Mining for New Kinds of Data in Rocky Markets - Barron's

If last week’s selloff felt worse than a 5% decline, you can probably blame the lack of clarity behind the move. Each day, good news became bad news and vice versa. Then, late in the week, the New York Federal Reserve president called the rout “small potatoes,” which the market took to be big potatoes, accelerating the selloff. Call it a tater tantrum.

Even as traditional metrics and analyses failed to make much sense of the week’s volatility, an emerging set of alternative data was proving quite useful. That refuge of coherence came from broad-based measures such as web traffic, Google Trends, and Twitter mentions. Those numbers offer a good look at customer behavior, and they usually don’t lie, at least when put in the right hands. For investors, the trick is finding areas where the data deviate from Wall Street assumptions. Sentieo, a financial research platform, has been honing that skill for the past six years.

Last week, Sentieo flagged such disconnects for Snap (ticker: SNAP) and Twitter (TWTR) ahead of each company’s earnings results. Sure enough, both companies beat analysts’ estimates, driving huge gains for the stocks. Sentieo’s data also predicted the big beats this quarter from Netflix (NFLX) and Skechers (SKX).

Sentieo takes large data sets—say, visits to Twitter.com—and compares them with corporate results. By crunching historical numbers, Sentieo can identify the data that are most predictive. We don’t know Twitter’s revenue until a month or so after the quarter ends, but we can see real-time visits to Twitter, which correlates with revenue. In the latest quarter, traffic was soaring, but Wall Street’s revenue estimates stayed flat. Buying the stock based on that information drives alpha—the outperformance craved by money managers.

I’ve been using Sentieo’s platform in my reporting for the past three months. The firm has given Barron’s access to its data; in return, we’ve reported on some interesting findings, sourcing Sentieo along the way.

Last week, I shared the Twitter disconnect with Barrons.com readers the day before the social network reported earnings; I’ve never felt quite so confident in predicting an earnings beat.

But writing that story—and seeing the result—got me thinking. How could something that looks so obvious get overlooked by investors? Why has a supposedly efficient market missed the compelling patterns being surfaced by Sentieo?

“Most folks still aren’t looking at the data,” says Alap Shah, Sentieo’s co-founder and CEO. And those who do still have to square the information with their preconceived notions.

In the case of Twitter, portfolio managers have been burned by the stock for many years. “Some weren’t willing to stick their necks out,” Shah says. (Even after a 95% gain in the past six months, Twitter shares are still less than half of their 2013 high.)

As data science explodes in popularity, the findings are battling years of human emotion and memory. Pushing back against those assumptions requires a contrarian take. You can be sure that investors will eventually make the leap. Successful money managers find ways to reorient themselves when there’s “alpha” to be had.

Shah notes that decades ago, investors realized combining analyst opinions could provide an edge heading into earnings reports. Today, that consensus is commoditized information that investors take for granted.

For now, though, alternative data remain a niche tool on Wall Street. “The top-performing hedge funds are deep in this data today,” Shah says, “and the rest of the market is rapidly moving to come up with their own strategies and positioning around it.”

Shah, a former consumer analyst for Viking Global Investors and Citadel, is hoping Sentieo can lead the way. The platform is being used by hedge funds, investment banks, mutual funds, and financial advisors.

“We’re still in just the second or third inning of the adoption of these data sets within the market,” he says.

Everyone is looking at Big Data to solve problems these days. Promising areas like medicine and retail, though, need new people and processes to apply the data. With investing, it’s far easier. Investors get the data and make a buy or sell decision. “In many ways, the market is the bleeding edge for data science,” Shah says.

NEW TECHNOLOGIES DIDN’T ALL LIVE UP to their billing last week. Robo-advisors got their first look at volatility, and the platforms didn’t exactly shine. Clients at robo pioneers Betterment and Wealthfront couldn’t access their accounts for short periods of time on Monday, as the stock market tumbled.

It’s probably a good thing those customers weren’t able to make rash decisions, but robos have made big promises to investors, particularly novice ones. Outages are something robos will need to solve if they’re going to keep pushing their disruptive narrative.

Jay Shah, the CEO of robo firm Personal Capital, told me last week that its platform stayed up throughout the selloff. Interestingly, log-ins to Personal Capital were stable during market hours, Shah says. But the firm did see a 40% spike in traffic on Monday night after the market closed. He attributes the rush to smartphone alerts about the selloff.

“It’s like a smoke detector,” Shah says. “People got emotional and wanted to see if something was burning.”

Check back next week for the answer. 

Email: alex.eule@barrons.com

Follow @eulelog