Consumers value access to ad-supported content when they feel that there’s a fair trade between the content and advertising. The study, “Consumer Sensitivity to Audio Advertising: A Field Experiment on Pandora Internet Radio,” examines the balance in this relationship. Specifically, the research analyzes the correlation between the number of advertising breaks, the number of ads, and content consumption. While Pandora sells display, video, and audio advertising, this study focuses on audio ads. Today most of Pandora’s advertising revenue is generated from audio advertising.
Jason Huang, David Reiley, and Nickolai M. Riabov created an experimental design using more than 30 million Pandora listeners. The design included nine test groups (1% of listeners) and a control group (10% of listeners). Each test group received either three, four or six ad interruptions per hour. And each interruption was comprised of either 1, 1.5, or 2 ads. By varying these two dimensions, the research assessed whether listeners prefer more commercial interruptions or longer commercial interruptions. The control group received Pandora’s normal advertising experience: four interruptions with 1.5 ads per interruption. The level of audio advertising remained consistent for each test group for the 21-month duration of the research.
The findings identify aclear relationship between the quantity of hours listened and the number of audio ads per hour. Interestingly, the number of commercial interruptions per hour showed little impact on listening behavior. Reactions in the test groups in the final month support these findings.
The findings identify a clear relationship between the quantity of hours listened and the number of audio ads per hour. Interestingly, the number of commercial interruptions per hour showed little impact on listening behavior.
Results across all nine test groups included:
- A 41% percent decline in the number of active Pandora listeners,
- A 41% decline in the number of days per active Pandora listener, and
- An 18% decline in the number of hours per active Pandora listener.
Of the three reactions, the most common (82%) reaction involves a decline in listeners or listening sessions.
Further, the decline in demand for Pandora’s ad-supported product that was triggered by an increase in ads suggests a potential opportunity to transition consumers to an ad-free alternative. The findings show that the correlation is highly significant and increases the older the consumer. Therefore, the probability of leaving Pandora entirely due to an increase in ad load decreases with age. It turned out that older consumers were just as likely to switch to an ad-free paid subscription as they were to leave the Pandora service entirely.
Huang, Reiley, and Riabov use of an experimental design allows for a better understanding of causal relationships. Each independent variable, the number of ad interruptions and the number of ads per interruption, is important to evaluate. Examining the impact of the independent variables and the correlation with listening time (the dependent variable) offers a dynamic learning environment. Interplay between the elements identifies the variable impacting a change in behavior.