Can search results, online advertising, likes and follows predict an election?
By Hayley Tsukayama,
Election Day is here, and the final polls are rolling in as the country tries to get a final snapshot of who will be spending the next four years in the White House. The Washington Post’s The Fix notes that the latest Post-ABC poll shows President Barack Obama is up by 3 percent over Republican candidate Mitt Romney making this still a “margin of error” race.
That seems to be the general take of the social science/statistician community, as Wonkblog’s Ezra Klein notes Tuesday morning — that Obama is likely, but not guaranteed to win.
But what does Google say? Or Facebook? Or Twitter? Can you use these (in some ways) more populist tools to learn anything about voter sentiment?
MediaBistro highlighted a study from StateTech magazine that highlighted five races — the gubernatorial races for Louisiana, California, Ohio and Texas, and the Chicago mayoral race — where the winners had more Twitter followers than their opponents. The graphic then went on to point out that Obama, as of Oct. 18, had more Twitter and Facebook followers than Romney. That’s not particularly surprising, as the Post noted last week, given that Obama was a visible early adopter when it came to campaign social media.
For example, there’s search. The Wall Street Journal noted in a recent article that Google searches of “Obama” turned up customized search links about the president when users then searched for “Iran,” “Medicare” and “gay marriage.” Meanwhile searches for “Romney” did not turn up related links — a strange quirk of Google’s algorithms.
Larry Kim, a marketer at WordStream, wrote that he suspects that more people are looking up Obama’s stance more often, which he said could be a measure of how engaged Web users are with the Obama campaign.
In fact, Kim said that online engagement metrics indicate Obama will win “by a landslide,” based on a read of Web sentiment and the pure amount of money both campaigns have spent on online advertising.
Kim said that Obama appears to have outpaced Romney in Google ad spend, budgeting twice as much for search advertising and three times as much for display advertising.
“Display ads are terrific for increasing awareness and building your brand, both of which are key when it comes to rallying and growing your voter base,” Kim wrote. “[So] Obama could be getting tons of leverage out of this advertising spend.”
Finally, Kim also took a look at the social media front. In addition to the Twitter and Facebook numbers StateTech highlighted, Kim looked at YouTube for an indication of how engaged voters are with each candidate. Obama had both more views and more subscribers — people who want to be regularly updated on videos from a certain candidate — than Romney.
“In this day and age, when everyone and their grandma is on Facebook and has an iPhone, Internet presence really matters. Romney’s relatively weak Internet presence and comparatively low spending on online marketing channels could be very bad news for the Republican party,” Kim concluded.
These, of course, are all just guesses — even if they are based on interesting data. It’s a stretch to say that any of these things can predict the election, though whether or not they are correct could say something about the efficacy of social media and online advertising.
It’s at least a little more interesting than 7-11’s “unabashedly and unscientific” coffee cup race — which also puts Obama ahead, at about 60-40.
Related stories: The politics of Facebook ‘likes’ Election 2012: The tweets that matter to swing state voters Pundit accountability: The official 2012 election prediction thread