How Real-Time Tracking and IoT Sensors Are Revolutionizing Fragile Goods Transportation
A shipment of wine may be delivered on time by a truck with a functioning GPS, only to find out upon delivery that the wine is ruined. The truck and GPS performed as intended, but the GPS would not have shown that the pallet had fallen off the truck several days ago at the rail yard, or that it had been sitting in the sun on a dock for four hours with nobody watching over the temperature. The issue is not that data was not captured. The issue is that the people who needed to know did not know that the data needed to be connected.
Tracking the truck is not tracking the product
Shippers of time-sensitive or delicate goods, say in the food or pharmaceuticals industry for example, have already adopted GPS tracking for their trailers if they are trucking the freight, or their containers if they are intermodal or ocean, sometimes down to the level of individual pallets. The bad news is that as one manufacturing executive put it, “If the product is going to be damaged in transport, it will arrive at my customer’s distribution center that way whether I knew where the truck was every minute of the journey or not”. The trailer was never “lost”, and the container holding the pallet that was punctured by the forklift operator in the warehouse was never “damaged”.
Wine is a particularly egregious example of this technology disconnect, as a case of bottles can easily survive a five day coast-to-coast road trip and then be ruined on a siding for six hours because the driver did not see a warning that the load had been exposed to excessive heat during the trip. The GPS got the truck to its destination on time. The wine is ruined and undeliverable. This is an extreme example, but the underlying problem is not an outlier.
The damage happens at the handoffs, not on the highway
The exposure to theft or damage is not static along a journey. Long-haul trucking as a transportation mode is actually one of the safest links in the supply chain, as the driver has an interest in smooth, steady, and predictable driving, and the cargo is undergoing few handling events. The exposure increases at the edges, at the handoffs between transportation modes. Whether that takes the form of truck to rail to port to cross dock to final delivery, the moment that your freight is transferred from one carrier’s possession to another’s is when things are most likely to go wrong. And at the edges, the digital visibility is often switched off. The trucking company’s TMS stops tracking the freight as it leaves its facility, and the rail company’s system does not start tracking it again until it is loaded onto the train. There is a period of time, potentially as short as minutes or as long as a day, in which the freight is not digitally visible to anybody, even as it is physically handled by somebody.
If you were to graph the time and place of a damage claim along a journey, there would be spikes at the transfer points between carriers. That is where attention should be paid.
GPS was never meant to be used indoors
Satellite tracking technology is excellent in places where it is allowed to be used, but the moment that your shipment is inside a steel box, under a ship’s hull, in a tunnel, or in a storage yard buried under other containers, it is often rendered useless. These are all situations that arise during the course of a multimodal shipment and during which GPS tracking is often ineffective or completely unavailable
Many leading shippers have understood this and are already combining GPS tracking with yard gate sensors that register a container’s passage through specific points, LPWAN networks such as LoRaWAN or Narrowband IoT that consume less power while still allowing for positional tracking through long periods with limited connectivity, and dead reckoning accelerometers that allow for some level of positional and event-based tracking even when GPS is unavailable. There is nothing particularly exotic about any of these technologies. The unifying factor is that they recognize the need to have multiple sources of information about a shipment’s condition and location available so that there are no blind spots when a shipment transitions from one connection point to another.
One sensor, four readings, and why that matters
The ability to detect four different environmental conditions with one sensor can make invisible forces visible, and it is more valuable than you might think.
Let us return to the wine shipment that was ruined by a forklift. By all accounts, the temperature was fine, so the cold chain was not at fault. There may have been a high-g event, but a temperature sensor alone would not have sounded an alarm. Some corks were loosened and some bottles were broken by the high-g movement and impact of the forklift dropping the pallet, so that while the temperature sensor registered a “normal” shipment, the multi-sensor fusion of temperature, high-g, tilt, and humidity would have shown a sudden spike corresponding to the exact moment at which the pallet was dropped by the forklift, which would in turn correspond to the loosening of the corks and the breakage of many of the bottles.
Sensor fusion is the difference between “I do not know what happened to my shipment” and “I know exactly what happened to my shipment”.
This is a legal issue masquerading as a technology issue
As a shipment transitions from one transportation mode to another, so too does legal ownership of the shipment, according to the terms of the bill of lading and the contract between the shipper and the carrier. And when a shipment is damaged, the first and most important order of business is establishing who was responsible for the shipment at the moment and location at which the damage occurred. If the sensor logs do not show it clearly, the dispute can take weeks to sort out, and potentially never be truly resolved.
This is where many of the best-in-class supply chains in the world still stumble.
The truck carrier has his logs, the rail carrier has her logs, and the warehouse has a third set. The three sets of logs do not talk to each other, and the visibility into the shipment they provide is fragmented. When a claim comes up, one of the three will have to take a closer look at the logs and make an educated guess as to what happened and where based on events recorded in the logs. It is not a technological issue, but a coordination one, with real costs in both time and money spent.
Why the fix is API-first, not another dashboard
When confronted with the prospect of losing visibility, many companies fall back on the same solution, which is to buy another dashboard. This is rarely the right answer. Buying a fifth TMS tracking tool does not make the handoff between TMSs any smoother, and only adds complexity to an operator’s working day.
What actually works is choreographing the events between TMSs, gate out, gate in, loaded, discharged, etc, using bidirectional APIs rather than a series of CSV exports and imports that get processed into whatever spreadsheet format a human can read three days too late
If the base carrier’s system is set up to automatically push an event of “loaded” the moment the metal flaps of the railcar are shut, and the taking carrier’s system is set up to automatically receive that event, then data and custody are transferred at the same moment. That is the fix, and while it is not elegant, it is necessary.
High-value, delicate goods carriers are slightly further ahead on this issue, if only because there is no other choice. High-value, delicate-goods shippers tend to gravitate towards either fully or partially dedicated carriers who offer a fully integrated transportation management system around a single set of events, rather than a series of disconnected ones, and who dedicate storage facilities to combined-temperature storage and multimodal shipping rather than using generic warehouses. Carriers that specialize in wine logistics in particular tend to design their handoffs as products that they sell to their customers rather than as an afterthought to a trucking contract. That is a great model to adopt for any high-value, delicate-goods carrier looking to solve this problem.
Bad handoffs quietly kill your ETAs
Predictive ETA algorithms are only as good as the events that they use to calculate them. Using planned rather than actual gate-out events as a predictive tool for when freight will arrive at the next destination will lead to errors at every handoff in the journey that the algorithm does not know about.
If you combine three or four handoffs in a single multimodal shipment using this method, the error margins compound, and the final predicted ETA during the last leg of the journey can lead to either an under-staffed warehouse, retail out-of-stocks, or excessive cushion time for a wine distributor.
Feeding actual sensor events into a predictive model rather than using planned departure events virtually eliminates that risk. The improved downstream ETAs then serve as a useful tool for all stakeholders, including warehouses and retailers, who can use them for capacity planning rather than always planning for the worst-case scenario.
Where the ROI shows up first
The return on investment for higher-tech sensors and better data integration may seem to be mostly in the realm of operations rather than finance, but in practice, the claims and insurance side of the ledger tends to see the fastest wins. With stronger sensor data, a claim can be settled in days rather than months, as the data can show where and how a damage event occurred with much greater specificity. The claims data in turn can lead to more favorable treatment from insurers and carriers in the long run, as they recognize that there is less opportunity for human error to cause a damage event.
Roll it out by fragility, not by budget cycle
Unless you are in the business of shipping pennies, it is unlikely that you need to have sensors on every SKU that crosses your warehouse floor. The more sensible approach is to have sensors on your most fragile and valuable goods, and to roll out the program one SKU at a time. Wine is an obvious candidate for this type of strategy, as it is both physically fragile and relatively high-value, and the value can be measured in both cases. The physical fragility is more immediately apparent, as wine is fairly sensitive to both temperature and mechanical shocks. Both of these have easily measurable effects on the integrity of the bottle, and if those have a measurable effect on the damage rate, then the program has immediate ROI. Installing IoT sensors on random SKUs across an entire fleet is generally a recipe for generating data that nobody has the time or resources to analyze.
Push for standards, not proprietary lock-in
Another thing to keep in mind is that the concept of a “visibility event” is becoming standard across the industry, in part for the same reason that API connections are more valuable than sensors. It is much easier to adopt a new visibility standard if the destination knows what to expect. The new DCSA track and trace standard for ocean freight contains a specification for “visibility events”
Shippers looking to adopt a tracking tool as a means of standardizing data across their network should ask about the standardization of the outputs before asking about the accuracy of the sensors. A sensor whose data can only be viewed inside a walled garden is no better than one that is publicly accessible but inaccurate.
The sensors have been good enough for long enough. What has been lacking is the plumbing into which they connect, and that is the part that actually needs to change.
