Pandora ran a large-scale randomized controlled experiment to measure the extent to which our audio advertising caused listeners to enroll in a trial of Pandora’s premium subscription service. We found audio advertising tactics to drive significant levels of effectiveness, leading to a 50% increase in trial starts over a control group. We also discovered that last-click attribution, a common method of using observational data to infer causal effects of ads, would have underestimated the true causal effects of the audio ads by a factor of 3. Audio ads are particularly susceptible to this underestimation because they capture listeners’ attention and influence their behavior without necessitating a click.
Get your copy of 'The Power of Audio & the Weakness of Last Click Attribution' today.