Sending the right insights to the right person at the right time
Next Big Sound was a data analytics platform at Pandora aimed at empowering artists with the knowledge and insight to grow their careers. We defined insight as a remarkable, bite-sized piece of content that drives engagement on Pandora. And over the years we developed some notable insights as seen here, here, and here. But developing these insights was only half the challenge. We needed a vehicle to deliver the right insights to the right person at the right time.
At the time, our team's target KPI was to increase the amount of weekly insights consumed. We considered notifications as well as emails as the vehicle and decided to bet on emails. The infrastructure for emails was ready for use, and we were seeing open rates already outperform the industry standard. And so, emails became the prime target to help increase our weekly insights consumed metric.
We started with this hypothesis:
And success would be determined by increased open rates, increased remarkability, and higher overall email satisfaction.
With that, we began compiling the research. We looked at bi-annual survey results, feedback from customers, and conducted user interviews. We noticed a handful of patterns and themes:
There were distinct behavioral differences for users following 1 versus more than 1 artist
The value of insights ranged on a spectrum of time-sensitivity
Context is key to understanding performance
With these findings, we kicked-off a series of cross-functional workshops with product, tech, and data scientists. We leveraged a couple of frameworks that would enable us to better make and guide our decisions. Those were:
The CREATE framework
Contextual. Remarkable. Educational. Actionable. Trustworthy. Exclusive. This allowed us to measure and weigh each feature based on how insightful it would be.
The DIKI framework
Data. Information. Knowledge. Insight. This data pyramid provided the framework to think about and organize data. Each level built upon another by adding context.
With these frameworks, we devised a series of emails, each for different purposes. Users were also able to subscribe to whichever series were more pertinent to their use cases. For example, if you were a solo artist managing your career, you would get a more in-depth weekly email about your performance. Compare that to a manager overseeing a few acts, they would receive a weekly report of the top line metrics for their artists, getting an overview of performance.
We launched the following email series:
Weekly reports for a single artist
Users following a single artist focus solely on that artist's performance. We designed an email for them that looks in depth at the various aspects of performance across streaming platforms and social media.
Weekly reports for more than one artist
Users following more than a single artist needed an overview for the artists on the top of their minds. We derived which artists were most relevant for users by their recent activity on NextBigSound.com. The emails were a templated view of the top metrics for a user's most recently interacted with artist profiles.
Notifications
Notifications were designed to alert users to notable streaming and social spikes for the artists they followed. So any artist or artist manager would know when something has happened, understand what it was and catch the opportunity to re-engage with their audience.
Pandora Programming Notifications
Getting programmed by the Pandora curation team was a shareable moment for many artists. But there was no easy way to know when that happened. To ensure that no moment was lost, these notifications would automatically alert users to when their artists were added to a new genre station or playlist.
We measured each of these email series against the CREATE framework to ensure that they met a minimum threshold of remarkability, actionability, and exclusivity through user testing and surveys. The results were that not only did we see an increase in open rates in emails, but it caused a flywheel effect as it drove traffic back to our web product as well.