Clorox Launches Cold & Flu Pulse, a Socially Predictive Cold and Flu Tracker

Based on a study of online flu conversations that showed a strong correlation between these conversations and actual flu cases last year, Clorox® launched a cold and flu conversation tracker that is a social predictor of the flu. People tweet about the flu before they treat it, so they may be sharing more than 140 characters offline.

The Clorox Cold & Flu Pulse analyzes millions of conversations in real time and reports how cold and flu is trending online, including what topics are being discussed, top cities where it’s trending, and tips related to the current virality of the cold and flu. Clorox is using the Bottlenose Stream Intelligence™ system to find the words, hashtags and topics about cold and flu that define a trend and measure conversation volume. Visit to see what’s trending and remind others to #stopthespread of germs. What you talk about online can make a difference offline.

In the study of flu conversations, Clorox worked with Jonah Berger, social media expert and bestselling author of “Contagious: Why Things Catch On.” By identifying trends in flu conversation online, Clorox will help consumers identify predictors of increased flu activity and take steps to #stopthespread of the flu offline.

“We aimed to uncover what people talk and share about online when it comes to the flu and see if there is a link to flu virality offline,” says Berger. “Our findings reveal a strong correlation between the two. Online flu conversation tends to spike a week earlier than number of flu cases confirmed by the Centers for Disease Control and Prevention.”

Online conversations about flu prevention also peaked before the flu was most severe, underlining the importance of prevention throughout cold and flu season, which can begin as early as October and last until as late as March. With the Cold & Flu Pulse, you can help spread the word instead of spreading the flu. Key correlations between flu virality online and offline include:

• Starve a fever… or not? Was mom’s advice right all along? People really did gravitate toward chicken soup and tea when flu cases spiked.
• Binge watch much? When flu spiked, many of us turned to TV. Last year, every 200 cases of flu led to a tweet about watching TV and flu. #OITNB
• Burning the midnight oil. Sorry boss, but I have to leave for my health. When conversation about staying late at work was higher, it correlated to higher flu rates in the next week.
• All I want for Christmas is NOT the flu. Last year, cases of the flu spiked around the holidays and every five cases of the flu is associated with a tweet about holidays and the flu.
• Flu or a late night? A lot of conversation about flu and alcohol was humorous, but vodka doesn’t count as prevention. When conversations about being hungover and the flu were high, it correlated to higher flu rates the next week.
• Sweating it out. Could a place you go to get healthier be making you sick? When conversation about exercise was high, it correlated to higher flu rates in the next week.

Dr. Tanya Altmann, pediatrician, author and mom, recommends following simple steps to help #stopthespread this cold and flu season.

“Vaccination is the first step in flu prevention. Even after getting vaccinated – and particularly when you see people talking about the flu virus online – you should still take simple steps to stop the spread of germs,” says Altmann. “Throughout cold and flu season, wash your hands frequently, cough and sneeze into elbows and disinfect germ hot spots with a disinfectant approved to kill cold and flu viruses, like Clorox® Disinfecting Wipes. ”

Clorox® Disinfecting Products kill 99.9 percent of germs that can live on surfaces for up to 48 hours, including Cold and Flu viruses, Staph, E. coli, Salmonella and Strep.

Jonah Berger and Ketchum Global Research & Analytics reviewed weekly data on flu prevalence from the CDC’s U.S. Outpatient Influenza-like Illness Surveillance Network. They analyzed weekly mentions of various flu and other related terms in social media using the tool Brandwatch. In addition, they performed textual analysis and manual coding on the top 100 flu tweets from the past year, as well as a random sample of 1,000 flu tweets. Additionally, more than 1 million tweets related to flu using the Brandwatch tool were analyzed and overlaid on flu data from the CDC. The top 100 most viral tweets about flu within the Brandwatch data were analyzed to determine what drove sharing.

They also used a lagged data structure to examine not only the co-incidence of flu-related discussions and flu severity, but also whether one preceded the other. A series of correlations and regression models were developed to examine the relationship between the confirmed incidence of flu and social media. They then ran a series of regressions to confirm the relationships between flu incidence and social media activity. The regression beta coefficients were converted into relatable numbers for the study. Additionally, a time-series model to test the hypothesis that mentions of the flu on social media in one week can be used to predict the actual number of confirmed flu cases in the next week was developed.

Source: The Clorox Company