While measuring a heated political race in Toronto, all social media factors pointed to a particular candidate to win - except one. Is YouTube an indicator of political influence more powerful than other social media outlets, or is there something else at play?
During the last week of the Toronto mayoral elections, an interesting discovery was made in reference to Internet Popularity. We had monitored the elections and reported each day over a five day period prior to the election which candidate was more popular online. 9th sphere's Toronto Mayoral Web Popularity Poll, looked at social profiles like Facebook, Twitter, as well as online activities like YouTube video views, and positive versus negative postings on the first page of the search engine results. We looked at these items as support for positive reinforcement that may provide insight into which candidates were more popular and therefore win the election. Interestingly enough, the results that supposed one candidate over the other for the entire reporting period was the candidate who didn't actually end up winning.
What We Found
In the five day period before the election, one candidate had more Likes on Facebook, followers on Twitter, searches on Google and the least amount of negative postings on the first page of the SERPS. These results all favored this candidate for the period in which the results were gathered. One would think that this same candidate would be the winner of the election if web popularity was indicative of the physical world of votes. However, this candidate lost the real election and by a large margin - over 47%, a difference of approximately 93,000 voters (source: Globe and Mail). The winning candidate surpassed his opponent in only one of our web popularity poll criteria - YouTube Channel Views. He had a 27% margin of difference. There is a well-known scientific testing methodology which states that all factors need to remain constant except for one to accurately measure the reaction of the changing factor. In our web popularity poll, one candidate was the dominant figure in all measured criteria, except YouTube video views. The only anomaly in our findings, coincidentally was clearly in favor of the actual winner. This either allows us to see the potential influence of this one criterion, or leaves us wondering about chance and statistical relevance. Canadians online users are both large in numbers and a more web savvy bunch than others. Over 80% of the population has Internet access and more so in Toronto. About 51% of Canadians between the ages of 16 and 34 purchased a product online in 2009 (source). According to many statisticians, the demographics of the early polls for Canadian voters for the candidate who won the mayoral race was largely made up of seniors and baby boomers. The candidate that lost, the one who had many more signs of web popularity, had voter demographics largely consisting of university voters (source). So while university students and this age bracket have a greater web presence, and likely contributed to the candidate's online popularity, they do not have the ability to surpass the large population of seniors and baby boomers in actual votes at the polls. One could conclude that when looking at positive web popularity criteria, Facebook, Twitter and the number of searches don’t necessarily correlate to popularity in the physical world. However, YouTube Channel views would. For argument’s sake, if there is a correlation between voting in the physical world and popularity online, YouTube video views would be more indicative of strong support for a winner. Watching videos is one of the most popular online activities. In our theory, those that engage in social networking are typically stronger brand advocates and part of the conversation. These supporters are loyal, but fewer in numbers. Most voters are not as engaged and won't make as strong a connection to the candidate. Online videos are more similar to TV; they reach a much larger market, demographic and have a bigger impact. If people are not sure who to vote for, they may prefer to watch a video online and then read their platform materials. Although the circumstances of this election illustrated a very telling story to social media marketers, we will be watching the next elections to see if this correlation continues.