Personalization of Audio: Shorter Audio Ads

At Pandora, we’re constantly seeking to improve listener experience and advertiser performance on our platform. We truly believe that what is good for our listeners is also good for our advertisers. It was in this spirit that we conducted a test with Nielsen Entertainment to understand audio ad effectiveness between broadcast radio and digital radio; however it was the ad length findings that caught our attention.

This test revealed findings that shorter audio ads (8 seconds in length) prompted high recall with younger demos, specifically 13-24 year olds. Although the study wasn’t 100% conclusive, it made us pause and consider how we might apply these findings on our platform.

There’s no denying communication is different than it once was. The hand written note gave way to the email, which soon gave way to the text. Now, we rely on acronyms instead of phrases, emojis instead words, and gifs to make us really “LOL”. With interpersonal communication constantly evolving, should our marketing messages follow suit?

As a leader of audio, Pandora strives to provide informed, data-backed recommendations to our advertisers to best communicate with their audience in today’s world. To this end, we’ve set out to conduct a series of creative audio tests, the first of which focuses on the effectiveness of the shorter audio ad length. We’re collaborating with strategic partners, such as Orkin, ZipRecruiter, and Subway to understand how spot lengths can impact listeners and advertisers.

No two Pandora experiences are alike. We leverage the Music Genome Project, which analyzes over 450 attributes per song, to stream music that caters to individual users and their personal music taste. As such, our approach to ads is also just as personalized and flexible. Variables such as advertiser vertical, message objective, and target audience are unique to each of our advertisers. So, we’re now asking ourselves, why should audio length be one - or two - size fits all, given these differentiators?

Today, our hypotheses is that there is value in utilizing a shorter length audio ad as a part of a media mix of varying ad lengths. We are looking to understand audience trends and recommend optimal ad lengths based on both our client’s target audiences, messaging needs, and campaign objectives.

We’re heading into our fourth week of our shorter ad test, collecting first, second, and third party data with our partners. So far, what we’re seeing is in sync with our hypotheses:

1. 10 second and 30 second audio ads both successfully drive ad recall: While lifts at this point are not as strong as 30 second versions, we’re seeing double digit lifts in ad recall amongst 10 second formats. These shorter form ads do drive impact and we recommend using them as part of a broader media mix.

2. Shorter ad formats seem to resonate better amongst younger demos:  Ad recall rates were highest amongst the 25-34 demo.

3. Shorter formats do not necessarily equal less time spent with the brand: In one test, we found that the shorter ad format actually drove higher time spent on the landing page versus those who were exposed to the 30 second ad.

As the industry continues to think about shorter ad formats, we think marketers shouldn’t just be focused on buying time, but now also focusing on buying effectiveness as well. If a message can be communicated in a shorter time period, you may want to consider that option. Should you squeeze a 30 second ad into a 10? Probably not, as there are many variables to consider - but we think your media mix might look different in the years to come. Our ongoing audio ad length testing will allow us to provide our advertisers with smart, informed recommendations around spot lengths, best practices for when and how to run which spot length, and most importantly - how it can map back to our partner’s KPIs.

To learn more about Pandora’s suite of advertising offerings, get in touch with us here.

^{Source: Pandora Internal Metrics}