Singapore start-up Parcel Perform is collaborating with Amazon Web Services (AWS) to use machine learning for improving the delivery experience of online shoppers.

Providing the latest delivery date of online purchases is a crucial driver for online shoppers’ satisfaction, but arrival date predictions by logistics carriers are often not available, inaccurate or out-dated, leaving consumers frustrated with the experience.

Parcel Perform is leveraging machine learning to predict more precisely when customers will get their online shopping items.

Parcel Perform is a carrier-independent delivery experience SaaS platform. The machine-learning algorithm can predict date of arrival of any ecommerce delivery. It was developed in collaboration with Amazon’s machine-learning laboratory.

The underlying machine-learning algorithm trains on delivery pattern data from millions of past orders to predict the date of arrival for any ecommerce shipment. This allows consumers to receive predictions of when their parcel will arrive, which is critical for improving the customer experience.

Today, 95 per cent of logistics carriers don’t provide this information to the recipients. With consumers shopping online, providing delivery predictions before and after checkout is essential, and offers opportunities for retailers to differentiate their delivery experience.

Parcel Perform’s ecommerce customers want to leverage their fulfilment and shipment data, and convert it to a specific, time-bound promise delivery date. Having started the cloud-native business on AWS infrastructure, Parcel Perform turned to AWS to help build a solution to this problem. They collaborated to build a flexible, scalable combination of AWS’s expertise in machine learning with services such as Amazon SageMaker, with the specific requirements of the logistics domain.

“With our date-of-arrival prediction, consumers now know when their parcel will arrive, instead of just where it is at the moment,” said Arne Jeroschewski, founder and CEO of Parcel Perform. “This date-of-arrival prediction truly makes the difference to their delivery. The experts in Amazon’s team have been working closely with our development teams in advising and improving the machine-learning algorithms that drive this service to make it a real value add for our customers.”

Michelle Lee, AWS vice president, added: “Our work with Parcel Perform enhances the customer’s post-purchase experience by using machine learning to predict when a customer will receive their items. Combining our expertise in machine learning with the power of Amazon SageMaker, we were able to help build a solution that can scale to meet the needs of Parcel Perform’s ecommerce customers around the world.”

Parcel Perform integrates with more than 600 carriers worldwide and is used by Australian brands including Nespresso, Catch Group and Mecca to improve delivery experiences.