American Idol Upset Trumps Simon, Social Media and Search

The votes are tallied and America has spoken – Kris Allen is the next American Idol. In our post on Mashable yesterday, we set out to see if Google Trends could predict the winner. Looking at search data – volume, geographic and related search trends – in the months leading up to past season finales, clear patterns began to emerge.

The eventual winner typically had a larger volume of search queries than the runners up in the weeks before the finale. Beyond that, geographic trends revealed “swing states” that would likely lean towards one of the final two when that state’s popular and eventual third place choice was voted off. We also looked at related searches to predict where an ousted contestant’s votes might end up.

This year, the data (and most pundits) favored Adam Lambert as the clear winner, but in the end America swung the other way. Even American Idol’s king of predictions, Simon Cowell, must have been shocked by last night’s result, as he recently told Oprah that Adam would win – and that furthermore, Kris Allen doesn’t have the vocal chops to win Idol.

We looked to Google Trends as our central indicator since it reflects a mass audience and provides a deeper level of insights into geographic and related searches relative to other online channels (like Twitter) used to analyze trends. But it certainly wasn’t just Google Trends indicating Adam’s overwhelming popularity.

Even today, the morning after Kris was declared the winner, Adam has nearly twice as many Facebook fans as Kris. The top three most populated groups for Adam currently boast more than 200,000 fans; Kris’ groups –in aggregate – contain less than 100,000. Adam is also beating out the winner on the iTunes download charts. His version of “Mad World” currently sits atop Allen’s rendition of Kanye West’s “Heartless.” Moreover, Nielsen Buzzmetrics found that Adam received 16 percent more blog buzz than Kris following the May 19 finale showdown. 

So why the discrepancy between search and social indicators and last night’s outcome? For one, Adam was clearly the more controversial contestant of the two. In early March, photos surfaced of Lambert dressed in drag and posing affectionately with other men. The buzz surrounding this scandal likely fueled a portion of his search volume during the season.

Previous finalists were arguably less divisive and polarizing than Adam, so at least some of the searches were not coming from those favoring Adam. Nonetheless, his popularity and fanbase on Facebook, coupled with his larger volume of iTunes sales did seem to indicate that Adam was the favored contestant.

Additionally, Adam was the contestant favored most by the judges throughout the season. On several occasions the judges praised his talent and declared him an early front runner. This initial hype may have also contributed to more people inquiring about him on the search engines.

So maybe search alone can’t predict the future, at least not all of the time. In our analysis, the Trends held up in the past three seasons, so it stands to reason that using Trends data is directionally accurate most of the time. No method of analysis – especially one as fluid as search – can be perfect, but it does provide a unique window into what people are interested in at a given point in time.

Armed with search data, as well as information gleaned from buzz monitoring across consumer generated media such as blogs, message boards, forums and social networks, marketers can gain a much truer picture of their customer more quickly and efficiently than ever before. This immediate insight into the interests and passions of their customers can reveal important information about how consumers perceive their brand and its products, as well as those of their competition.

We may never know why America’s votes swung towards Kris (remember: that 13-year old girl can vote an unlimited number of times – and some professed to have voted “100 times” for Kris), and we’re not the only ones pondering how last night’s upset came to be:

• Celebrity blogger Perez Hilton is blaming political motivations as well as a poor finale performance by Lambert
• The New York Times says the outcome might have something to do with the “moods and mores of the country” or with the public’s taste for the pop genre
• The Huffington Post suggests that the outcome may have had something to do with religious undertones

In the end, Google Trends isn’t a crystal ball when it comes to matters of the heart – and reality show voting. American Idol results are not based on sheer number of supports alone – they’re also heavily influenced by the passion and zeal of voters. And as this video from last year demonstrates, some groups of voters are more, let’s just say … enthusiastic … than others.