How to Turn Fleet Data Noise Into Signal

The modern commercial vehicle is one of the most prolific data generators in the transportation industry. Between ELDs, cameras, trailer sensors, reefer monitors, engine diagnostics, and OEM telematics platforms, a single truck can produce tens of thousands of data points every day. Multiply that across a fleet of hundreds or thousands of vehicles, and the result is an unmanageable quantity of noise that most fleets can’t often use.

Loaded and Rolling host Thomas Wasson sat down with EROAD’s David Blackwell, ELD and Transport Product Manager, and Akshay Kumar, Senior Product Manager, to discuss why fleets with more data than ever are still struggling to act on it, and what the path forward actually looks like.

Blackwell, who is based in East Tennessee and has spent more than 20 years in telematics and product management working with enterprise carriers across North America, says the problem is straightforward. 

“No matter how much data people have access to, things aren’t changing on the ground,” Blackwell said. “You can have hundreds of alerts, dozens of metrics or KPIs. You can have data from trailers, cameras, OEMs; everything looks important and so you can’t really act on anything.”

That paradox is at the heart of what EROAD is trying to solve. According to Blackwell, the question that data alone cannot answer is the one that matters most to the person sitting in the back office, and that question is, “what do I actually need to do right now?”

“When you get your filtering right, understand what to look at, and get the right information to the right people, that’s when trucks and data stop shouting at you and start telling you something you can use,” Blackwell said. 

Kumar, who is based in Houston and joined EROAD four years ago after building back office systems for naval vessels and aviation operations, brings a mission-critical perspective to the conversation. In those environments, data had to be 100% actionable or it served no purpose. He’s applying that same philosophy to trucking.

“A single commercial vehicle can have tens of thousands of data points a day,” Kumar said.

The layers stack up fast, he explained. There’s the ELD data: speed, odometer, engine hours, location pings, and driver status events. Then safety data like harsh braking and cornering. Camera events for distracted driving or following distance violations. Trailer and cargo data, including reefer temperatures, door open-close events, and trailer mileage. Engine diagnostics. Route schedules. Stop-level execution data. The volume is staggering for a fleet with hundreds or thousands of assets.

“If you’re running an operation of thousands of trucks, you’re not just managing a fleet anymore,” Kumar said. “It feels like you’re managing a data center.”

The real issue, Kumar says, isn’t the data itself, but the systems and processes that are supposed to turn that data into decisions. Back office users have been inadvertently saddled with a role they were never trained for.

“Dispatchers are stuck trying to decide in a few seconds which of these 50,000 data points is a crisis, and which is background noise,” Kumar said.

Blackwell says that the most common and most counterproductive response to data overload is to implement more alerts. It’s a pattern he encounters constantly among the enterprise fleets he works with.

“So now to fix overwhelming data, fleets wind up overwhelming staff with alerts,” Blackwell said. “Now, dispatchers have too many notifications and people start ignoring those too.”

It’s a cycle that feeds itself. The fleet sets up email or text notifications for the things that seem most urgent, those alerts multiply until inboxes are flooded, and the people who were supposed to be acting on the information tune it all out. Everybody is getting notified, but nobody is listening.

Being more effective with fleet data, according to Blackwell and Kumar, means having the right signal reach the right person at the right time. 

“The hard part is knowing which part of the data conversation to pay attention to,” Kumar said.

No conversation about freight technology can avoid AI for long, but Blackwell says that EROAD is focused on practical applications rather than hype. He recommends that fleets start with what he considers the clearest use case for AI: predictive maintenance.

Predictive maintenance, he says, is not the same as a fault code triggering a repair order. It’s the ability to detect patterns in vehicle and trailer data (small drifts in temperature, voltage, or pressure) that indicate a breakdown is forming before any warning light comes on.

“AI shines in being able to pick up those patterns that really no person could pick up because of the amount of data,” Blackwell said. “The right AI tool can flag those issues early.”

Maintenance is the fourth-largest cost category for most fleets, and predictive maintenance has the potential to reduce that expense by 10 to 15% by shifting from calendar-based or reactive service schedules to data-driven, condition-based ones.

“What if I could tell you with 85 to 90 percent confidence that your trailer will break down in two weeks?” Blackwell asked. “That’s the kind of predictability we’re talking about.”

AI-powered temperature simulation for refrigerated trailers is another practical solution some fleets are starting to implement. Traditional reefer monitoring provides periodic temperature readings of the supply and return air inside the trailer. The AI layer goes further, predicting when a load will drift out of its required temperature range before it happens.

“We can now simulate what the box of frozen hash browns’ temperature is,” Blackwell said. “That’s the difference between making a change proactively and losing a load.”

That capability converts directly into margin protection when it comes to high-value, temperature-sensitive freight. A single rejected load can wipe out the profit from multiple successful deliveries.

Dispatch execution, according to Kumar, is where the financial stakes of poor visibility become especially clear. He broke down delivery operations into three core functions: routing, dispatch, and driving. All three, he says, have to be in sync for the system to work.

Without real-time visibility, dispatch teams are left reacting to problems after they’ve already cascaded into costly consequences. The margin of error on certain sensitive freight means any mistake can be massively costly.

“If a truck misses a site window, you aren’t just late,” Kumar said. “You might be throwing away four-thousand dollars’ worth of ready-mixed concrete, which is no longer compliant. Add that to the new load needed, idle crews waiting for the next truck to come, extended equipment rentals, delayed payment, and you know, the cost can easily exceed twice that of the original load,” he said.

EROAD’s answer to this is its Route Manager, which ingests planned routes from a fleet’s routing or optimization system and overlays them with live execution data. 

Dispatch tends to move from what happened to what needs attention, Kumar says. The goal with Route Manager is to give dispatch a single consolidated view of how routes are actually performing against plan and allow them to intervene early rather than perform post-mortems.

The value of telematics data increases exponentially when it’s connected to the other systems a fleet depends on, such as TMS platforms, maintenance management systems, and customer-facing tools. The integrations themselves aren’t new, but the maturity of what they enable has changed dramatically.

“When TMS job data and live telematics are integrated, I’m now able to give my customers an ETA of when their load will arrive,” Blackwell said. That kind of capability helps carriers meet SLAs and differentiate on customer experience.

Both Blackwell and Kumar advise fleet managers to start working on overwhelming data by doing less, not more.

Kumar recommended picking a single workflow that matters, whether that’s something like dispatch visibility or vehicle availability, and reducing it to the handful of signals that drive action. Get dispatch, maintenance, and drivers looking at the same picture.

“You can anchor that first effort around one of the four biggest fleet cost categories: payroll, fuel, equipment, or maintenance,” Blackwell said. Before changing anything, he says, set a baseline and give it 30 days.

“Just pick one metric, like unnecessary idling,” Blackwell said. 

Not all idle time is bad, but unproductive idling is a significant and measurable money burn. The same principle applies to reefer operations, where pre-cooling hours early wastes fuel that never needed to be spent.

“Make sure that you’ve got a system that allows you to have that data in a way that’s easily consumable and able to get it to the right person,” Blackwell said. 

Click here to learn more about EROAD.