328 KBDaily warehouse operating metrics per distribution centre and warehouse.
fact_sla_events.csv6,841493 KBService exceptions, SLA breaches, claims and penalties by customer.
fact_cost_ledger.csv6,318297 KBDaily cost leakage postings by category with preventable share.

Metric coverage

Monitored metrics per surveillance module

Trucking & Fleet

9 metrics monitored
  • Fleet Utilizationpct
  • Empty Mile %pct
  • Fleet MPGmpg
  • Idle Time Sharepct
  • PM Compliancepct
  • Unplanned Downtimehrs
  • Driver Overtime Sharepct
  • Driver Detention Hourshrs
  • Lane Margin %pct

Warehouse & Distribution

11 metrics monitored
  • Inventory Accuracypct
  • Order Backlogcount
  • Pick Cycle Timemin
  • Pick Accuracypct
  • Pack Cycle Timemin
  • Order Cycle Timehrs
  • Units / Labor Hourratio
  • Warehouse Overtime Sharepct
  • Space Utilizationpct
  • Damage Ratepct
  • Fulfillment Ratepct

Dock & Loading

9 metrics monitored
  • Average Dock Dwellmin
  • Wait Before Dockmin
  • Load Timemin
  • Unload Timemin
  • Appointment Compliancepct
  • Dock Door Utilizationpct
  • Detention Minutes / Truckmin
  • Load Ready on Arrivalpct
  • Units / Dock Labor Hourratio

Delivery & Customer SLA

7 metrics monitored
  • OTIFpct
  • On-Time Deliverypct
  • Average Delivery Delaymin
  • Failed Delivery Ratepct
  • POD Turnaroundhrs
  • Claims Valueusd
  • Customer Complaintscount

Cost & Margin Leakage

0 metrics monitored

    Interpreting exposure

    What the numbers do and do not claim

    Exposure is conditional

    Every dollar figure answers one question: what does this cost if the current trend continues unchecked for the stated horizon? It is not a booked loss and not a forecast of the P&L.

    Avoidable share is explicit

    Each slippage carries a preventable percentage reflecting how much of the exposure responds to operational intervention rather than market or structural conditions such as fuel price.

    Severity is relative

    Severity is assigned by rank across the current board, so the list stays triaged. A watch item is still real exposure — it is simply behind twelve larger ones.