.jpg)
Knowsis is the first service provider that can prove analytics, based on listening into social media, actually works. Rigorous testing by the data and analytics firm has established that a trading strategy utilising social media sentiment data significantly outperforms one that does not. The results have been corroborated by a number of leading quantitative funds. It is the first time proof of this kind has been provided in a real trading environment.
The tests were based on a portfolio of stocks. Each stock was given a sentiment rating based on social media activity and this was used to devise simple trading strategies. All trading strategies generated returns over and above those achieved by major indices for the same time period. One strategy built using Knowsis' analysis, for example, produced an average annualised return of 46% compared to 13.6% from the S&P 500.
Knowsis’ success is down to the way it mines, manages and analyses the vast quantities of data produced by social media. Its methodology is based on one overarching question: will sentiment make a difference to asset prices? Crucially, Knowsis’ analysis looks to identify underlying behavioural trends rather than instigating a “trade by tweet” approach. This allows for the data to be validated and provides the best possible control mechanism against false or misleading information.
Knowsis CEO, Oli Freeling-Wilkinson, says: “The markets are beginning to open their eyes to the potential in the vast quantity of data produced by social media. Some have been spooked by recent events and there have been a few false starts, but the results of our extensive testing and end of day approach prove that there is now a genuine way to extract significant value from this big data source.”
A trader from a quantitative trading firm says: “Previously I was sceptical of the role that social media, and the sentiment derived from it, could play in the financial markets. I was not sure it had a place in a quantitative trading strategy. So when Knowsis asked me to review their data, I proceeded with great caution. As a quantitative trader I rely on hard facts, empirical evidence and statistically robust data. Knowsis’ data stood up to this and presented a clear correlation between validated social media sentiment and improved returns. Our extensive testing and analysis confirmed what Knowsis had already proved: social media sentiment can significantly enhance a trading strategy.”
The quantitative trading firms performed statistical tests to verify Knowsis’ hypotheses that social media sentiment signals could predict market movements.
A representative from a second quantitative trading firm supports the initial conclusions: “I approach any new data source with curiosity – never disregarding it or taking it at face value. I had heard about social media sentiment data but hadn’t yet come across something that would work for me. I was intrigued to see if Knowsis’ approach to sourcing and validating social media data was different. For me, the litmus test was whether it picked up on signals when controlling for other significant indicators. We conducted preliminary tests, albeit detailed, which indicated that Knowsis’ data does indeed provide alpha.”
Knowsis has been developing its technology and methodology for mining and analysing social media sentiment for over two years. Its analytics are used to support trading, investment and risk management decisions across the breadth of the financial markets.
