Opening Cross: Don’t Undervalue the Unusual

What kind of “unusual” datasets? Esoteric documents or data about industries and sectors off the beaten path, to be sure, but even social media and sentiment—now commonly accepted by some, while avoided by others—can be a challenge to get hold of, at least in a demonstrably usable format.

Only a few years ago, a panelist from a European bank at one of our conferences scoffed at the idea of using tweets as an input to trades, saying that a regulator would never accept Twitter data as reliable enough to justify a trading decision that might impact, say, a pension fund. Shortly thereafter, the US Securities and Exchange Commission reversed its position on social media, deciding that it was an acceptable form of communication to publicize market moving information, after Netflix chief executive Reed Hastings tweeted important information about subscriber figures.

And if you doubt that traders take Twitter seriously, remember the so-called HashCrash, when hackers took over the Associated Press’ Twitter feed and tweeted false reports about an attack on the White House. The AP quickly regained its feed, but the damage was done: the tweets had indeed moved markets. 

In some cases, vendors such as Nous have developed trading games that encourage players to submit their price forecasts in return for cash rewards, while the output of those forecasts is a feed of information to trading firms that shows them where and when the retail market is headed, so they can plan services to meet demand, or plan to trade against it. 

The latest in this vein is crowd-sourced earnings estimates provider Estimize, which has unveiled Forcerank—a tool that allows users to rank stocks against one another in pre-selected portfolios to create a crowd-sourced ranking of what investors believe will perform well or badly, giving others the option of following those recommendations. Meanwhile, social media analytics vendor Eagle Alpha is expanding the proprietary content that it offers, and has begun sourcing data on housing and auto manufacturers. For example, negative sentiment around an auto manufacturer may indicate issues with the manufacturer’s cars that could expose it to delays and expensive recalls.

All of which leads me back to AlphaSense, since trying to get value out of social media data is a great example of Big Data at work: And as more sizeable Big Datasets appear, search will continue to become increasingly important to help consumers understand what data exists, where, and how it can be used. 

Since traders are now using items such as FDA approvals, earnings announcement dates, and crowd-sourced stock rankings to drive investment decisions, I wonder whether other types of rankings, such as awards, would also have value as a predictive indicator of a company’s success, value and stock price? Specifically, I’m referring to this year’s Inside Market Data and Inside Reference Data Awards, for which the voting and call-for-entry processes are about to begin. You’ll notice some changes to this year’s awards: we’ve dropped a few categories, added some new ones, and moved some from the online poll to the call-for-entry section, which we hope will make the awards more dynamic and give companies outside of the very top vendors a chance to shine. So if it’s your time to shine—or you know an individual or company worthy of recognition—get voting and nominating now!