Two new academic studies cast significant doubt on the marketing of Twitter. The implications for marketers are that Twitter is harder to use and less truly representative of the on-going conversation than we thought.
Marco Camisani Calzolari, a professor in Milan, studied 10,000 random tweets for each of 39 well known brands including Dell, Blackberry, Coke, Ikea and Vodofone. He discovered, based on memes and language used that 45 percent of tweets are likely to be non-human bots. Basically half of all tweets are bogus.
We already know a small minority of activist tweeters generates the vast majority of tweets. Add to this anxiety, a huge volume of brand-centric tweets generated by machine and this calls into question both the reach and credibility of Twitter. And while its understood and expected that marketers will try to game any channel or platform, these numbers are scary. It raises the prospect that trending stories are heavily manipulated and that Twitter, as a barometer of genuine popular sentiment, may be seriously flawed.
This data comes at a time when Twitter is seeking to find new ways to monetize its platform and brands, feeling overwhelmed with the 24/7 always on torrent of tweets, are re-evaluating the reach, influence, utility and value of this social network for marketing purposes.
Another mathematical academic study by Twitter’s own research scientist Jimmy Lin, concludes that trending topics can quickly change from one term to another as the conversation builds momentum or as new information or news breaks.
On average, only 17% of the top 1000 query terms for one hour were still in the top 1000 in the following hour. The pattern of words or hashtags is not easily predictable, even as topics or stories work their way through the news cycles. “There does not appear to be cyclic patterns in day-over-day churn” in the use of keywords or hashtags on any given topic.
The implication is that its much harder to keep up with, influence or leverage trending topics because “the phrases or hashtags become trending topics primarily as a result of velocity – the rate of change – not the sheer volume.” The pattern is like a super nova. A topic bursts in to the stream with high energy then dissipates almost as quickly as it appeared. And while a topic may decay at a slower rate than the words used to describe it, trying to keep up with it in the moment or over the course of its run, in ways that amplify, intersect or influence the conversation is increasingly more difficult.
Bottom line: Marketers need to be wary about the real volume and real human inputs behind tweets. And they need to understand that it may require enormous effort and resources to keep current, involved or influential on trending topics or news stories.