Engagement with Political Viral Videos

With the United States Presidential election just over a month away, the airwaves of television and the tubes of the Internet are full of messages from the candidates. Luckily for them, these messages are being received, as there seems to be wide-spread interest in this particular election. Marketingvox has a nice recap of the viewership numbers from the Democratic and Republican National Conventions. McCain may have had 500,000 more television viewers than Obama during the convention, but how did he fare on the Internet?

Not as well, it would seem. Visible Measures collected data on viral video views just after the conventions and had some interesting, but unsurprising results. As of September 8, 2008, Obama had 1.3 million more views than McCain. Unsurprising, because there is no doubt that social media tends to leans towards the left. Jeremiah Owyang, an Analyst at Forrester, conducted a small survey that highlights this phenomenon (also check out Josh Bernoff’s post on the matter).

In his post, Owyang states that “frequency [of viewing videos] isn’t telling… sentiment is.” Here at OTOinsights, we whole-heartedly share this opinion. The core application of Quantemo is to measure a person’s emotional reaction with media.  As such, we decided to throw on our political hats and conducted an experiment to see how engaged people are with political advertisements.

Thirty participants were shown nine videos on YouTube. These included a music video, a smear ad, and a regular ad from Obama, McCain, and a neutral ‘control’:

	McCain	Obama	Neutral
Music	Raisin' McCain	Yes We Can	Where  is Matt?
Ad	Global	Hands	Airbook
Smear Ad	Taxman	Seven	Lenovo

As with social media, the participants were either moderates or liberals (93%) who planned on voting for Obama (63%). Even though most of the sample was in Obama’s court, we had some interesting findings.

The Quantemo Engagement Index (QEI) is OTOinsights’ proprietary index to determine users’ engagement with media.  In short, it is a combination of the users’ opinions of the aesthetics of the media, their emotional reaction, and their physiological reaction.  The QEI results for the political viral video study can be found below:

It was not surprising to find that the McCain videos had the lowest engagement scores. But it is interesting to note how the QEI continually drops as we progress from a neutral music video to an ad with malicious intent. This trend occurred with both candidates and the ‘control’. Additionally, we found that no smear ad elicited a positive response.  Perhaps smear ads are not the best way to engage one’s constituency.

What makes an Internet video viral is its ability to highly engage the individual so that he wants to share it with his friends. These music videos seem to have engaged the participants more than either of the advertisements did (when comparing McCain to McCain, and so forth). We asked the study participants whether they would send the video to a friend, and again, the music videos would have been sent more than the ads, as shown below:

So, in summary, it would seem that not only are the music videos more engaging to people, but they have a better chance of getting sent a friend. Additionally, since social media tends to skew to the left, one would expect more of Obama’s videos to get sent around the world, and this is definitely the case. You can keep track of the current numbers over at techPresident. As of September 25, Obama had 51.5 million more YouTube views than McCain. That’s a pretty telling figure. A word of advice for McCain: if you want to engage more users on the Internet, it may be time to create more viral videos that users will want to send to their friends.

In the near future, we will be compiling the full results of the study and will make them available here on the OTOinsights blog.  If you would like more information about the study, OTOinsights, or Quantemo, please be sure to contact us.

Comments

[Jeffrey Bardzell]

This is some interesting stuff. There certainly is lots of conventional wisdom that suggests "going negative" gets results. It's encouraging to get some data that at least confirms what most of us feel, which is that smear ads are a huge turnoff. It still leaves the mystery about why, if we are so turned off, we kinda sorta start to believe them anyway.

The findings about music-based ads is also extremely interesting. Music videos is a popular form of Internet video to be sure, and it's an important finding that preferences for music-based videos translate to a tangible behavior--forwarding the video.

Maybe the singing smear ad is the next major innovation in political viral videos!

Seriously, one promising implication of your work here is the possibility that genres of Internet video may, in the future, start to affect political advertising. Given the wit and self-referentiality of a lot of Internet video, that would be a very interesting development--and certainly an improvement over propagandistic smear ads!

Nice work!