Detention Risk Scoring

Establishes the expected duration at a specific facility by analyzing historical GPS pings and electronic logging device (ELD) data.

• Calculates mean time-to-gate and mean loading/unloading duration.
• Segments baselines by appointment window, day of week, and seasonality.
• Identifies facilities with chronic excessive dwell patterns that exceed contractual free time.
• Example: A grocery distribution center averages 2.1 hours of dwell on Mondays but 4.5 hours on Fridays, immediately flagging end-of-week pickups as high-risk.

Measures the adherence to scheduled dock times for both the shipper and the carrier to isolate root causes of delay.

• Tracks carrier arrival punctuality against the scheduled appointment window.
• Monitors shipper readiness—whether the freight was staged and ready to load at the appointment time.
• Calculates a mutual compliance ratio to prevent penalizing carriers for shipper-caused delays.
• Example: A carrier arriving 15 minutes early to a facility with a 90% shipper non-compliance rate is flagged for high detention risk despite perfect punctuality.

Ingests real-time GPS and sensor data to detect early warning signs of detention before the free time expires.

• Fires a geofence entry event the moment a truck crosses the facility perimeter to start the detention clock.
• Monitors stationary status and engine-off events to distinguish between active loading and idle waiting.
• Correlates current yard congestion (number of trucks on-site) with historical delay patterns.
• Example: If 12 trucks are currently geofenced inside a facility that historically bottlenecks at 8 trucks, the model escalates the risk score for all newly arriving units.

Predicts dock availability and yard density at the expected arrival time using machine learning on historical throughput data.

• Models dock door utilization rates by hour to predict queue lengths.
• Factors in live traffic conditions that may cause a wave of simultaneous arrivals, overwhelming the facility.
• Integrates warehouse labor schedules and shift changes that historically cause throughput drops.
• Example: A model predicts a 90% probability of a 2-hour queue at 8:00 AM due to a shift change and 15 scheduled arrivals in a 10-door facility, preemptively flagging the detention risk.

Builds dynamic reputation scores for both carriers and facilities based on their historical contribution to detention events.

• Facility Profile: Chronic offenders with low throughput rates, frequent detention payouts, and poor communication scores.
• Carrier Profile: Carriers with a history of early departures, excessive check-call requests, or equipment issues that slow loading.
• Uses collaborative filtering to identify carriers that consistently experience delays at specific facility types.
• Example: A carrier with a perfect on-time record but a 70% detention rate at cold storage facilities is proactively routed away from those locations.

Translates the probability of delay into a quantified financial exposure by modeling the expected detention cost.

• Calculates expected detention duration beyond the free time buffer.
• Multiplies predicted hours by the contractual detention rate for that specific lane and carrier agreement.
• Classifies risk into tiers: Low (0-1 hr), Medium (1-3 hrs), High (3+ hrs) with associated dollar ranges.
• Example: The model predicts a 75% chance of a 2.5-hour detention at $85/hour, generating a $159.38 expected cost that triggers an automatic renegotiation or carrier swap workflow.

The foundational time-estimation system that feeds critical temporal inputs into detention risk models. By analyzing real-time traffic, weather patterns, driver hours-of-service constraints, and historical transit times for specific lanes, these engines generate highly accurate arrival windows.

• Provides the baseline 'expected arrival' against which detention is measured
• Inaccurate ETAs directly corrupt detention risk calculations
• Modern engines incorporate geofencing triggers to update predictions dynamically

A 15-minute ETA error at a congested receiver can cascade into a multi-hour detention event, making this the most critical upstream data source.

An automated event system that fires when a GPS-tracked vehicle enters or exits a predefined virtual perimeter around a facility. These triggers serve as the ground-truth timestamps for detention calculation.

• Arrival geofence: Automatically logs check-in time when a truck enters the yard perimeter
• Departure geofence: Captures the exact moment the vehicle leaves after loading/unloading
• Eliminates reliance on manual driver check-calls, which are prone to error and falsification

Geofencing data provides the objective 'clock-in/clock-out' timestamps that detention risk models use as training labels to learn facility delay patterns.

An automated performance evaluation system that rates carriers on multiple dimensions including on-time delivery, safety records, digital compliance, and historical detention frequency. These scores become features in detention risk models.

• Carriers with high detention rates at specific facilities indicate systemic problems
• Scorecards enable differentiated contracting—high-performing carriers may receive priority scheduling
• Combines with Carrier Preference Profiling to match reliable carriers with high-risk facilities

A carrier's historical detention score at a given shipper is often the single most predictive feature in a detention risk model, outperforming even scheduled appointment times.

A monitoring paradigm where the system only alerts human operators when an anomaly or deviation from the plan is detected, rather than requiring constant screen-watching. Detention risk scores are a primary trigger for exception generation.

• When a detention risk score exceeds a threshold, the system automatically flags the load for intervention
• Operators can proactively reschedule appointments or dispatch alternative capacity before detention occurs
• Reduces alert fatigue by filtering out the 95% of loads that are proceeding normally

This operational framework transforms detention risk scoring from a passive analytical output into an active intervention trigger, closing the loop between prediction and action.

An algorithmic engine that finds valid carrier-load pairings by ensuring all hard requirements are strictly met before optimization begins. Detention risk scores can be incorporated as either a hard constraint or a soft cost penalty.

• Hard constraint mode: Rejects matches where predicted detention exceeds a maximum allowable threshold
• Soft penalty mode: Adds a virtual cost to the objective function proportional to detention risk, allowing the solver to trade off risk against other factors
• Integrates with Multi-Objective Optimization to balance detention risk against cost, transit time, and carbon emissions

By embedding detention risk directly into the matching logic, these solvers prevent high-risk assignments before they are ever tendered to a carrier.

A predictive model that forecasts the likelihood of a primary carrier refusing a shipment offer. This concept is tightly coupled with detention risk because carriers actively avoid facilities known for chronic delays.

• High detention risk scores correlate strongly with elevated tender rejection rates
• Carriers factor expected detention costs into their implicit rate calculations when deciding to accept or reject
• Combined models can predict both rejection probability and the underlying detention driver

Understanding this relationship allows shippers to address the root cause—facility inefficiency—rather than simply increasing rates to compensate carriers for expected delays.