A number of studies have been done over the last decade looking at whether social media can in some way predict behavior in financial markets. The results are mixed with some techniques working better than others. Most studies, however, measure success by whether or not their data correctly signaled price behavior the following day. It may be that these studies focus on this metric because it is so easy to use, as well as being helpful to trading strategies. But it is not so helpful to investors. Traders are looking for information about the short term price direction of securities whereas investors are concerned with earnings, cash flows, and financial strength of the issuers of securities. Many investors consider only financial statement information and market data when making investment decisions. But in order to evaluate corporate earnings going forward, it’s necessary to understand what makes companies tick—what their products are like, what the management is like, what the culture is like, what the opportunities and risks are. This involves a lot of qualitative assessment and judgement. When it comes to the valuation of financial instruments, market data helps establish a range of expectations, but ultimately a judgement call is made based, again, on qualitative assessments.
It is often difficult to pinpoint the most helpful sources of information that investors and analysts use to form opinions about companies. But undoubtedly all types of media, including social media, play a role. The fact that social media, and all types of digital media, are measurable makes them useful analytical tools. And the social reaction to traditional media, the liking and sharing of news articles for example, provides additional insights. Social media data is particularly helpful in monitoring brand equity strength, consumer interest in products, and public moods and opinions.
Brand equity is viewed as highly important to companies. Branded products have proven their success and there is a willingness on the part of companies to purchase brands. It has been argued that an increasingly large part of a successful firm’s value lies in its brand. Social media has revolutionized how companies manage their brands with Facebook, Pintrest, and Twitter standing out as leading platforms for brand management. From Interbrand 2013 State of the Industry Report:
- 4 in 10 social media users have bought an item after liking it on Facebook, Twitter or Pintrest.
- 80% of the Internet users between the age of 18 and 50 use social media.
- 80% of social media purchases take place within three weeks of sharing or liking an item.
- 29% credit Pintrest with getting products on their radars in the first place.
- Most Facebook and Twitter buyers say they were already thinking about buying the items they liked or shared.
But companies use social media for more than advertising. They use it to maintain a dialogue with customers about the quality of their products. They listen to their customers, take in suggestions, and provide customer service. Some companies engage better than others. Social media engagement metrics include company Twitter handles, the number of followers of those handles, the number of tweets exchanged, the number Facebook fans, the number of Facebook page posts, the number of likes and shares of Facebook page posts, etc. It may be difficult to quantify these engagement levels in terms of dollar increases in brand equity, particularly because the methods for valuing brand equity are open to debate. But engagement nonetheless does affect brand equity. If building and maintaining brand equity is so important to companies, it’s difficult to imagine analysts not using social media to monitor their success.
The only study I have seen linking brand popularity to stock price performance was done by Arthur O’Connor in 2010. He used data on number of Facebook fans, number of Twitter followers, number of YouTube views and certain keyword searches. This data for three publicly traded companies–Starbucks, Coke and Nike–was found to have a high correlation with stock price performance over a 10 month period from April 2010 to February 2011.
Google Trends, which provides search volume data from January 2004 to the present on a weekly basis, is a treasure trove of data. An often cited 2009 study by Choi and Varian showed that when forecasting near-term values of economic indicators such as automobile sales, unemployment claims, travel destination planning, and consumer confidence, the use of relevant Google Trends variables resulted in more accurate predictions. Google also offers Google Correlate, which operates likes Google Trends in reverse. For a particular data series, Google provides the queries whose frequency follows a similar pattern. The user interface is so simple that wide scale adoption of it by analysts is likely to occur.
The link between media sentiment and financial market behavior has been demonstrated in a number of studies. More recently studies have focused on the linkage between financial markets and sentiment expressed in social media. Twitter’s emergence in 2006 dramatically transformed the media landscape as it became evident that it’s possible to monitor moods and opinions of a large swath of the population in real time. A 2010 study (O’Connor, Balasubramanyan, Routledge, & Mellon) showed that the content of tweets closely matched the results of sentiment surveys, further legitimizing Twitter as a source for monitoring public sentiment. Extensive studies have been done correlating Twitter moods about stocks and subsequent market behavior. These studies generally focus on daily market behavior, occasionally including periods a bit longer. There are scarce studies to date examining Twitter mood trends and correlating it to longer term trends, such as consumer outlook for employment or willingness to take on credit. But that doesn’t mean that the data to examine these questions doesn’t exist, and it doesn’t mean that this data should be ignored.
Twitter is not the only social media source for public sentiment and opinion. A 2009 study by Eric Gilbert and Karrie Karahalios from the University of Illinois used a data set of over 20 million weblog posts from which was estimated a level of anxiety, worry and fear. They found that increases in expressions of anxiety were correlated with downward pressure on the S&P 500 index. Another recent study (Preis, Reith, & Stanley, 2013) analyzed changes in Google query volumes for search terms related to finance, finding patterns that the authors conclude may be interpreted as ‘‘early warning signs’’ of stock market moves. An October 2013 draft of a study by Karabulut from Goethe University in Frankfurt showed that Facebook’s Gross National Happiness (GNH) index had the ability to predict changes in both daily returns and trading volume in the US stock market from January 2008 through April 2012. Web blogs, search, and Facebook could all be helpful for longer term trend analysis.
The Journal of Finance in 2008 published an interesting study by Fang and Peress of INSEAD that found that stocks with no media coverage earned higher returns than stocks with high media coverage. The affect was particularly strong for smaller cap companies. I have yet to see a study looking at the frequency of Twitter mentions of small cap stocks and the corresponding correlation with stock returns. But this linkage would be very easy to explore.
Social media can provide a wealth of information useful to investors and financial analysts. It offers the ability to track public moods, opinions, and attitudes about issues that may affect corporate earnings. It helps analysts make better sales forecasts and better evaluate brand strength. The question is how to get social media information and how to use it. There are number of social media analytics providers and the techniques they employ vary widely. Data filtering, cleaning and sampling can have huge impacts on outcomes, as can the type of language processing used. It will likely be some time before standards and best practices develop in the field of social media analytics and this can make investors reluctant to use social media research. It is inevitable, however, that social media analytics tools will work their way further into the investment process.