Convoy reduces theft and double brokering to less than 0.001% of loads
Convoy News, Shippers, Tech & Visibility • Published on September 7, 2023
Real-time fraud detection system continually monitors carrier risk and proactively blocks fraud before it occurs.
Today we’re excited to announce a new real-time fraud detection system that has reduced cargo theft, double brokering, and other forms of fraud in our network by 90%, even as the industry experienced a 41% increase¹. This new system, which we’ve been testing with customers since January, combines machine learning technology, forensic behavioral data, automation, and industry collaboration to monitor risk in real-time, proactively block fraud before it can occur, and more quickly apprehend perpetrators. As a result of this work, we reduced the rate of theft and double brokering in our network to less than 0.001% across hundreds of thousands of shipments in Q2 2023. And at the time of this announcement, we haven’t experienced a single incident of theft in the second half of 2023.
A growing industry-wide problem
Every year, cargo theft costs shippers hundreds of millions of dollars—in 2022, the total value of stolen freight within the US was nearly a quarter of a billion dollars according to CargoNet. This industry-wide problem has only accelerated in recent years, with cargo theft up 57% year-over-year and trailer theft up 17% for the second quarter of 2023. Attacks have also become more sophisticated, with thieves using technology to target high-value loads and often working as part of larger cargo theft rings.
Traditional solutions to fraud prevention fall short in several ways:
1. Over-reliance on onboarding
Vetting carriers during onboarding is a necessary step to protecting against fraud. However, risk at onboarding is a single static data point, while a carrier’s risk profile and behavior can change over time. Effectively evaluating risk requires a more durable solution that relies on real-time data to continuously verify carrier legitimacy.
2. Failure to identify fraud networks
Traditional theft mitigation is akin to a game of whack-a-mole, in which individual actors are pursued for prosecution, but their connections to larger fraud rings can’t be established. A more systemic approach is required to identify these networks and proactively block fraudulent activity across all known members.
3. Insufficient driver data
Fictitious pickups, in which criminals use false identities to pose as legitimate carriers, were up 600% in 2022. Yet most solutions, including prominent industry data sources, lack information about individual drivers, which can be critical in securing check-ins at facilities and mitigating impact when adverse events occur.
4. Impartial assessments
Traditional risk assessments are often black-and-white decisions based on broad blanket policies and arbitrary data that have little correlation with cargo fraud. This results in frequent false positives that stifle the growth of legitimate carriers by blocking them from working with established brokerages. Instead, these carriers often resort to working with less credible brokers, exposing them to non-payments or misdirection schemes.
The problem is made worse by the fact that most fraud prevention efforts today rely primarily on manual process and human analysis, which can’t scale to effectively protect a freight network.
Introducing the industry’s first real-time fraud detection system
Over the last six months, our team has been working in collaboration with Fortune 1000 shippers and industry partners to pilot a unique real-time fraud detection system that addresses these shortcomings. At the core is machine learning technology that continually verifies the trustworthiness of carriers and identifies relationships between carriers, along with a dedicated team that provides decision oversight and input into the model.
Before a carrier can join Convoy’s network, and once they’ve satisfied mandatory insurance, operating authority, and FMCSA carrier contact verification, they’re run through a proprietary risk assessment model. Borrowing concepts from graph theory and fuzzy logic, and using publicly available and proprietary data, this clustering model identifies latent connections between carriers and their personnel and then makes recommendations based on the risk level of those links. Identifying these connections is critical to preventing fraud since perpetrators often operate in networks or through the use of multiple identities.
Carriers flagged as high risk are immediately blocked from bidding on loads within the Convoy app. Those flagged with potential risk are asked to submit additional information to prove their legitimacy. And all carriers flagged with any level of risk are reviewed by a member of Convoy’s Trust & Security team to make the final decision on network membership.
Of course, a carrier’s risk profile can change over time. Carriers who appear to be trustworthy during onboarding can quickly evolve into high-risk entities based on economic changes, sales of motor carrier companies, or onboarding of poorly vetted drivers. For this reason, our solution continually vets every carrier in our network multiple times per day before they can accept each load, using data from all interactions Convoy has with the carrier as well as publicly available information. Additionally, Convoy has built a multi-layered driver verification process to validate the accuracy of provided identification, with the capability to provide the validated assigned driver directly to facilities to verify before releasing a shipment.
Once a load has been assigned to a carrier, our systems track every step of the shipment lifecycle, looking for anomalies and proactively alerting our team to any suspicious activity. GPS tracking enables our systems to automatically determine if a shipment isn’t progressing to its destination as expected. Similarly, cargo sensors automatically alert us if a shipment is being unloaded in an unexpected location, enabling us to take immediate action.
Our solution is also an effective defense against the increasing threat of double brokering. At onboarding and throughout the course of their membership in Convoy’s network, carriers are evaluated for attributes typically associated with double brokering, behavior in accessing and navigating Convoy portals, and other attempts to mask their identity. Then, as carriers book shipments, our system evaluates their schedules for practicality. If deemed infeasible or high risk, loads are automatically removed from the carrier.
Ensuring a positive, fair experience for carriers
In building a modern fraud prevention solution, we knew it was also critical to provide the tens of thousands of legitimate small carriers and owner-operators in our network, including those with new MCs, with a system that doesn’t obstruct the growth of their business. This is in contrast to many traditional solutions, which are often indiscriminate and haphazard in their approach to carrier vetting—using broad policies and arbitrary data to block by default rather than using precise, up-to-date forensic and behavioral data to block only when necessary.
For example, a well-known tactic in the industry is to use DOT inspection count as a deciding factor in whether to block a carrier from accessing their shipments. This unfairly penalizes new carriers entering the industry, as the median time for a small carrier to receive their first inspection is 154 days (more than five months). In addition, Convoy data indicates that just 56% of single-truck carriers are inspected each year, with 75% inspected over a two-year period. Blocking these small carriers based solely on an issue like inspection count unnecessarily stifles the growth of their business and exposes them to increased risks of unscrupulous broker behavior, as noted above.
Our approach was to build a solution that proactively monitors risk in real time, combining the automated decision making model described above with a dedicated team trained to provide technology oversight and to make fair and consistent decisions. We began by modeling behavioral details and other latent attributes that are more directly correlated with fraud than generic data points like inspection counts and location. We then employed custom heuristics to fine tune our recommendation engine in order to minimize the number of false positives.
To further minimize false positives and other similar issues, we designed a human-in-the-loop process where an audit team reviews our model’s decisions when there’s a higher degree of uncertainty. The team follows a thorough standard operating procedure to request additional information from carriers and make a decision on each case. These decisions are then fed back into the model to continually improve future recommendations.
Through these efforts, we’ve created the industry’s first always-on fraud prevention system that significantly improves cargo security without negatively impacting the tens of thousands of hard-working carriers in our network.
The critical role of industry collaboration
The development of our model, mitigation, and reaction techniques wouldn’t have been possible without the trend analysis, data, and products from companies including CargoNet, Central Analysis Bureau, Persona, and others. These organizations exemplify how technology innovation can be used to reduce losses from fraud and foster a greater sense of security and trust within our industry.
In addition, law enforcement continues to play a critical role in the investigation, apprehension, and prosecution of bad actors—in the last six months, they’ve tracked down and apprehend members of multiple cargo theft rings across the US. Additionally, law enforcement adjacent professionals, industry groups, and our private investigator partners continue to provide incredible support to the development of industry-wide solutions to combat fraud.
As we continue to navigate the increasingly complex landscape of fraud and cargo theft, we look forward to expanding these relationships, which we believe can ultimately benefit every shipper, broker, and carrier in our industry.
If you’re interested in learning more about how our fraud prevention solution can help your transportation team reduce supply chain risk, drop me a line at firstname.lastname@example.org or contact your Convoy account manager.
¹ Convoy internal data comparing H1 2023 vs. H2 2022, CargoNet Quarterly Supply Chain Risk Trends Analysis 2022, 2023