Who’s listening to Canadian social media conversations, and why

I’m willing to bet that when someone stuck on the 401 takes out their iPhone and posts something like, “This gridlock sucks!” they have no idea they are contributing to one big massive research project. Be aware, however: everything you say on social media services may be tracked, collected and ultimately, judged.

Earlier this week IBM released the latest results of its “Social Sentiment Index,” which uses complex analytical software to take the pulse of what people are saying about particular subjects, and how they collectively feel about things. In this case, the subject was traffic congestion. The mind-blowing news? People really hate it! Of course, there’s a little more to it than that.

The Social Sentiment Index showed, for example, that people in Toronto were particularly irritated by lack of parking, while people in Montreal were relatively blasé about the volume of cars on the road. IBM canvassed not only Twitter but blog posts to capture the prevailing moods, which the company suggested could help local governments identify opportunities to make improvements in the way major routes are set up or in public transportation services.

Social listening

This concept of “social listening” is an interesting new field that I’ve seen primarily used by marketing departments to have a better sense of whether their advertising messages are being perceived the way they want, or by customer service organizations to repair any reputational damage a company could suffer if their products or services malfunction, somehow.

To IBM, however, social listening could potentially do even more than that. Jean Francois Barsoum, who is one of the leaders of IBM’s Canadian Smarter Cities initiative, said such analytics could influence how much of a particular product to make available at a given time, what areas need better delivery of products and services or a chance in the size or features of a product or service.

“Social media information isn’t going to replace the traditional loop detectors telling you there’s someone at a particular intersection or that you need to change the light,” he said by way of example. Surveys and other forms of proactive research is still important. “You will still need to balance the two.”

You’ve also got to approach social listening in such a way that you also don’t totally creep people out. That said, many people are blissfully unaware this it's going on. A few weeks ago another company in this area, Netbase, held a Webinar featuring Jackie Anderson of J.D Power and Associates who presented some research that showed 32 per cent of people don’t know their social media comments may be followed by outside organizations.

They also have some firm opinions about how they want to interact with such organizations. “Sixty-four per cent want companies to respond to social comments only when spoken to.” For instance, if you post a comment on Twitter like, "Hey @Ford, you should make better engine lights!" it’s OK for Ford Motor Co. to reach out directly via the same channel. That’s a lot different than monitoring, Big Brother-style, for anything anyone says about a topic that’s associated with your company’s interests.

There are two other issues worth thinking about here. I would suspect that social media, by its very nature, tends to be a channel more likely to feature negative sentiment than positive. I’m a regular user, and maybe there are lots of people out there tweeting, “Loving my commute today!” but I never see it. There is a potential bias in the heuristic that companies will have to include as part of their analysis of social conversations.

The other issue is more about signal versus noise. Not everyone is on Twitter, Facebook or similar services, and even if they are, they may be choosing not to speak. On April Fool’s Day, for example, I saw a lot of people write that they were staying off social media because they had no patience for all the phony jokes and pranks that were being posted. If it’s that bad on April Fool’s Day, how many phony or inaccurate sentiments are being conveyed in social media on an average day, and how do you filter that out of the analysis, and what of the people who aren’t active online?

We all pay attention to road rage but there’s a lot of other sentiment among drivers that never manifest itself. Ultimately the biggest challenge facing the evolution of social listening is silence.