Racing to predict the future in transportation logistics

Steve Beda (pictured), executive vice president of customer solutions, Trax Technologies, explores the importance of embracing analytics for best-in-class operations.

Not that long ago data analytics was the purview of political wonks, dusty academics, and financial analysts crunching numbers to beat the markets. Then things changed as business leaders felt increasing pressure to find new ways to improve performance, deliver shareholder value and meet skyrocketing consumer expectations for products and services. Marketing teams seized analytics as a new, more sophisticated tool for understanding customer sentiment and predicting buying patterns, while operations leaders found ways to use production line sensors and collated data to foresee line failures and areas for improvement. Sadly, transportation logistics has been slower to adopt analytics at scale to the point where, even today, some operators continue to make potentially enormous and impactful decisions for their business based on, in some cases, little more than custom and practice and a well-trained gut. That too has to change.

It is simply not enough anymore, in the ultra-competitive world of transportation with soaring gas prices, global and real-time trucking and shipping complexities and the long shadow of pandemic supply-chain interruptions are still creating staff, capacity, and container shortages. To understand the root causes of all of this, to identify market influencers and potential problems before they bring down an entire network, global transportation executives now need access to both comprehensive data and sophisticated analytics capabilities. Embracing both will do one crucially important thing for every business: help us see the future. Or at least make data-driven decisions today that lead to efficiencies tomorrow and success for the long haul.

As we begin the process of due diligence and adoption, let’s take a look at the four core types of data analytics and just what they do for supply chain efficiency.

1. Descriptive Analytics – a Foundational Understanding

You can’t look ahead without understanding where you came from and there’s absolutely no way to improve performance without pinpointing what’s working well or identifying where problems or weaknesses may be emerging. For both of these critical performance benchmarks, we need Descriptive Analytics. This is quite simply an assessment of current state using historical data. A collection, curation, and analysis of data that shows us what happened over a period of time. The number of deliveries. Volume of goods. The number of carriers, routes, modes, and so forth.

It’s the foundational baseline on which more complex data analytics practices build insights into performance and quality of service.

2. Diagnostic Analytics – Why did that happen? What does that mean?

So we have a list of what happened. But understanding why things occurred, pinpointing the variables and influencers for more nuanced strategic learnings – that’s the role of Diagnostic Analysis. Diagnostic looks beyond baseline operational outcomes to measure performance against goals, deadlines, by mode and geographies, and many other variables. As part of this process look for three key measurement strategies:

Carrier Scorecards

Carrier scorecards are a critical and comprehensive tool for understanding why outcomes happen. For decades scorecards were inconsistent, non-standardised paperwork that captured intelligence but made it incredibly time-consuming and inefficient to analyse or draw strategic insights. Ideally shippers today should enforce the use of electronic, standardised formats with their carrier partners to make analysis consistent and meaningful. From the collection and analysis of scorecard information you should expect to rapidly understand:

  • On-time Delivery: how many packages arrived on time versus missed deadlines – and where? Which carriers? Which markets?
  • Damage Tracking: how many packages arrived broken, damaged, or otherwise contaminated, affecting replacement costs and long-term reputation?
  • Billing Accuracy: do final invoices match estimates? What is the percentage variance? Which carriers are less accurate with their billing and why?
  • Fallouts: how many loads were delivered to warehousing or other points in the network only to be returned due to limited storage and handling capacities?
  • Tender acceptance: what is the acceptance rate among carriers for contracts? If this is low, there may be a perceived problem with stated terms, compensation, or another factor that must be worked out in future negotiations.

Functional Performance Metrics

Quite simply this is the process of data collection and analysis to provide a detailed understanding across the operation. When measuring performance in this way, key indicators typically include shipping times; order accuracy (did the right product ship to the right customer at the right price?); delivery times (are there unforeseen delays between shipping and delivery?); transportation costs; warehousing costs; number of shipments; inventory accuracy and inventory turnover. There can be many other variables too, but measure against these and you’ll build a comprehensive picture of your success in the marketplace.

Cost Analysis

Lastly, how much did everything cost? Is there a discrepancy between budget and actual cost? And where is the variance?  Understanding these factors will help identify weakness in the supply chain for renewed operational improvements in the short term.

3. Predictive Analytics – the impact of external data on planning and validation

The data race is on – with all kinds being collected, curated, and analysed across industries, organisations, and businesses. As external data, beyond our logistics operations, becomes more readily available and with real-time information becoming instantly accessible, the role of analytics is rapidly becoming even more critical to day-to-day operations. In predictive analytics, analysis combines external data with owned company data to create context for how we may operate in the moment or over a period of time within geographies, modes, channels, and even by carrier. Even more significantly, predictive analytics, when employed correctly, can predict the outcome of a strategic or operational change and help validate the change prior to execution, limiting risk as much as possible. It’s important again to ensure that data content and capture quality is high to enable accurate modelling using the available data inputs and known constraints to make the most accurate and meaningful prediction.

Major measures for predictive learning include:

GPS Tracking: GPS tracking was one of the earliest data points for operational efficiency analysis and is still important. Real-time tracking of trucks, trailers, and other equipment creates a broad baseline oversight of operations. Knowing at any time where carriers and other equipment are can lead to greater efficiencies and resource optimisation. But it also enables driver and safety performance monitoring, reduces administrative time to manually check-in and monitor locations and progress, and ultimately leads to more end-to-end quality customer service.

Weather Data: Knowing when weather constraints may affect performance, cause delay or even endanger drivers or other assets can be critical, especially when dealing with global shipments across countries and regions. But weather can also impact the supply and demand balance too so planning for it is important even in the short term. Tornadoes and hurricanes, flooding, and deep freezes create massive demand for generators, water, heaters, and lumber to protect against storm damage. A surge in demand creates supply challenges, can spike prices, and have other knock-on effects down the supply chain.

Traffic Data: Look at how major coastal storms can cut off whole towns and cities, or even a major sporting event in a downtown area can cripple a traffic system for hours. Understanding traffic problems, road or modal outages, parking availability, port congestion, or volume issues is an increasingly important factor in best-in-class operations. There’s a huge industry focus happening right now on the use of AI to provide real-time route optimisation – the local, national, and global re-routing of shipments and deliveries based on changes to traffic patterns in the moment. Rightly so, using external traffic data to map clear, efficient routes or make urgent changes can save time, money and protect reputation with customers.

Predictive analysis monitors and models against all of these factors and more to create an impact analysis – what we need to know, do and change to avoid network, shipping, or delivery problems. This enables strategies and data-driven decision-making that will ultimately help reduce the negative impact of external factors beyond our control while optimising operational efficiencies to meet deadlines and stay within budget.

4. Prescriptive Analytics – where are we headed?

The ultimate goal for any best-in-class transportation logistics operation is the transition to, and full adoption, of a Logistics Control Tower strategy. A seamless collection and analysis of high-quality external and internal data collected over time, modelled and used to provide a prescriptive analysis of actions needed and changes to be made.

When the Suez Canal – the world’s largest shipment gateway – was blocked last year, the effect on the local, national and global supply chains was immediate and lasting. Often when a crisis like this occurs, the teams managing logistics just don’t have the manual capabilities to deal with the complexities, scale, or scope of the problem – or the impending and far-reaching impact. They look at reports, shuffle paper, and do their best to make decisions, but only sophisticated data-driven modelling can rectify massive outages or high-impact events.

Or picture this. You are a manufacturer with a delayed ship in Long Beach with raw materials impacting three different finished goods. What do you do? Fortunately, your control tower can see that two of the three will be out of stock before the inbound arrives. One of the finished goods is a high-margin item and sold to one of your top ten customers. Your system assesses the cost for air freight and determines this is just enough to cover the top customer and automatically weighs the cost against the loss of margin on the sale. That is the future of transportation logistics.

Prescriptive data analysis will continue to play an increasingly pivotal role in day-to-day operations, helping make and prioritise decisions. When product inventory is severely limited, can we prioritise delivery to customers that would be most impacted by failure to deliver? Can we instantly measure, understand and rank cost to customers and then prioritise by region for fastest delivery? In this way, shippers can fulfil high-impact orders to big-box retailers like Target while managing other options to cater to smaller customers with a less urgent need. Or is it advisable to ship highest margin products first when carrier capacity is constrained? And if we are facing constraints, is it better to continue with a low-cost carrier or allocate budget to pay spot rate shipping costs? All of these questions and more can be managed and answered with prescriptive analytics using best-in-class master data management standards and AI to predict outcomes and prescribe action.

Take Steps Now To Be Ahead of the Curve

So, the ultimate question: Where is your operation in the analytics race? Are you still reporting on what already happened in the past? Are you beginning to explore the abilities to predict what will happen? Has your organisation fully embraced the transition to a data-driven, analytics-led strategy, and do you have the tools to prescribe the necessary actions or a recommended path forward? Wherever you are on the curve, four fundamental factors must be kept in mind:

  • Be honest with your current state If you are still in “hindsight.” Own it and plan for how to make the necessary changes to “insight.” Don’t try to skip a step in the progression of analytics.
  • Plan how to get access to all needed data (internal, external providers, and public data). You may have to subscribe to services, turn to logistics providers and expert partners, etc.
  • Hire a data scientist or find a good partner to help prove out a few test use cases. When you start out you should always choose the project that has the highest value (return) for the lowest effort (investment). Make sure that you are solving a real challenge that drives value.
  • Take your impact and go after more.

Making a transition to a data-analytics-based operation isn’t simple or instant, but the investment will be invaluable. Embracing data is essential to future success, particularly given the events of the past two years. If the Russia-Ukraine War, global pandemic, supply chain shortages, Suez canal, and increasing environmental events tell us anything it’s that the landscape for our industry has changed fundamentally. Cause and effect is the new norm. Expect obstacles, challenges, and a fluid future where custom and practice must give way quickly to prediction, prescription, and inarguable data-driven action to succeed tomorrow and down the road.

Steve Beda is executive vice president of customer solutions for Trax Technologies, a global leader in Transportation Spend Management solutions. Trax elevates traditional Freight Audit and Payment with a combination of industry leading cloud-based technology solutions and expert services to help enterprises with the world’s more complex supply chains better manage and control their global transportation costs and drive enterprise-wide efficiency and value.

The post Racing to predict the future in transportation logistics appeared first on Logistics Business® Magazine.