"It was a lot of fun," said Quantifind CEO Ari Tuchman,
referring to the CIA pilots. "It wasn't clear there was a business
there."
The company Mr. Tuchman and his co-founder John Stockton are
building currently has 40 employees -- including nine Ph.Ds and a
former Facebook engineering director -- and is now effectively
coming out of stealth and looking to build its sales
organization.
Its premise is to pair social chatter with financial data --
sometimes internal sales data from clients -- to detect
correlations with revenue and speech patterns that are suggestive
of real purchase intent. (For example, it's not feasible to predict
whether people intend to buy a movie ticket if they've said they
loved the trailer; but if they also mention babysitters, they're
expressing intent, Mr. Tuchman said.) Quantifind is positioning
itself against the plethora of social-listening companies that tout
buzz and sentiment.
One potential use case would be to examine a viral event on
social media that appears to be unleashing negative sentiment
toward the brand; Quantifind could perhaps find that the purchase
intent hasn't changed, so no crisis campaign is needed. Another is
that a movie studio discovers that the group with highest purchase
intent for an upcoming release is different from the target
audience of its marketing.
"Sometimes the conclusions you reach are unexpected," Mr.
Tuchman said. "A shampoo brand might think people are buying the
product because of its scent, but it's really because of damage
control."
Quantifind currently has about a dozen customers, four of which
are movie studios, which is the vertical they launched the
marketing business with in 2012. Under Armour has been a
customer.
Military connections
Messrs. Tuchman and Stockton had a long and circuitous journey
toward building a marketing business. They left Stanford -- where
they had been doing postdoctoral research -- in 2007 and began
filing patents and applying for grants. They won two in 2009 from
the National Science Foundation and the U.S. Army. The latter would
have had them building landmine and IED detectors, an extension of
research they had done at Stanford. They ended up choosing the
former and taking a seed round from Andreessen Horowitz.
"When we thought about our backgrounds in experimental physics,
the real game, the real power, came a lot more from the signal
processing and data analytics than it did from building better
hardware," Mr. Tuchman said.
From there, Quantifind worked on several disparate projects, not
settling on a business model. It did the pilots for the CIA but
also worked with the Stanford Bioinformatics Group to research a
blood thinner called Warfarin, which can be dangerous when
prescribed at the wrong dosage, and patients' tolerance for it
varies. Quantifind applied its algorithms to gauge which gene
mutations in individuals might impact what their optimum dosage
would be.
Then in the fall of 2011, it raised a Series A round led by
Redpoint Ventures for an amount it's never disclosed and started
hiring more coders to grow its staff of seven. It also had a
serendipitous meeting at Andreessen Horowitz with T.J. Marchetti, a
senior digital marketer at Disney at the time. He mused about whether
the Quantifind technology could help movie studios sell
tickets.
"That was the first sense we got that there was actually a
commercial application for the stuff we were doing," Mr. Tuchman
said.
Pivot to marketing
The company began working with customers other than movie studios
about a year ago, largely through referrals from its venture
capital firms -- which now also include U.S. Venture Partners,
which led a Series B round for an undisclosed amount in September.
It's worked with marketers spanning categories but hasn't had one
in finance or health care yet. Mr. Tuchman also recently met with
the chief technology officer of "one of the big" U.S. political
parties but says Quantifind will keep out of politics for the time
being.
Though marketing is proving to be a promising area to build a
business around, Mr. Tuchman ultimately has bigger goals for the
core technology. In the case of health care, for example, he
envisions applications like the company's algorithms being used to
process unstructured data from the notes of doctors and nurses
who've recently admitted a patient and then detecting patterns
related to how that patient will progress.
"Marketing has been an awesome way to demonstrate this and bring
value, but our goal is to expand it to other total verticals," he
said.