TMC Archives

Transportation Management and the Promise of Machine Learning

Transportation Management Systems | By Steve Banker • 12/17/2018

Transportation management systems have a proven ROI. Primarily, a TMS can save companies money by lowering their freight spend. An ARC Advisory Group strategic report on the ROI of TMS found that respondents indicated freight savings of approximately 8% with the use of a TMS application. But that does not mean that there is not room for improvement. ARC is excited about the promise of machine learning to allow a TMS to better handle competing objectives and discover nonobvious impacts on performance.

The primary reason companies buy a transportation management system is for freight savings. These freight savings can be attributed to simulation and network design, load consolidation and lower cost mode selections, and multi-stop route optimization.

But few companies would buy a TMS if it would lead to declining service levels. A transportation management system maintains the service levels by understanding the origin to destination lead times and using that as a constraint during the optimization run. There are also analytics associated with the system. For example, a shipper can analyze which carriers are too often late, and which lanes and destinations often receive late shipments. Consequently, it is not surprising that most companies using a TMS maintain or improve their service levels.

This sounds complex and sophisticated, and it is, but machine learning promises to allow us to go deeper, and to capture nonobvious tradeoffs.

Let me give an example. TMC, a division of C.H. Robinson that provides managed transportation services, published a white paper called Multi-stop Trucking: How It Affects Load Acceptance and Pricing. What it shows is that for multi-stop truckloads, every additional stop lowers the on-time delivery level.

“On single-stop truckload shipments, 80 percent of loads that picked up late still delivered on time. Multi-stop loads are different… The more stops there are, the worse the on-time delivery percentage if one of the early pickups is delayed or late… Trucks that picked up late on a three-stop load, for example, averaged on time delivery only 71 percent of the time.”

Existing TMS solutions will calculate the many situations where multi-stop loads save money. The TMS understands the lead times. It assumes the lead times will be adhered to and the loads will be delivered on time. But the transportation management system does not show in-line analytics to a planner that say, for example, “if you go forward with this shipment, there is only an X% chance the last customer on the route will receive their load on time.” Existing TMS solutions are just not built in a way where these kinds of relationships can be discovered and easily acted upon.

But TMSs – particularly network-based solutions like those from TMC, BluJay Solutions, and Transplace – are rich in data. Machine learning depends on Big Data sets. The problem above would probably be solved using supervised learning. In supervised learning the system is provided with a raw data set but then also provided with a target. In this case, the system is asked to predict on-time deliveries (OTDs) based on a variety of data inputs. How does OTD change based on multi-stops? Based on region? Based on type of customer (customers classified as grocers, for example)? Based on real-time milestone data. And potentially based on lots of other data sets, even data sets that exist outside the TMS. Then, the algorithm attempts to understand how to match the input to the output and whether an input has any impact at all on the desired outcome.

Regression is a form of statistics that can find relationships. But here is the difference between regression in statistics and regression in machine learning: when regression is used in machine learning, there is a feedback loop and the system will continue to learn and get smarter over time.

It is not necessarily just freight costs and service levels that need to be traded off. There are any number of metrics that can be targeted for improvement. I’d argue, for example, that a balanced set of metrics also should include some measure of how popular a shipper is with carriers, whether the shipper is a shipper of choice. First tender acceptance rate for the lead carriers on your lanes is a good proxy for performance as a shipper of choice.

But the Big Data accumulated in TMS and complementary solutions can lead to a fuller understanding of how all sorts of policies and practices affect important performance metrics. It is inevitable that TMS solutions will be built with Data Lakes and supporting analysis tools that will allow these kinds of tradeoffs to be discovered and surfaced to planners.

Why Do Some Shippers Receive Better Prices from Carriers?

Procurement/Sourcing | By Steve Banker • 06/13/2016

When it comes to getting a good price, there are things that one would reasonably suppose would lead a shipper to get better pricing when dealing with carriers. But do they?

For example, one might reasonably expect that the way a shipper treats drivers when they arrive on site to pick up or drop off a load, could affect price or service levels. For example, is there an onerous driver check in process? Do the drivers have lounges and restrooms they are allowed to use? The TMC division of C.H. Robinson provided a large data set from their managed transportation operations to researchers at Iowa State University.

In an article entitled Do “Favored Shippers” Really Receive Better Pricing and Service?, the answer turns out to be “no.” The way a carriers’ full truckload drivers are treated does not affect the prices shippers pay those carriers.

But it turns out that the ability to turn trucks around with less than a two hour dwell time does affect carrier pricing.

I’m a big fan of C.H. Robinson’s sponsored research. I’d rather define a “best practice” as something that has been definitively proved to be true, rather than something that just seems like it should be true. For this reason, when ARC conducted its’ latest survey on the ROI associated with managed transportation arrangements, when we included questions on transportation best practices, we paid close attention to what C.H. Robinson sponsored research had identified as quantitatively proven practices.

TMC Sponsored Research

So how common is it that our shipper respondents reported that “carrier dwell time at our facilities is consistently less than 2 hours?” 66 percent of respondents reported that they followed this best practice, a number higher than I expected.

If you are one of the shippers not doing this, what can you do to improve? The article above makes several suggestions.

“Leverage Live Load Flexibility. Freight may be ready early for pickup. If a dock door is open and the shipper can proactively communicate this to …. (their yard jockeys), they can obtain a truck before a scheduled appointment. Live loading provides the flexibility to deal with fluctuations in demand.”
Pre-stage freight on the shipping dock.
Offer appointments for loading and unloading.
Palletize freight.
Maintain adequate staff for loading and unloading.
For large facilities, provide a good map or directions to the appropriate dock door.
Use drop trailers.

Functionally rich yard management systems (YMS) can facilitate live loading, dock door appointments, and make it easier and quicker for yard jockeys to do their job. A robust dock door scheduling system does not hard schedule a truck for a particular dock door. Rather, the YMS should fill the open time slots in such a way that the dock doors are fully scheduled, but also allows for more flexibility later if trucks arrive early or late.

A robust warehouse management system provides visibility to work on the warehouse floor that allows workers to be moved to high priority tasks – such as not enough people working the receiving and shipping docks.

In conclusion, the vast majority of a truckload rate can be explained by the origin/destination pairings and the distance traveled. Some origin/destination pairings create too much traffic moving in one direction, and not enough in the reverse. That leads to higher prices. But there are things shippers can do to improve their rates at the margin; reducing driver dwell time is one of them. Other best practices include intelligent lane aggregation for low volume lanes and doing strategic procurement with carriers at least annually.

The Total Cost of Ownership Advantage of Cloud Solutions

Cloud/Software-as-a-Service | By Steve Banker • 03/14/2016

The most important reason to go with a cloud solution is that is has a better total cost of ownership (TCO). These savings are driven by faster implementations, easier upgrades, better system longevity, and lower IT costs. Traditionally these advantages applied to public clouds, increasingly they now apply to private cloud solutions as well.

So what is the difference between a public and a private cloud? A public cloud solution is a piece of software code used by multiple customers. Google Maps would be an example. It is the same code that many users from many companies are using. A private cloud solution is a solution that is hosted but one in which each customer has their own version of the software.

A public cloud system is a true commercial-off-the-shelf (COTS) solutions. Over the years many on premise solutions have claimed to be COTS. But historically, even companies that have implemented vanilla versions of the “off-the-shelf” supply chain software will talk about only having 10 or 12 or 15 customizations. With a public cloud none are necessary.

The transportation management system (TMS) market has had public cloud solutions offered by C.H. Robinson TMC, LeanLogistics, MercuryGate, and Transplace for several years. These were true COTS solutions.

In theory, an on premise solution could be put in with no customization. I’ve just never seen it. While customization is also possible with private cloud solutions, there is also a real trend toward foregoing any customizations with these solutions, particularly among small and medium sized (SMBs). Recently I’ve seen presentations by, and talked to, companies that have implemented private supply chain cloud solutions with NO customizations. This is a new development.

I am a very big believer in software that is truly off-the-shelf. Over the years I’ve talked to many logistics or IT executives who thought their operations were unique, and therefore they needed some degree of customization. Other companies felt that they should customize the software to fit the way they conducted business rather than changing processes to fit the configuration workflow. Both are mistakes.

When implementations have dragged on and on, were buggy, and delivered less than the promised results, customization was often a core contributor. Even companies that have had fairly successful implementations, when they finally get around to doing the first upgrade, admit that many of those customizations were a mistake. They subsequently spent time and money to eliminate customizations they initially thought their company just had to have.

In other cases, companies have chosen not to upgrade because of the long and costly upgrade process that would result from their customizations. The warehouse management system (WMS) market provides a perfect example of this. Historically, companies who had customized their WMS software would support the software for some years, even as newer versions of the software with better functionality was introduced. Eventually, the companies decided it was too painful to support the “legacy” software anymore. When they looked at the price of an upgrade, they discovered it would cost as much as a new software license and implementation. At this point, most companies ended up looking at WMS solutions from all the leading suppliers, and often selected a solution from a supplier they had not done business with before.

Because cloud upgrades are so much easier, it is likely companies will retain their existing cloud solutions for longer than they would with traditional solutions. That means users will also continue to have access to new features and functions.

With public cloud solutions, and some private cloud solutions as well, if companies want to continue to use the system, upgrades are mandatory. Users report that many security patches and bug fixes are invisible, a mini-upgrade occurs that users never notice. Larger upgrades may occur once or twice a year. This does require some planning and effort, but much less than would be the case for traditional on premise software. In short, a true COTs cloud solution forces companies to do what is inherently good for themselves.

The lack of customization does not mean that the cloud solution can’t be tailored to a particular company’s operations. They can. Modern enterprise software configuration engines are a marvel. It is even possible to have one instance of modern software manage operations at different facilities, divisions, and regions, with quite different processes.

There are also savings from a cloud solution in the IT department. Companies don’t have to spend money on new servers, or spend money on IT professionals to manage them.

Customization has long been the bane of the enterprise software industry. Fortunately, supply chain and IT professionals increasingly understand that customization should be avoided at all costs.

I recently talked to Fred Erb, the Director of Transportation at SPSCI. SPSCI, a mid-sized company, is a distributor and value added processor of steel products. Their company recently went through an Oracle Transportation Management Cloud implementation. This is a private cloud solution. In many ways, that conversation perfectly encapsulated many of the points I’ve made in this article.

SPSCI began their journey to a transportation management system (TMS) implementation in 2014. One of their divisions had a dedicated fleet and was performing satisfactorily, but the other divisions were struggling with managing their carrier networks. There was a driver shortage and flatbed capacity was tight, especially during produce season when they competed with agribusiness for trucks. Mr. Erb was tasked by the President on how to better manage transportation and make deliveries in a timely manner.

SPSCI looked at several software vendors. The company examined the possibility of a managed transportation service arrangement in which their dedicated carrier would manage all their transportation planning and execution processes. This Third Party Logistics firm used an on premise version of Oracle Transportation Management (OTM). SPSCI particularly liked the OTM user interface.

“Ultimately it came down to whether transportation should be a core competence,” Mr. Erb said. “As a distributor, we decided it should be.”

Once they decided to implement Oracle Transportation Management, they needed to decide how to implement it. Would they go cloud or on premise? As a midsized company, their IT resources are tight. The IT group encouraged them to first look for the best solution,

but was also completely supportive of a cloud based service if it fully met the business’s needs. They decided they wanted to get a TMS up and running without burdening their IT group.

“Cloud was also more cost effective,” Mr. Erb said.

There was absolutely no customization, “that is great.” From a resource perspective, Mr. Erb, an IT analyst, and a transportation analyst were the primary internal resources that worked on this project.

Despite the fact that the different divisions were using the same instance of the software, the software was configured to meet each division’s distinctive needs. According to Mr. Erb, “One division groups orders together so we configured it to bring in the grouping information so the transportation person could create the shipments off of the order groupings. Another division runs more like a manufacturing facility and needed the ability to create shipments off of orders that were not produced at the time they needed to arrange transportation, so we gave them the ability to assign several orders to a shipment, exceeding shipping weights, but then updating the shipments when we produced actuals in our ERP system. OTM was very flexible in adapting to our business requirements.”

Also, “having one instance (of the software) is nice.” The company’s ERP system is utilized differently across each division, so information is re-ported differently. By having one platform for TMS, one division’s data is the same as another division’s data. “From a reporting perspective,” Mr. Erb reports, “that is a real advantage.”

“If cloud was not an option, I’m not sure this project would have happened.” It would have been difficult to free up the necessary internal re-sources, and the business case would have been a much “tougher sale.” Oracle’s cloud solution was “perfect timing.”

Guest Commentary: Global Logistics and the Transportation Management Center: The Next Phase in Network Agility

Guest Commentary | By Fabrizio Brasca • 05/30/2013

In my previous guest commentary on the evolution of transportation management, Driving Value from a Transportation Management Center: Think Global, Act Local, I discussed how transportation best practices were evolving from more rigid, centralized control towers to more flexible and collaborative Transportation Management Centers (TMC). The TMC merges technology, organization and processes to establish a set of services leveraging the scale and efficiency of consolidated network planning, procurement and unified metrics with the agility of localized execution. I said that the next phase in this evolution would be to expand the scope of the TMC from a largely domestic view to the complexity of global transportation operations. I will pick up that thread here.

The evolution to global transportation management is possible because the coalescence of three trends is creating additional value for companies able to master the complexities of global logistics. These trends are:

Consumer Demand – Consumers now have greater control over the omni-channel buying process and their preferences change quickly, putting increased pressure on supply chains to respond agilely without raising costs.

Shift in Control – Just as shippers have taken control of many domestic transportation functions from suppliers and third party service providers to drive greater efficiencies, they are now taking control of the more complex global transportation processes to gain increased agility and cost control in responding to global consumer demand.

Technology – Fortunately, more sophisticated technology is now available to enable shippers to more agilely and cost-effectively plan and manage their global transportation.

When considering the complexities of transportation operations in many different markets, cultures and regulatory environments, shippers must address questions such as:

Where to draw the line between planning and execution?
What should be the scope of control centrally versus at each location?
Where can I best leverage scale?
Ultimately, what mix of cost, risk and agility will create the greatest value for the organization?

To answer these questions, shippers must examine both global process issues and the application of technology.

Navigating Global Processes
Taking control of global shipments is a lot more involved and complex than routing a domestic truckload. There are many more legs to the trip, typically with a mixture of modes, many more parties and languages involved, and hundreds of regulations to comply with. Shippers must make important decisions about “who does what” in this complex, multi-step process, and must evaluate a series of complicated financial trade-offs.

Some of those decisions may be straightforward, such as understanding that carriers in many parts of Asia tend to be smaller operators with varying equipment types and sizes, so it makes sense for those decisions to be made locally. Other decisions are more complex, for example, whether accepting shipments under CIF or FOB Incoterms will best balance risk and financial benefits. You will also need to weigh where transportation management technology adds value versus contracting services from third party providers, such as using customs agents to clear cargo through ports.

The key to making these global process decisions is understanding who has the best upstream and downstream visibility to information critical to each decision and what technology do they have to support the process.

Applying Technology
The complexities of global logistics, the volume of data involved, and the frequency of changes and disruptions make it difficult to gain sufficient agility and maximize potential value using spreadsheets and manual processes. They lack the real-time visibility and capacity to expeditiously balance risk, cost and service to select the optimal solution for each shipment, or to re-optimize shipments when disruptions occur.

Domestic transportation management systems have sometimes been extended to support global movements as well. But these systems are seldom robust enough to take you to the next level of agility and optimization. They typically are not designed for the unique requirements of each point of origin, port or consolidation site across vastly different countries, cultures and regions around the world. Instead, global transportation management systems developed by vendors who have the “feet on the street” in each market are needed to fully understand and leverage the intricacies and best practices for each market.

Why is this important? Because these systems must both empower the local planners with the situation-specific information to optimize local decisions and provide complete visibility and control along the entire journey so TMC planners can make intelligent decisions across the network.

For example, using Incoterms to balance costs versus risk, the TMC planners may change how they group orders into shipments, spread shipments across outbound or destination ports, and select carriers based on sailing dates and service levels. Because of the long lead times for sea legs, needs and priorities may change during the shipment. Planners must be able to re-plan subsequent legs to satisfy priorities and service commitments.

As I hope you are getting the sense of, these are not infrequent or trivial decisions. Managing global transportation involves important process design decisions, the use of sophisticated centralized planning and local execution processes, and support from advanced transportation management solutions specifically designed for the global complexities and local intricacies of international movements. The rewards are greater agility, visibility and control with reduced risks and costs.

As Vice President of Global Logistics at JDA Software, Fabrizio Brasca (@FabBrasca) is responsible for developing innovative transportation and logistics strategies across all industry verticals, strengthening executive-level relationships with JDA’s key customers and prospects, and advising companies on best practices to become more profitable. He holds an Honours Bachelor of Mathematics co-op degree with a specialization in business and information systems from the University of Waterloo, Waterloo, Ontario.

Guest Commentary: Driving Value from a Transportation Management Center: Think Global, Act Local

Guest Commentary | By Fabrizio Brasca • 02/07/2013

To centralize or not to centralize? That is the question – or at least it is a major discussion point for many large, global companies considering how best to manage transportation across today’s vast supply and distribution networks. The argument to centralize usually includes the advantages of economies of scale when dealing with carriers and opportunities for better utilization through back-hauls and continuous moves. The argument for decentralization often involves strong relationships with local carriers, faster handling of the inevitable hiccups in execution, and long histories of successful delivery.

So which is the right answer? Fortunately for everyone involved, the right answer might be both. No, I’m not equivocating in an attempt to keep everyone happy. The truth is that there are advantages to both approaches. The problem is that people have assumed it is an either/or decision. It is not.

The Control Tower
Often the either/or decision comes about when companies feel it is time to replace less efficient, multi-site manual transportation functions with a transportation management system (TMS). Logic dictates that the new system should be run from what is typically called a load control center, or more recently, a control tower. The thinking is that the control tower has visibility to the transportation needs of all sites as well as to the carrier resources available to complete the moves. The control tower can more effectively negotiate carrier rates through a centralized bid procurement process and more efficiently match loads to the least expensive carriers using the TMS.

If your distribution sites are in Pittsburgh, Youngstown, and Cleveland, for example, this centralized control tower approach makes sense, especially if you are shipping similar products. The shipping profiles and carrier resources will be familiar and well-known, so matching loads to carriers will be straightforward.

But what if your business is something like this: you ship electronics from China, component assemblies from Thailand, fabrics from South America, manufactured parts from Mexico, apparel from Europe and finished goods from the United States? Suddenly the centralized control tower approach seems less certain. There are real advantages to local resources in each country making transportation decisions.

The Transportation Management Center
A different approach, and one that I am a particular fan of, is the concept of the Transportation Management Center (TMC). A TMC is an evolutionary step beyond just centralization. It is a merging of technology, organization and processes that establishes a set of services leveraging the scale and efficiency of consolidated network planning, procurement and unified metrics with the agility of localized execution.

The concept of the TMC is to provide a set of services that leverage the visibility afforded by a centralized view of all business activities. The TMC enables the business to think holistically about its transportation needs across all functions and locations, and to plan accordingly. It also allows the business to go beyond traditional transportation planning activities to include value-add services such as network modeling, centralized procurement and a wide range of analytics. This helps companies to better utilize all transportation assets, be more agile in adapting to changing requirements, be more resilient in responding to disruptions, and improve contract and governmental compliance.

Just as important is what the TMC doesn’t do. It doesn’t ignore local relationships or customs. It doesn’t handcuff local transport operations when last-minute changes need to be made. Yes, the TMS will still plan and execute 85-95 percent of all loads, but local resources can manage the exceptions and peculiarities.

Of course, the key to making a TMC successful is having the right technology supporting it. The technology must be comprehensive enough to cover the myriad of transportation modes and methods across all locations while also being flexible enough to support local execution deviations. It must offer global visibility with built-in analytics. And it must offer a simple, yet powerful user interface that cuts across organizational silos so all appropriate parties can use it.

The Next Stage
As the TMC is the next evolution of the control tower, I think there is an additional iteration on the horizon. Current TMC-type logistics operations tend to be regionally focused. This makes a lot of sense considering the value of creating geographic centric scale for such processes as procurement and network route consolidation. As the consumer-driven supply chain continues to grow, however, the need for new definitions of network agility will grow as well. This will force organizations to extend the notion of the TMC beyond continental boundaries and to provide capabilities beyond visibility, creating a more dynamic environment where previously existed segmented rigidity.

To centralize or not to centralize? Maybe it’s not the question at all.

The Science Behind the Benchmarking of Lane Rates

Business Intelligence | By Steve Banker • 07/09/2012

Gathering benchmark data to drive better procurement is a very valuable exercise. In the realm of transportation, this means getting benchmark data on what companies are paying to move freight on particular lanes. However, for a benchmark to be valid, you need a sufficiently large sample size, which is not a problem if you are moving freight from New York City to Chicago, but what if you have regular flatbed shipments between Helena, Montana and Evansville, Indiana? There is little chance you can find a sufficient number of companies shipping on this lane to have any sort of validity to your benchmark data.

In a benchmarking exercise, you also need to make apples-to-apples comparisons. What this means for lane benchmarking is that you need one origin-destination benchmark for reefers, a different benchmark for flatbeds, and so forth.

Yet CH Robinson, LeanLogistics, and Chainalytics provide truckload lane benchmark data for the United States. All three are using a similar methodology that I am convinced is valid even for lanes with little traffic. How can this be?

This is where the magic of clever mathematical modeling comes in.

I attended a presentation on this topic by Kevin McCarthy, CH Robinson’s Director of Consulting Services, at TMC’s Interactive Client Forum last November. I recently had a follow up conversation with him about benchmarking, and also with Chris Johnson, the VP of Research & Development at LeanLogistics. Both CH Robinson and LeanLogistics are Logistics Viewpoints sponsors.

TMC is a division of CH Robinson that uses its own software-as-a-service (SaaS), platform-style Transportation Management System (TMS) to plan and execute loads for its clients. Some customers have chosen to purchase the company’s TMS and not use TMC’s managed services. LeanLogistics is also a provider of platform TMS and transportation managed services.

Both companies have a nice data set. CH Robinson is one of the largest 3PLs in the world and it has extensive brokerage operations. The company is leveraging its internal lane data and supplementing it with data from TMC customers that choose to opt in. If TMC customers don’t provide data, they don’t get the benchmark data.

LeanLogistics is a division of Brambles, which also owns CHEP. CHEP makes a very large number of transportation moves, and that data can also be leveraged. At LeanLogitics, new customers sign a contract where they are told their data will be aggregated for benchmarking purposes. All LeanLogistics customers have access to the benchmark data in the form of in-line analytics that show the current lane benchmark as well as the historical trend.

All three suppliers create a regression model to create a benchmark rate for a low volume lane. They begin by looking at whether an origin or destination is a location with balanced demand and supply; whether it tends to have many trucks coming in full but going out empty (IN); or whether it has a lot of shipments going out (OUT) but has less demand for trucks coming back.

Going back to my Helena-to-Evansville example, there might not be any shipments in the data set between this O-D pair, but the model can use all the data on Helena outbound flatbed shipments to all other destinations to see whether this is an IN location or an OUT location.

I suspect Helena is an OUT. Helena probably has lots of trucks containing lumber going to other parts of the country, but far less demand for inbound flatbed shipments.

Because getting a load for the return trip will be difficult, a carrier needs to charge a premium that will cover the empty miles it has to travel in both directions. For example, once the load is delivered, the carrier might have to drive to St. Louis to get a load to Denver. The miles from Evansville to St. Louis and Denver to Helena are empty miles the carrier needs to be compensated for. By looking at what the average national rate per mile is and comparing it to the rate per mile of a particular origin or destination, the model can understand whether the location is an IN or an OUT and the extent to which that is true.

Kevin tells me TMC is getting r-squared values in the high 80s and low 90s. Although statistics was my minor in grad school, I’ve forgotten a lot of it, but I know that r-squareds that high indicate very, very high predictive value. Young professors trying to get tenure would kill for this kind of a result.

In addition to using this benchmark data for procurement, it is also extremely valuable for network design projects.

Chainalytics is the granddaddy for doing benchmarking in this area. But there are some advantages to getting this kind of data from a TMS provider with a SaaS platform. First of all, if you are a TMC or LeanLogistics customer, the reports are free. Secondly, this is not a manual process where you have to send your lane data to a third party on a quarterly or semiannual basis; LeanLogistics and TMC already have your data. And finally, while lane rates overall don’t change all that much over the course of a year, some rural locations can have a significant volume shift that would affect a lane benchmark. With TMC and LeanLogistics, the lane data is constantly refreshed, it is not one quarter out of date.

One of the things I love about platform-style TMS solutions is their ability to monitor network data and drive value in new ways. This is a prime example of that.

(For related commentary, see “The Missing Link in Transportation Business Intelligence”).

A complete analysis of logistics trends, technologies, and services

Archives

Login