85% accuracy in predicting customer churn

Who are you?

Sorted.com came about in 2010, inspired by the frustration of a slow Post Office queue. We’re a parcel delivery comparison website designed to make sending parcels as simple and as hassle-free as possible. We also give individuals and small businesses the kind of bulk discounts and services that bigger companies achieve with their buying power. Today, we deliver over one million parcels a year for 100,000 customers via 14 different couriers to over 200 countries.

What was your challenge?

When we came across Peak, our sales were stagnant, and we were spending a lot of money on advertising without being sure that it was working. We also had a lot of data that we weren’t really using and knew we could do more with it. We wanted to understand who our best customers were so that we could focus on optimizing our service for them.

What did Peak do?

Peak used AI techniques to help us optimize delivery times, improve the online experience for our customers, and reduce costs. Its platform enabled us to drive orders and revenues in a fiercely competitive market. Through this initial work developing the understanding of our customers and optimizing our service, Peak also identified an opportunity for us to diversify.

It found that businesses accounted for 1% of our customers, but 50% of our revenue, so we built a product specifically for that segment. In addition to providing a specific, tailored service for that market, it allows us to track all key metrics, predict the lifetime value of customers and look for signs of churn or high-spending among customers so that we can act on that info. We can also see which campaigns are working best for which users.

What’s the upshot?

Harnessing our data with Peak has given us a far more in-depth understanding of the business than we realized could be possible, helping us to both optimize our business and develop a new revenue stream.