Not seeing aScroll to Top Button? Go to our FAQ page for more info. LOOKING FOR DATA IN ALL THE WRONG PLACES: MINING TWITTER FOR CONSUMER PURCHASING INTENT – refine+focus

LOOKING FOR DATA IN ALL THE WRONG PLACES: MINING TWITTER FOR CONSUMER PURCHASING INTENT

January 30th, 2015 by

posr-261

Emerging research on the political implications of social media suggests that, when it comes to forecasting election results, Twitter hype matters more than tweet type. Social scientists at Indiana University looked to Twitter as an alternative to traditional polling, combing the site for election-related posts from the run-up to the 2010 congressional races. They found that candidates with higher raw shares of Twitter mentions won by roughly the same margin of votes, regardless of whether the tweet reflected positively or negatively on the candidate. Remarkably, for hopeful politicians in competitive races, even negative press was good press on Twitter.

Studies on Internet buzz and consumption suggest that the predictive power of tweet frequency also applies to market outcomes. For example, in one 2010 HP Labs study, analysts found that the rate at which people Tweeted about new movies correlated to real box-office revenues. Furthermore, these Twitter-based predictions systematically outperformed in accuracy those of the Hollywood Stock Exchange, the industry’s classic gold standard of information markets.
twitter-photo
But, do negative tweets contribute to positive buzz for consumer products in the same way they do for politicians? Findings on the relationship between tweet sentiment and consumer purchasing behavior are less clear-cut. A 2011 study on Twitter mood and stock market trends found that, while the general positive or negative sentiment of tweets contributed little to forecast accuracy, the relative “calmness” of Twitter users did correlate to daily fluctuations in the closing values of the Dow Jones Industrial Average. Thus, while the emotional polarity of tweets did not make much difference, other dimensions of tweet sentiment were still relevant.

For brands seeking to attract and retain consumers on Twitter, proactive interaction, rather than passive buzz, seems to matter as well. A review of tweets from consumer electronics brands found that UK shoppers exposed to brand tweets were 65% more likely than the average Internet user to add products to their carts on those companies’ websites. Overall, Twitter users are 50% more likely to purchase from brands they follow.

Ultimately, determining the optimal balance between tweet quantity and quality will depend on the answer to a more complex question: what comes first, the consumer or the Tweet? Are box-office hits well-tweeted because they are already well-liked? Or, does the mere exposure effect—by which familiarity breeds favorites—make Twitter a real determinant of purchase intent? When the platform is flooded with an average of 58 million tweets per day, Twitter presents tricky terrain for those hoping to make a lasting impression on potential fans, followers, and consumers. But, as the findings above suggest, mining the Twittersphere for hints of socialized consumption can provide insights for the architects of social media campaigns.

 

Determining the optimal balance between tweet quantity and quality will depend on the answer to a more complex question: what comes first, the consumer or the Tweet? Are box-office hits well-tweeted because they are already well-liked? Or, does the mere exposure effect—by which familiarity breeds favorites—make Twitter a real determinant of purchase intent?

 

Leave a Reply

Your email address will not be published. Required fields are marked *