Three ways data analytics and machine learning will enhance ecommerce parcel shipping in 2021 and beyond

Optimizing the ecommerce delivery experience while controlling transportation costs is more important than ever for online retailers. Consumers today are growing more demanding when it comes to online delivery. Not only do they expect a wide range of delivery choices and insist on their orders arriving perfectly on time (or to be notified in advance of any possible delays), but they also expect to receive these benefits for free, piling even greater pressure on shippers to drive efficiency and contain costs.

Meeting these expectations is a tough ask and retail parcel shippers are using every means at their disposal to give themselves an advantage. This includes turning to supply chain business intelligence, data and analytics to identify ways to improve the delivery experience and to uncover new ways to generate efficiencies. In 2021 and beyond, we expect to see a continued increase in data source obtainment, business analytics platform expansions, and machine learning to support strategic decision making to meet these objectives.

One particularly exciting area is the availability of, and demand for, new external data sources to augment transportation management data within Business Intelligence (BI) platforms. Industry surveys repeatedly find organizations are prioritizing visibility and predictive analytics within supply chain technologies. The growth of technology providers that focus on visibility and the customer delivery experience validates retailer pressure to offer a flawless customer experience. Supporting this increased focus, budget funding to support these capabilities is also on the rise.

Here are three examples of how additional external data could be integrated with transportation management data to help retailers make informed parcel shipping decisions.

1) Weather data:  predicting how major weather events will impact carrier performance

Using external weather data feeds, shippers can model and evaluate the extent to which major weather events such as hurricanes or snowstorms have impacted past carrier performance.  With these data analyses, organizations can then predict what will happen when similar future weather events occur. Taking things to another level, Transportation Management Systems (TMS) for parcel shipping with geographic carrier hub and spoke networking capabilities are able to use machine learning algorithms to best route shipments to appropriate shipping points. They can optimally meet the Service Level Agreement expectation of end customers and minimize or ‘bypass’ weather related risks. By analyzing how carrier partners have performed at the service level compared with their competition, shippers are able to make more informed decisions about which carriers, services, and origins to choose for orders moving through weather-impacted regions.

This would be incredibly valuable to ecommerce companies and goes way beyond standard carrier information, which typically only tells you information such as “a delivery from A to B takes two days and costs $10.”

For example, past performance analyses may reveal that on average, orders shipped from either of two distribution centers to a destination will arrive on time, with 97 percent accuracy. However, a pending winter storm event is coming. With weather data analytics and machine learning, the shipper knows from experience that the type of pending storm will likely add a one to two-day delivery delay when routed through one distribution center over the other.  Use of weather data can have key routing benefits to meet customer expectations. 

2) Traffic data: comparing the impact of traffic flows on parcel delivery journeys 

Similar to the weather example, BI platforms can integrate available road traffic data against in-transit carrier deliveries. Historical benchmarking of traffic data can augment confidence levels about potential delivery delays or failures.  This enables shippers to assess how traffic flows may affect deliveries and proactively notify end customers of potential delays. 

Carriers are expanding geo data availability during delivery processes, and its consumption and interpretation will provide key last mile delivery insights.  As consumer demand for delivery event visibility expands, and technology used by carriers evolves, it will be critical to expose and share delivery events.  Using social media data as an early warning of incidents that could disrupt parcel deliveries is a new practice that carriers could, and likely will, deploy.  

Another fascinating area of use is anomaly detection through social media activity tracking. It’s the power of the people. For example, a large spike in unusual activity on a social platform in a particular location or region could be a signal that there may be a disruption affecting carrier networks or delivery processes. Algorithms already recognize these spikes in activity, which often occur before carriers are aware or traditional news media reports them. The actual incident might be a major accident, an unplanned protest march, or even a riot.  While technology may not be able to immediately identify the specific nature of the event, social media spikes can be used as a trendline against carrier performance in that region during the same time.  Marking these incidents as impactful or not, helps machines learn through these incidents over time.  Which carriers were most affected? What were the impacts?

The challenge of co-mingling client data

In an ideal world, BI platforms with parcel transportation management systems could provide even more accurate insights if they were able to consolidate both carrier visibility and performance data, as well as these types of external data, across different customers. However, most of the key global parcel carriers’ contracts prevent this kind of co-mingling of data for analyses. 

But data ownership and usage are a topic for another day. Suffice it to say, if it were possible to secure clients’ and carriers’ permission to aggregate performance data across a parcel transportation management system vendor’s largest customers, this bigger data sample would produce more robust analytics, accurate KPIs, and broader trend results. It remains to be seen if this will ever happen.

Despite this limitation, however, we continue to see an increase in the use of external data feeds to augment BI systems within parcel transportation management software. And a properly integrated supply chain technology stack, complete with BI, will support strategic decision-making by providing insights that enable greater flexibility to adapt to today’s rapidly changing supply chain landscape while keeping costs in check.

Mike Eisner, vice president of business intelligence and data analytics, Logistyx Technologies

Source link