Unlocking Supply Chain AI To Meet Customer Delivery Expectations

At Convey, we’re committed to helping our customers improve the delivery experience, including preventing negative customer experiences when delivery expectations are missed. This holiday season alone, retail sales are reaching record highs. Sales are predicted to skyrocket to create the first-ever trillion-dollar season. As a result, retailers are sending a record number of shipments to customers. Every day, an average of 6% of parcel shipments will face issues in-transit, such as missed pick-ups or transfers, that will cause them to miss their Estimated Delivery Dates (EDDs).

When this happens, retailers won’t receive an updated delivery estimate from the carrier until the EDD is missed, eliminating the opportunity for retailers to notify their customers. This leaves consumers empty-handed, wondering where their shipment is. They are significantly more likely to call the retailers asking “Where is my order?” (otherwise known as a WISMO call). This can rack up operational costs for the retailer, or worse. 83% of customers will churn from the retailer following a negative delivery experience.

A New Convey Insight: Shipments Predicted To Miss EDD

Convey set out to provide our customers with a way to proactively resolve these customer experience (CX) failures. We examined our datasets of over 300 million shipments to determine if it was possible to predict when a shipment would result in a delay.

This research resulted in a new Convey Insight: Shipments Predicted to Miss Their Estimated Delivery Date (PMEDD, for short). PMEDD uses machine learning models to assign each in-transit shipment a risk score, indicating how likely the package will miss its EDD.  The PMEDD Insight allows retailers to notify customers of a missed delivery before carriers even report it. This proactively averts customer service nightmares where customers are left package-less and empty-handed.


How The PMEDD Insight Works

To find shipments that are likely to miss their EDD, our machine learning models identify data points from over 300B shipments while they are in-transit. To do this, we look at both the data and events reported by the carrier. And more importantly, we look at the data and events not reported by the carrier when they should’ve been.

Then, through our machine learning classification models, we compare these data points to patterns we expect from on-time deliveries. When there is a high level of certainty that a shipment will miss its EDD, Convey tags it as ‘PMEDD’ and assigns the shipment with a risk level. These percentages translate to the probability that each shipment will end up missing its EDD. 

Risk levels

Every day, our AI model flags approximately 6% of parcel shipment volume as PMEDD. We report PMEDD by a probabilistic risk level. At all risk levels, shipments tagged in our system as ‘PMEDD’ are more than 96% likely to miss their EDD.

When there is higher risk, such as during peak season when delivery volume spikes, our machine learning models only get more precise. During these times when delays are especially risky, we accurately predict that shipments will miss their EDDs 99% of the time. 

For example, during Cyber Week 2019, when both volume was high and winter storms slowed delivery, 17% of shipments were flagged to as ‘PMEDD.’ Had these shipments gone undetected, it would have caused WISMO calls and potential churn from over 850,000 customers.

Because Convey reports PMEDDs by carrier, service level, and location, retailers can detect and diagnose larger network problems in real-time. Often, delays are concentrated in a few locations, so this insight helps retailers quickly identify and take action to solve a delayed carrier facility, lost truck, weather event or other major event causing delays.

Leveraging Supply Chain AI Is Retail’s Competitive Advantage

Predicted to miss
This means a retailer has 12,126 customers that will be disappointed tonight, and 12,126 potential WISMO calls.

Every shipment flagged as PMEDD represents a potentially disappointed customer and a risk of customer churn. We know that customers strongly prefer to be notified when delivery expectations change: Convey research shows that 98.3% shoppers want a notification in some form if their delivery is late – up more than 10% from last year (87.8% in 2018). Shoppers overwhelmingly prefer to be notified via email (55%) and SMS (32.5%) if their package is late.

Even more, operating costs increase when retailers choose not to alert customers when their packages will be delayed. Customers who see that there are delivery exceptions such as delays, but who aren’t notified, are more likely to have a WISMO call.  Alerting customers when their shipments are expected to miss their EDD reduces this call volume.  

Accurately Predict Delays Caused By Weather And High Volume

 Convey’s PMEDD Insight detects delays no matter what causes them, but it can be especially effective in cases of weather events and high volume.  Between Thanksgiving and the second week of December, Convey detected over 56,000 delays that later received weather delay exceptions from their carriers, but not until after the delivery expectation was already missed.  On average, carriers reported the issue over 27 hours after Convey detected the problem.

PMEDD allows Logistics Managers and Supply Chain Managers to detect and act on network events at the micro-level using exact location reporting:

Weather-related delays
This high concentration of delays in Glenn’s Ferry, ID showed a large weather-related issue before carriers reported any weather alerts, exceptions, or delays.

And at the macro-level, using real-time network reporting, logistics managers can see extensive delivery problems in the northeast caused by the 2019 winter storms:

Heatmap of predicted delays

Detect Delivery Delays During Peak Season 

Frequency of PMEDD

During peak season, both customer expectations and the pressure on delivery networks amplify. During Cyber Week 2019, Convey saw five million parcel deliveries.  Among those were 850,000 packages detected as PMEDD that ended up missing their estimated delivery date, and none of these received an updated EDD from the carrier until after the EDD had already passed. This means that 17% of deliveries did not arrive when the customer expected them. The retailer broke its delivery promise.

While we see an uptick in shipment volume and exception rate across the board during peak, EDD performance is most heavily affected. The table below shows the increase before and after Thanksgiving for shipments, exceptions and PMEDDs.  On the busiest delivery day of Cyber Week, Thursday December 5, there were more PMEDD shipments than all the carrier reported exceptions combined. 

Peak vs off-peak 2019

When shipping networks are stressed by extreme volume and weather, delays are the most common consequence. Yet, these delays are rarely accounted for until after the delivery expectation has failed. 

The stress on delivery networks also leads to a decrease in customer happiness. During Cyber5, when shoppers check tracking pages more often per shipment, we saw a sharp, negative swing in feedback from customers who ordered shipments with missed EDDs. Before delays trigger customer disappointment, shippers can leverage Convey’s PMEDD insight to mitigate the main cause of bad experiences during peak. 

Tracking page feedback

The Supply Chain Manager’s 6th Sense: Predicting Issues Before They Impact Customers 

No matter the size of the retailer, bottlenecks within the supply chain can affect customer expectations. This famously happened to Apple in 2017 when launching its iPhone X, resulting in angry customers. More recently, fans of Popeyes were forced to endure a long wait for its anticipated chicken sandwiches after high demand, resulting in a shortage of chicken. In the last mile, the PMEDD allows retailers to detect problems in fulfillment networks as they happen and handle supply chain bottlenecks before they have a negative impact on customers. 

In a world where eCommerce expectations are only getting higher, it’s more important than ever to make and keep bold fulfillment promises. It’s no longer enough to rely on parcel carriers to inform retailers and customers about distressed shipments.  Doing so leaves an unacceptable number of customers in the lurch when their package doesn’t arrive on time.

Convey’s new Insight, Shipments Predicted To Miss EDD, is a subset of Convey’s new Discover reports — 30 out-of-box reports made to answer pressing delivery questions in real-time and over time. Learn more about Convey Discover here, or learn how Stalled Shipments can cause WISMO.