This week’s CBC tech column was a look at how tricky it is to measure the direct impact of social media on elections. There’s a version up at, and one below, for posterity. You can also listen to an MP3. [audio:]


Some called it Canada’s first “social media election.” But now that it’s all over, I have to wonder: did any of the tweets, Facebook groups, or YouTube videos actually make a measurable difference?

Certainly, many of the candidates had active social media accounts, and the #elxn41 Twitter hashtag was on fire during the election. But online activity is very difficult to link with the offline behavior of marking a ballot with a golf pencil. Ultimately, once you’re behind that cardboard partition, voting is a private action. A secret ballot is, after all, secret.

In many ways, this stands in stark contrast to how things work on the web. Social media marketing and advertising isvery measurable. Hyperlinks can easily be tracked to generate a myriad of stats and reports on our collective online behaviour. For example, this link to the Elections Canada site displays all kinds of information about itself. Recently, on Spark, we explored the profiling potential of personalized email coupons. This sort of trackability is widespread online, but doesn’t follow to the ballot box.

So if it’s not measurably effective in getting politicians elected, why all the social media hubbub? According to researcher Daniel Kreiss, it’s because social media can be useful in tracking “other forms of electoral influence” that are very measurable. For instance, campaign donations. All of the national parties have online donation pages. Traffic to these pages is easily measurable. Another metric campaigns can track is the number of voters who sign up for email newsletters, which can easily be targeted and customized, riding by riding. Or, mobilization statistics – how many people signed up online to volunteer. This is all trackable, measurable and useful, even if it’s not directly linked to actual votes. Of course, parties generally keep this kind of information pretty close to their chests, so we don’t necessarily hear much about it.

There are other, less quantifiable benefits of social media, too. Towards the beginning of the campaign, I spoke to digital strategist Mark Blevis, who talked about using Twitter as a focus group: “In fact, it might even be what I would argue to be the best focus group, largely because it’s happening in real-time. It’s happening quickly.”

Another factor that’s not directly measurable — but is definitely worth looking at — is the use of social media to encourage voter turnout. I saw a lot of this on Twitter during election day. Politicians saying, “Where is your voting station? Do you need a ride to the polls? Call us!” As of early Tuesday, Elections Canada reported 61.4 percent voter turnout, an improvement over the 2008 numbers, but still the third lowest in Canadian history.

Could a tweet change your mind about who to vote for? Probably not. But could a tweet encourage you to get off your duff to cast a ballot? Maybe.

I think it’s important to remember that compared to traditional media channels (like TV advertising, for example), most new media election budgets are relatively small. The focus is still on old media, with its large reach. But according to Daniel Kreiss, “There has been, by leaps and bounds, an increasing recognition that new media can deliver major returns on investment in particular areas, namely fundraising and volunteering.” Success in those areas is much easier to measure in a quantifiable way.

So then, could we make the impact of social media on election results more measurable? Of course. For the past several elections, there’s been debate surrounding the idea of online voting, in which voters would log into a website instead of a marking a paper ballot.

Imagine a hypothetical scenario in which Canadians can vote electronically during a federal election. A candidate could post a trackable “Click here to vote” link on their Facebook page or their Twitter account. Those links could easily be tracked, and the link between a candidate’s social media strategy and the act of voting could be made much clearer. Now, I’m not saying that scenario should happen. In fact, I’m seriously creeped out by the thought of it. But again, it’s technically possible, and not much different than existing online marketing campaigns and conversion metrics.

So, we may never be able to truly measure the direct impact a tweet, Facebook post or YouTube video has on an election’s outcome. But I’m fine with that. So is Daniel Kreiss.

“That’s sort of one of the beauties of politics. That it can’t be totally rationalized in that way, of knowing exactly what is going to influence the electorate,” he says. “You know, it’s not quite a science. But it’s not quite as messy as an art. It’s somewhere in between, I think.”