Personalization is an especially important driving force in the competitive music industry, where consumers have a wide range of music streaming and purchasing services to choose from. That’s why Ponomusic, a music download service and dedicated music player founded by Neil Young, turned to Datameer, a big data analytics platform, to ramp up its web site with highly personalized recommendations. The results were immediate and dramatic, and included:
- A spike in album sales by 15% to 20% within the first week;
- More customers experimenting and buying albums they normally wouldn’t buy; and
- Praise from customers on the accuracy of their personalized recommendations.
Ponomusic offers high-resolution audio files ranging from CD quality to master studio quality, which allows customers to hear music as it was recorded in the studio. But since digital music catalogs aren’t exactly tangible items with SKUs, it was difficult for the company to track customers’ purchasing history, a key element in personalization.
“For us, because digital music catalogs are so vast and we literally have no means of [assigning] SKUs, we had to work with a partner who would be able to handle those challenges,” said Randy Leasure, VP Business Development and Content Marketing at Ponomusic in an interview with Retail TouchPoints. “It’s a very fluid data set that changes because of people’s purchase history. We’re trying to make that deeper discovery experience better for people. At the end of the day, they’ll buy more music, and will also keep coming back and be satisfied.”
Ponomusic leveraged Datameer to help the company make its “shopping experience richer and discovery path deeper for people to find what they’re looking for,” said Leasure. The company also wanted to introduce its customers to music that they didn’t know was available in high-resolution formats. “Every week we get new, re-mastered titles, and it’s important for us to bring those to people.”
Customers Pleasantly Surprised By Accuracy Of Data-Driven Recommendations
The company worked with Datameer on a “recommendations project,” which was designed to personalize recommendations based not just on what customers have purchased themselves. For example, if a customer purchased a hard rock album, but they also like classical and jazz music, the Ponomusic recommendation engine is able to take that combination of data and match it with similar purchases from other customers, to provide new recommendations for that specific user. The result has been a highly satisfied Ponomusic community.
“We mix that data together and come out with those recommendations and refresh it on a daily basis,” said Leasure. “Once [a customer] purchases something, and they’re logged in, they will see a fifth shelf appear on our homepage, which is the recommendation shelf. It’ll surface those recommendations to [the customer].”
“This type of recommendation solution we’ve built features a lot of frequency counting,” said Datameer Solutions Engineer Victor Liu in an interview with Retail TouchPoints. “Using frequency numbers, we look at all the users and what albums they’ve purchased. That [data] was already contained within the Ponomusic database. That information is then used to process how many times an album is bought concurrently with another album. Then we look at all the possible combinations.”
Following the implementation of the personalized recommendation engine, Ponomusic not only experienced a 15% to 20% jump in album sales, but the service also received praise from customers on the accuracy of the personalized recommendations.
“We’ve definitely noticed an uptick in sales,” said Leasure. “In our community, people are really vocal and talking about it. There was a ton of positive feedback. Customers were surprised how accurate the recommendations were based on what they purchased, and that some of the things that were surfacing were not normally what they would think of. People were experimenting and trying, and buying things that they normally weren’t buying.”