We have a recent article on audio mining and marketing, see here. It is an exciting topic. Unstructured data is increasing being analyzed. Audio data, a key type of unstructured data, has great potential to give insight.
People Use Their Voices To Persuade
People often use their voices to persuade. This can be positive or negative. (Grima Wormtongue, from the Lord of the Rings, is the first example that comes to my mind but that says more about me than the topic probably.) For our research we analyzed Kickstarter videos. In these an artist/creator makes a request for funds. They want to be persuasive. Clearly what they say is likely to matter but so is how they say it.
There is a significant literature on vocal tones. What do people generally find persuasive? Three features, we thought, were likely to be positives in a speaker’s voice. These were indications of focus/concentration, lack of excess emotions, and lack of stress. How to measure these?
Audio Mining And Marketing
We scrapped videos from Kickstarter in both the music and technology categories. These videos have accompanying texts which we performed an analysis upon, e.g., sentiment analysis. (For more on text mining see here.)
We then ran the audio from the videos (not the music) through audio mining technology from Nemesysco. This software considers the vocal tones used and then gave us reports on various aspects of the speakers’ voices. Three aspects corresponded to what we were interested in — concentration, stress and excess emotions.
Our analysis showed largely what we predicted. The strongest finding being an association between whether the speaker sounded like they were focused on the task and funding success.
From Association Towards Causation
The audio mining software had given us the associations. These would likely be enough for many practical applications. Managers may just want to know what correlates with success. Indeed, the advice we generated seems sensible, if not too surprising. Concentrate when you are making a pitch and, as you want to be persuasive, try not to sound too stressed or excessively emotional.
For an academic contribution we needed to show what was driving the sucess. The team (Mansur really) therefore did some experiments to track what was driving (mediating) the impact of perceptions of focus. (Mediating can be thought of as connecting up. So here we are looking at what links perceptions of focus with funding succces).
Consistent with prior research we showed that perceptions of competence drove success. With the right vocal tones you could demonstrate focus which led to perceptions of competence which led to funding.
Potential For More Research Is There
One of our key aims was to promote more audio mining research.
Fascinating new methods can quantify key elements of the human voice, promising benefits for numerous fields. While consumer research can take the lead in this area—gaining insight into a speaker’s thoughts, identifying presenters’ styles, and (as we do here) assessing persuasion effectiveness—any further research on both using and validating various audio mining techniques would be valuable.
Wang, Lu, Li, Khamitov and Bendle (2021)
We expect to see much more work in this field.
Read: Xin (Shane) Wang, Shijie Lu, Xi Li, Mansur Khamitov, Neil Bendle (2021) Audio Mining: The Role of Vocal Tone in Persuasion, Journal of Consumer Research, Advance Articles, https://academic.oup.com/jcr/advance-article-abstract/doi/10.1093/jcr/ucab012/6147021?redirectedFrom=fulltext