Adding emotional elements to a tweet doesn’t make it more viral

By February 11, 2012
eggs with emotional faces

We can now predict what articles get shared on Twitter based on a few features. Sources, subject and mentioning celebrities all help posts to get retweeted. Tone, however, isn’t a criteria.



A person’s clout has been generally accepted as being the main reason why a specific piece of content is shared or becomes viral and on Twitter. But a Research from HP Labswas conducted showing that content also matters and more specifically the source, the category, and mentioned names. Adding emotions to a tweet, however, did not prove very useful. In other words, the tone is not important; what matters is the content. Results were effective in predicting the range of popularity for a given tweet to approximately 84 percent accuracy. While it is tempting to focus only on viral tweets, a large number of tweets spread at a medium level and target interested and informed readers.

The tone doesn’t matter, content does

"Brand matters; information matters; tone, however, doesn't seem to make much of a difference when it comes to sharing," Megan Garber noted in her article about the research. The HP Labs team found that emotional components of a tweet didn’t seem to have much influence on how frequently it was shared. Subjective tweets and objective tweet seamed to be shared in the same way. In other words, subjectivity is not necessarily helping a tweet go viral, which is worth noting since symbols of emotions or mood in the shape of emoticons are widely used across social networks. Brands often build their brand image by adopting a specific tone and “voice” for the brand. Those elements are crucial for building a brand image and addressing a specific population, but as far as Twitter goes, they don’t suffice and solid content remains the most important for a tweet to spread online.

Reliability is crucial, but being a traditional source isn’t a necessity

The source of the tweet is important – a reliable source is always more likely to be retweeted than a not so reliable one. Articles were collected by Feedzilla, and broken into categories. Data covered articles from 1350 unique web sources, and each was scored on its historical success on Twitter based on the average number each article was shared, or its “tweet-density.” However, using various methods to construct a consistent tweet-density of a source, the research team found that top traditional news sources were not necessarily high scorers. While these sources (e.g. Christian Science Monitor, Wall Street Journal, Reuters) publish many articles, social web sourced articles (from sites like Mashable, the Facebook and Google blogs) had more tweets. 

Legal mentions © L’Atelier BNP Paribas