As Facebook went public last week, Twitter was set ablaze with tweets about the company’s stock. Is there a correlation between the two?
While the eyes of the world were on Facebook last week, Twitter was where people headed to discuss the IPO. Given the volume and sentiment of tweets, we wanted to know: could Twitter act as a “canary in the coal mine” and predict how the stock would perform during the first day of trading?
Using our social data platform, we analysed every one of the estimated 340 million tweets per day to identify Facebook IPO-related conversations and the positive/negative sentiment regarding the stock. Come IPO day, we ran our models and came to an interesting conclusion: in many cases, Twitter sentiment acted as an early indicator of a stock price shift.
Looking at the data from Facebook’s first day of trading more closely (see chart below), we found that nearly every time the volume of negative comments increased, Facebook stock price dropped within 20 minutes. Overall sentiment related to the IPO dropped just 20 minutes before the stock market opened and started recovering 30 minutes later when the chatter picked up. The price of the Facebook stock followed this pattern 20 minutes later, increasing to slightly over $40 around 10 am ET. At 12:30 pm ET, Twitter sentiment dropped and 10 minutes later, the stock price followed the same pattern. The same thing happened on Monday when the stock price took a dive.
DataSift analysis of the first day of trading:
DataSift analysis of the second day of trading when Facebook’s stock price dropped below the IPO price (click to expand):
When we saw Twitter sentiment preceding a stock price move, it was oftentimes on the back of a news story that led to an uplift in Twitter conversations. When public tweets around a news story carried a strong sentiment (positive or negative) and the story had a potential to move the markets, it usually did.
Of course, not every story will have an impact, but for news stories that have the potential to shake things up, Twitter sentiment frequently reveals whether the markets will move up or down.
Twitter could also give traders a few minute lead on everyone else. If Friday’s IPO is any indicator, there is at least a four minute and up to 25 minute delay in the market’s response after a strong Twitter sentiment is registered.
Facebook was not the first time that we were able to see Twitter sentiment predicting the response of the stock markets. Back in February, we did an analysis of the resignation of Research In Motion’s co-CEOs. The comparison of stock price and conversation sentiment on Twitter was startling. The news broke on a Sunday evening, when the markets were closed. Comments started spreading rapidly on Twitter and across other social platforms – indicating how important the information was.
The diagrams below show the overall sentiment of conversation and demonstrate how a switch in company sentiment from negative to positive was in this case a pre-indicator of the stock reaching its support level and starting to rise.
The idea of using public sentiment and publicly available online data for predictions is not new. Google predicts the flu season based on an increase in the level of searches, Derwent Capital Markets runs a “Twitter-based” hedge fund and commodity traders follow farmers on Twitter to predict the future price of corn based on the quality of their harvest.
What’s new here is the sheer volume of data that is available, or Big Data, as it has been dubbed, plus the availability of real-time analytics and historical data, and the technological advancements that enable companies to do much more sophisticated and useful analysis.
In this case, we have been talking about social media sentiment indicating which direction the price may move, but the ability to use billions of public social conversations as a form of focus group stretches much further.
As the stock price example shows, real insights come from viewing social data alongside other data sources such as stock price. For businesses, the insights you can get from social data alone are limited. But by analysing social data alongside product sales, marketing campaigns, web traffic or inventory – that’s where Big Data and business intelligence can create the outcome that every executive wants: actionable, valuable insights.
We have the data and we have the tools. Now it’s just a matter of making a sense of it all and finding real-life applications. Twitter trading might come first, just because there is a lot of money to be made, but the world is definitely not short of ideas. What’s yours?