By Bradley Hope
James Crawford sees money in weird places, including hidden in
the shadows cast by half-finished Chinese buildings.
The former Google Inc. engineer is one of a coterie of
entrepreneurs selling analysis of obscure data sets to traders in
search of even the smallest edges.
In many cases, the value isn't yet proved, but analysts say
there is a growing market among sophisticated investors for
snippets of information that aren't widely available.
Take the changing shadows of Chinese buildings, which Mr.
Crawford says can provide a glimpse into whether that country's
construction boom is speeding up or slowing down. Mr. Crawford's
company, Orbital Insight Inc., is analyzing satellite images of
construction sites in 30 Chinese cities, with the goal of giving
traders independent data so they don't need to rely on government
statistics.
Mr. Crawford, who ran Google's project to scan millions of books
and make them searchable, said the company's first clients include
several multibillion-dollar hedge funds, but he wouldn't disclose
their names or the prices he is charging.
Orbital is also selling analysis of satellite imagery of
cornfields to predict how crops will shape up and studies of
parking lots that could provide an early indicator of retail sales
and quarterly earnings of companies such as Wal-Mart Stores Inc.
and Home Depot Inc.
Other recent startups aim to use social media, crowdsourcing and
other largely unexplored data sets to provide traders with breaking
news and macroeconomic indicators.
"There is so much correlation in returns in the asset-management
world that it's becoming increasingly important to find ways to
find new types of data," said Ronnie Sadka, a professor of finance
at Boston College's Carroll School of Management.
However, analysts say traders would be wise to proceed
carefully.
"I put these into the experimental category," said Paul Rowady,
a senior analyst at Tabb Group, a market-research firm. "These
products are compelling on their surface, but there is a huge
challenge in turning it into a trading indicator."
For the parking-lot analysis alone, Orbital bought one million
images from satellite companies and built computer systems to
produce possible trading indicators out of the data.
"This is just the beginning," said Mr. Crawford, who also
previously was a senior vice president of Climate Corp., a San
Francisco company that uses weather data to help farmers predict
their crop yield. Monsanto bought that company for more than $1
billion in 2013.
Orbital Insight, whose early investors include Silicon Valley
venture firm Sequoia Capital, hopes to raise as much as $8 million
in its next round of fundraising in the next few weeks.
Dataminr Inc., which combs through 500 million tweets a day to
find market-moving news before it hits the wires, was among the
first to get traction in this field.
Founded by three former Yale University roommates, Dataminr is
one of a handful of companies to whom Twitter Inc. provides the
full "fire hose" of data that comes through its servers every
day.
Dataminr's systems categorize and analyze every single tweet in
real time, weeding out spam and comparing information against news
feeds, market prices, weather patterns and other data to determine
its significance. The systems also check if a particular user has
been reliable on certain topics in the past.
On Sept. 2, when independent journalist Brian Krebs tweeted that
Home Depot "may be latest credit card breach victim," Dataminr's
systems rapidly identified it as a "Notable Signal" for its
clients.
The alert went out to subscribers--which include 60 banks and
hedge funds--a full 15 minutes ahead of financial news wires and
before a 2% decline in Home Depot's share price.
"There is no way to ignore social media as a data set anymore,"
said Ted Bailey, Dataminr's chief executive. "A feed like Twitter
has so much information, and there is certainly value hidden within
it."
A senior banker who subscribes to the company's services said
his traders saw the first report of the Boston Marathon bombing in
April 2013 on their Dataminr desktop app. They helped some clients
put up hedges on the stock market, which began to decline soon
after news organizations began reporting on the attack.
"Clients have maybe 50 people they talk to everyday," he said.
"It's hard to differentiate yourself from the others. Having better
information can make a big difference."
Another startup, Premise Inc., makes small payments--often in
the form of credit for cellphones--to people around the world who
monitor prices of goods, giving the company an early idea of
changes in inflation rates and other economic indicators. The idea
behind Premise's macroeconomic data is that reality on the ground
is often different from the official narrative.
The company distributes much of its data free of charge, but it
also has some financial customers looking for hints on the
direction of inflation in places such as India and China. Premise
has people contributing data in 68 cities in 18 countries.
"In certain parts of the world, people don't have a trustworthy
source of information about economic conditions," said David
Soloff, CEO of Premise. "Our goal is to foster transparency."
Several hedge funds have also sought out data sets based on
where people go each day, from companies such as Placed Inc., which
gathers information from about 500,000 people who agree to be
tracked through apps on their smartphones.
Placed CEO David Shim said the interest from big investment
firms is "absolutely growing."
Access Investor Kit for Google, Inc.
Visit
http://www.companyspotlight.com/partner?cp_code=P479&isin=US38259P5089
Access Investor Kit for Google, Inc.
Visit
http://www.companyspotlight.com/partner?cp_code=P479&isin=US38259P7069
Access Investor Kit for The Home Depot, Inc.
Visit
http://www.companyspotlight.com/partner?cp_code=P479&isin=US4370761029
Access Investor Kit for Wal-Mart Stores, Inc.
Visit
http://www.companyspotlight.com/partner?cp_code=P479&isin=US9311421039