Frost Data Capital is like a modern mining company.
It looks for gold in the vast piles of data that companies of all sorts amass.
The payoff comes when one of the startups Frost Data fosters at its offices in San Juan Capistrano finds a new way to tease out pertinent data and make sense of it—discerning trends or identifying untapped opportunities in the marketplace.
The operation focuses mainly on the industries of technology, energy, manufacturing and healthcare. It creates startups—two dozen to date—that focus on what’s known as “predictive analytics,” proactively identifying problems and solving them.
It’s about to get bigger, too, as Frost Data works on a new fund of up to $250 million and expects to soon announce new corporate investors, according to Managing Partner Stuart Frost.
An example of the sort of startups created so far is NarrativeWave, which provides software for companies to manage their critical pieces of industrial equipment, glean data from them, and deliver insights to enable better operating decisions. Another is Sentrian, now in Aliso Viejo, where it’s engaged in preventive healthcare with the goal of keeping Medicare patients out of the hospital through software that uses “biometric monitoring” to remotely detect their conditions, such as temperature, brain activity and heart rate.
They and the other 22 startups launched from an initial fund of $55 million range from two to 62 employees.
Frost Data’s main partners, Fairfield, Conn.-based conglomerate GE Corp. and Mountain View-based software maker Symantec Corp., also are clients. Frost Data solves problems for the partners, as well as their clients. The partner relationships also involve other aspects: GE and Symantec invest along with Frost Data in some of the startups and collaborate on a range of strategies, from go-to-market development to access to customers.
The Road to Data
Stuart Frost grew up in an entrepreneurial family in England and obtained a degree in electronic and computer engineering. He started his first company, Select Software Tools, in 1988 when he was 26.
The company branched out to Irvine in the mid-1990s. He headed it for about 10 years, raised venture money, and guided it to a public offering on NASDAQ in 1996.
He left Select Software in 1998 when the company had about $30 million in revenue, and made a couple of stops before creating DatAllegro Inc. in 2003. He ran the analytics outfit until Redmond, Wash.-based Microsoft Inc. bought it for $275 million in 2008.
Frost spent two years with Microsoft before he started Frost Data in 2010 as a new challenge, he said. It’s not a venture capital firm or an incubator—most of the startups are formed with the expectation they will stay with Frost Data throughout their existence. He amassed the $55 million venture fund from high-net-worth investors in Orange County and around the world, he said.
Machine Learning
Three-quarters of the 24 startups Frost Data has created are still in operation. The majority of them employ machine learning, a type of artificial intelligence that enables computers to “learn,” without human input, to improve the algorithms they use over time.
“When you’re dealing with large volumes of complex, dynamic data, you can’t build a simple algorithm and expect it to work successfully and keep functioning in the same way for the next 20 years,” Frost said. “It may not be right even on day one. With machine learning, you build a model and let the algorithm adapt and change this model based on experience and the data.”
It’s really all about “human-augmented machine learning,” he added, in which various experts can “nudge” a system and help the machine make sense of the automated findings.
“This combination of the best aspects of machines—fast analysis of huge quantities of data and spotting anomalies—with humans’ knowledge and creativity, results in rapidly evolving models that can give better insights, more accurate predictions, and act as a force multiplier for any organization,” Frost said. “Perhaps the very best aspect of this type of approach is that each person is contributing to the overall improvement of the model, so institutional knowledge grows exponentially.”
Two Paradigms
The startups are divided between two paradigms. One type builds industry-specific software applications, such as analyzing oil pumps, turbine performance, or human health.
The other category creates broader software platforms that glean the meaning of any type of data or automate data science that used to be done by hand, such as predicting maintenance problems with equipment at windmill farms using live data.
An example of the broader software technology is Maana, a company Frost created about 3 1/2 years ago. He’s still the chairman of this company, which has offices in Palo Alto and Bellevue, Wash. Maana, Swahili for “meaning,” created and patented the “Maana Knowledge Graph,” which uses data analytics in an attempt to increase the profitability of companies. It does that by taking data from “silos”—information that’s under the control of a single department and typically isolated from the rest of an organization—and organizing it into a “knowledge graph” that can show associations and relevant information regarding a specific business area or context, similar to an information box that would appear on the right side of a computer screen during a Google search.
Maana provides the same type of contextual information and relationships on assets, customers, processes, or any relevant item in an organization, assembling knowledge graphs from a variety of data sources. It summarizes data to help companies in various ways, such as finding operational inefficiencies, predicting machine failures, and recommending process changes.
Maana just raised $26 million in a Series B round led by Saudi Aramco Energy Ventures, the corporate venture arm of Saudi Aramco, which is the world’s largest integrated energy enterprise, Frost said. The round brings Maana’s total funding to date to more than $40 million.
GE
GE has invested in some Frost Data startups, including Maana. These days 11 of Frost Data’s startups are actively engaged with GE in pilot or “proof-of-concept” projects—that aim to prove value before any investment is made—in fields including aviation, healthcare, energy and transportation, according to Chief Strategy Officer Anthony Howcroft.
Frost Data also maintains a close relationship with GE’s Predix team, a cloud-based platform that connects people and machines with big data and analytics to bring the industrial internet to businesses.
It’s in the very early stages of working with another GE division called Current toward a goal of “disrupting” the traditional energy market and creating the next generation of energy usage, Frost said. That involves a much broader application of solar panels on homes, batteries in houses to store excess energy, and selling excess energy back to the grid or to neighbors. Current is a $1 billion energy division that applies GE’s “lab experiments” in the real world, according to news reports.
Frost Data will work with Current to implement the new paradigm in “smart” cities, Frost said. That entails switching out traditional lighting in all streetlights and public buildings with LED lights, which saves money in annual energy and maintenance costs, he said. The savings are used to attach various sensors, such as cameras and thermal imaging, to lights to obtain more data.
“The first generation was getting data directly from something,” Frost said. “In the next wave, we’re starting to understand the context around it, which gives us a much richer data set to get value from.”
Frost Data does not yet have any contracts with “smart” cities, Frost said. Examples of cities with that intention include San Diego, New York and New Orleans, as well as several cities in Turkey, he added.
Goals
Frost moved the company to San Juan Capistrano after a few years in Aliso Viejo because he wanted to be in an area with more historical context, he said.
His future vision is to expand it into a “campus,” with Frost Data startups spread out in commercial office space all over the city, he said.
“The idea is to extend … through San Juan Capistrano and build a culture … where we can pull everyone together with team-building exercises and create an atmosphere … like a mini-version of Palo Alto, where you have venture capitalists passing term sheets across every table at every Starbucks and restaurant,” Frost said.
Frost Data is in discussions with a major landlord in San Juan Capistrano for more office space.
