According to a study carried out among UK finance professionals, general public sentiment, as expressed on the major social networks, almost certainly has an impact on stock prices but significant barriers to using this data to achieve accurate predictions remain.

Social Media Still an Unreliable Aid to Forecasting Stock Market Movements

Last year, researchers at the BI Norwegian Business School demonstrated a potential link between increases and decreases in stock market prices and tweets published by followers, whether positive or negative, directly on a company’s Twitter page. The findings of this study seem to be in agreement with experts on the subject. New research from Colt Technology Services, which provides an information exchange website for companies, reveals that 63% of all sector professionals – brokers, heads of trading desks and so on – believe that the valuation of individual stocks can be directly linked to public sentiment expressed through the social media channels.

Data should be used with care

However, the survey of 360 UK finance professionals found that 45% of respondents regard social media sentiment as a trailing (as opposed to a leading) indicator and even though this information may have its uses, investment firms still need to make a detailed study of the millions of messages sent by Internet users if they are to decipher the trend. Explains Hugh Cumberland, Payment & Settlements specialist at Colt Technology Services : “What’s important is working out how best to leverage the data mined from millions of social media messages. And addressing anxiety over data integrity requires confidence that the tools can accurately separate credible data from the general social noise along with maliciously generated content.” Some stock exchange trading firms are already using tools to scan social media data at random and then categorise messages into one of a range of ‘public mood states’.

Usefulness of sentiment-based analysis still to be proved

However, nearly a third of the professionals surveyed believe that the need to respond fast enough to social media sentiment, together with the questionable accuracy of the information itself, constitute barriers to the adoption of public sentiment-based trading. The sheer volume of data from social media clearly poses a major challenge and moreover only 7% of the professional respondents regard social media sentiment as a leading indicator. Many firms already seek to steal a march on their competitors by investing in analysis based on more traditional data mining processes so it seems unlikely that analysis of signals emanating from social networks will be fully integrated into a firm’s systems until they can be shown to be based on reliable, high-quality data.