The strategy was relatively straightforward: The system tracked whether a word such as “debt” increased in search frequency or decreased in search frequency from one week to the next. If the term was suddenly searched much less frequently, the investment simulation bought all the stocks of the Dow on the first Monday afterward, then sold all the stocks one week later, essentially betting that the overall market would rise in value.
If a term such as “debt” was suddenly searched much more frequently, the simulation did the opposite: It bought a “short” position in the Dow, selling all its stocks on the first Monday and then buying them all a week later.
Some words, "debt" in particular, worked pretty well between 2004 and 2011:
[B]asing a strategy solely on the search frequency of the word “debt,” which turned out to be the single most profitable term in the study, would have generated a profit of 326% over the seven years studied—compared to a profit of just 16% if you owned all the stocks of the Dow for the whole period.
Terms like "debt," "stocks," "restaurant," "portfolio" and for some reason "color" were pretty good predictors. "Ring," "environment," and "fun" were the least effective.
Of course, the flip side of using online activity to predict market activity is that one errant tweet can apparently throw the whole thing into chaos.
Spencer Platt/Getty Images