Skewed Sentiment in the Twittersphere

Twitter, the popular micro-blogging and messaging service, has become a big deal – and twitter-bird-blue-on-whitenot just among social media. By its sixth birthday in 2012, it had an estimated 500 million “active” users worldwide, 140 million of whom are in the US. Based on its latest round of financing, its implied market capitalization would be $ 9.9 billion.

Activity on Twitter has become a leading indicator and Twitter sentiment is often reported in newscasts. The very volume of Twitter activity has become newsworthy. For example, the Academy Awards telecast generated 8.9 million Tweets and the Super Bowl garnered 24.1 million. Of course, many minor events are tweeted about – it seems that every conference, meeting, or presentation I attend – begins with announcing its hashtag.

A number of services such as Twitturly, Twitscoop, and Monitter allow you to track Twitter activity about your brand, organization, or cause. This can be interesting, but how relevant is it? Twitter measurement is relatively cheap and easy. Still, should you judge a marketing campaign or prospects for your latest product by reception on Twitter?

The folks at Twitter seem to think so and want to sell you a “promoted account,” which solicits more followers for you. Presumably, more followers will lead to more mentions.  If that won’t prime the pump, you can also pay for Promoted Tweets, which you can send to audiences who don’t follow you.

The question still remains, is Twitter sentiment a good proxy for what the Universe thinks about your organization or product? An interesting study just released from the Pew Research Center points to significant selection bias in  reactions on Twitter, compared to responses in statistically representative surveys. For example, the 2013 State of the Union Address was substantially more positively rated in polls (42%) than by Tweets (21%). According to the Pew researchers ” Twitter users are not representative of the public.”

This may not be shocking, but it should give one pause. How many of your users not only have Twitter accounts, but are actively engaged. Unless you’ve restricted your market to Twitterati, you may be wise to do some of the heavy lifting involved in market research.

Please Tweet me your thoughts @4ourth.