Have you ever wondered how a given piece of audio will perform, but don’t have the time to run a full test? Perhaps your new TV campaign is launching in 12 hours? Or you’re meeting with the CMO in 20 minutes? Or you’re just debating with a friend over beers who sounds better for voiceover work? (Yes, that happens.)
Introducing the Veritonic Predictive Algorithm. Upload any audio asset or set of assets, and within seconds the platform will give you a prediction for how a General Population audience will score it. If you’re just looking for a quick read, or to settle a bet, you’re done. But if you need results from a specific audience — auto intenders, say, 18-34 year-old females in the Midwest — you can promote the tracks to a full test, with human Marketing Response Data.
How Does It Work?
First, the platform identifies what kind of asset is it: an audio ad, a voiceover track, and so on.
This ensures that predictions for voices are always based on voice data, while predictions for audio ads are based on audio ad data.
Then it uses sophisticated Machine Listening technology to compare the audio file to similar assets that the platform has collected data on in the past. The evaluation looks at the characteristics of the audio file, including the sound frequencies in the file, volume, and more, all using cutting-edge digital signal processing approaches. The algorithm then looks at the historical scores for specific attributes.
Lastly, the platform returns these attribute scores, along with an overall score, in a Baseline Report: the same easy-to-read report style that tests with human Marketing Response Data are presented in, along with the ability to then create a test if necessary.
Where Do I Find It?
The predictive algorithm is now built into the standard test-building process. This means that to create a Baseline Report, simply click on the “New Report” button in the left-hand nav. Follow the directions to upload tracks and you’re on your way.
What If I Want Different Attributes?
To maximize the training data set behind the Machine Listening algorithms, we’ve limited this initial release to the most common attributes our clients have used to test assets. Over time, as we refine the algorithms and grow the database we expect the set of attributes available in Baseline Reports to grow.
What If I Need A Specific Audience?
As with the set of initial attributes, we’ve limited the predictions to a General Population (GenPop) audience to maximize the system’s accuracy. If you need a specific audience, you can quickly promote it to a full test.
What Else Do I Need To Know?
You’re all set! This functionality is live in all Veritonic client accounts. If you have further questions or feedback, please contact your account manager or the client success team. We’d love to hear from you!
New to Veritonic? Get Started with Baseline Reports
Fill out the form below to get started using Baseline Reports for audio effectiveness analytics.