Jim Fruchterman, founder of Palo Alto-based nonprofit Benetech, was very relieved when, at the end of September, he received an email from the U.S. Department of Education.
With one day left to go in the federal fiscal year, the department had renewed funding for a longstanding reading program for the disabled, and it had chosen Benetech for the third time to run the five-year, $42.5 million project. That meant Benetech could continue to operate BookShare, which provides digitally materials to 500,000 students with reading disabilities, including blindness and dyslexia, and Fruchterman could forget about planned October layoffs at Benetech.
No less important for Fruchterman, a buoyant, 58-year-old Caltech graduate with a more or less constant grin, his frequent trips to Capitol Hill are over for now.
“It’s been a huge distraction to have to keep going back to D.C. and say, ‘Hey, we’re the largest program that serves blind and dyslexic kids. We’re 10 times more cost-effective than what you were doing before. Isn’t that great? Both Democrats and Republicans like kids with disabilities, right?” Fruchterman rolls his eyes and laughs at his own frustration.
In Silicon Valley’s hyper-competitive startup scene, Fruchterman is a highly unusual figure, and not just because he spends a lot of time in D.C. or likes to work on social good projects. Back when no one had heard of social entrepreneurship — the idea that the energies and skills of entrepreneurs can be used for social good instead of investors’ profits — Fruchterman was one of the very early pioneers. And unlike wealthy techies in the philanthropic ranks, like Bill Gates or Pierre Omidyar, Fruchterman didn’t wait to do good until he made billions. In fact, he has never made a fortune, and he intentionally dropped out of the Silicon Valley race-to-riches.
Fruchterman instead took a page out of the Silicon Valley playbook to address persistent social challenges, like helping the disabled to read. His approach is to use money from philanthropists in the same way founders use venture seed rounds to get concepts off the ground, and then, if they work out, raise more funding to drive to a self-sustaining, nonprofit revenue model.
BookShare followed this model to the T, starting with foundation grants to deliver reading tools — by instantaneous, digital means — for the blind and dyslexic who formerly depended on physical products like Braille books and the U.S. mail. Today, the U.S. Department of Education underwrites the service for students, and adults can also subscribe on their own for $75 a year.
Yet Benetech is about much more than BookShare alone. Fruchterman created the organization to house a number of ongoing projects, including around human rights, the environment and other fields. In fact, Benetech houses an incubator for new endeavors, like a current initiative to make social services as easy to discover as the closest pizza joint, as well as a consulting service for nonprofits in need of technology help.
“No one else does what Jim does with such focus and dedication,” says Sally Osberg, president and CEO of the Skoll Foundation, which has granted nearly $2.6 million to Benetech in the past decade, “His ability to bridge the commercial world of tech innovation and the potential for social impact is his real strength in the field of social entrepreneurship.”
The model works, too. Eighty percent of the $13.4 million annual revenue the 70-employee outfit enjoyed last year came from operating projects like BookShare. The balance came from donors, most of them in Silicon Valley. And in total, Benetech projects have taken in $9 million in nonprofit risk capital and have since attracted nearly $107 million in either follow-on grants or revenue.
I started from single-enterprise entrepreneur, to portfolio-of-enterprises ringleader, to guy who wants to help the entire Silicon Valley software and data ecosystem transform the world of disadvantaged communities and the social sector that serves them.
Fruchterman has accumulated lots of laurels for his work, including a MacArthur Fellowship and the Skoll Award for Social Entrepreneurship, but in a series of interviews with Fruchterman over the past year those accolades never come up. The conversation inclines to what he’s learning as he “plays CTO for an hour” to other nonprofits, which provides endless fuel for his key insight: the poor or non-existent use of technology in much of the nonprofit sector.
In a recent piece, Fruchterman called for a “software revolution for the greater social good” to address the fact that “the social sector has not fully benefitted from what Silicon Valley does best: Using software to achieve scale and using the data inherent in software solutions to continuously improve services, identify new opportunities and demonstrate impact.”
That’s the Silicon Valley thinker in Fruchterman: The market opening he’s after today isn’t simply helping the blind read, it’s marrying the good intentions to hard technology. “This is part of my career evolution,” says Fruchterman. “I started from single-enterprise entrepreneur, to portfolio-of-enterprises ringleader, to guy who wants to help the entire Silicon Valley software and data ecosystem transform the world of disadvantaged communities and the social sector that serves them.”
hirty-six years ago, the young Caltech graduate was no less ambitious, though his main interest then was rockets, or maybe winning a Nobel. It was the early 1980s, in a universe far, far away, when the Reagan administration made it legal for private companies to develop rockets and defense spending boomed thanks to “Star Wars” missile defense programs.
Fruchterman was enrolled at the time in the PhD electrical engineering program at Stanford. He and his friends liked to host entrepreneurs for dinner, and one evening Gary C. Hudson, founder of rocketry startup GCH, Inc., joined them at the Stern dormitory. Hudson was looking for engineers, and he had a simple screening test. He asked Fruchterman two questions:
“Who is your favorite science-fiction author?”
“Poul Anderson,” Fruchterman replied, citing the author of A Knight of Ghosts and Shadows, among many others.
Hudson approved and asked, “Who was your favorite Poul Anderson character?
Fruchterman replied that it was Ensign Flandry, the starship pilot-turned-James Bond of the late Terran Empire, in Anderson’s Technic History series.
Hudson shook his head. The right answer, in his view, was Nicholas van Rijn, the libertarian-minded adventurer, another figure in the same series.
Hudson gave Fruchterman a pass on the second question and offered him a job on the spot. Fruchterman couldn’t resist. What sci-fi loving engineer could? In 1981, he took a leave of absence from Stanford and went to work in Sunnyvale for GCH, which had backing from Space Services of Houston, and a real estate investor named David Hannah. Soon after, on August 5, 1981, GCH tried to test-fire its innovative rocket-engine design, Percheron, on a launch pad in Matagorda, Texas. The countdown reached ignition, but the engine blew apart on the launchpad.
Fruchterman picked up part of a rocket fin — a better startup memento than most — then he and a couple of his colleagues headed back to Silicon Valley to try to raise $200 million for a new rocket company. There were no takers, but Fruchterman decided nevertheless that he would not be returning to the PhD program.
“Gosh,” says Fruchterman, sitting in his Benetech office on leafy California Street, roughly a mile from Stanford’s campus. “I realized I found what I wanted to do; I liked to build stuff a lot more than do research.
That “gosh” thing happens strikingly often in conversation with Fruchterman, and it’s as disarmingly authentic as his Archie-comics good looks and sunny disposition. It’s hard to square that sense of Fruchterman with the stubborn, principled engineer who would not let go of a technology product idea that would eventually better the lives of millions of blind people. And this was long before BookShare and after the rocket disaster.
That part of Fruchterman’s story begins at Sirloin & Brew, the now-defunct Silicon Valley hangout on El Camino in Mountain View. Fruchterman and one of his colleagues from GCH, Dave Ross, met with Eric Hannah, a Hewlett-Packard chip designer, who had a new idea for a chip that could work with a scanner and software to recognize the characters on a printed page.
“I got really excited by the idea,” says Fruchterman, “because I had this idea in college to make reading machines for blind people using pattern recognition, and here was Hannah saying he had a way of doing that.”
Fruchterman joined Ross and Hannah to found what would become Calera Recognition Systems, where they launched one of the first OCR technologies that used machine learning to “read” almost any script, not just pre-programmed ones.
“What we did was gather several million examples of characters, compared them all, clustered them, then built an algorithm that would say, ‘This looks like an A,’ even though it had never seen that font before.”
Through his first seven years at Calera, Fruchterman kept thinking about how the OCR tech could help blind people. In 1987, after the OCR tech was working well, Fruchterman and Ross, who was the VP of engineering, decided to set up a skunk works to figure out if they could get the OCR technology to work with Votrax, an early voice synthesizer technology on a PC.
“It sounded really crappy,” says Fruchterman, “but we could scan a page and the voice synthesizer could read it back.” He and Ross finally had a prototype that could use a PC and scanner to read text aloud to a blind person.
Fruchterman took the demo to the Calera board, fully expecting an immediate green light to go to market. Instead, the board asked how big the market was. Fruchterman responded that it was probably a $1 million market, which in fact understated the market by 5x, though Fruchterman figured market size was not the issue – helping the blind read was the issue. The board didn’t see things that way. Calera had taken on $25 million from investors, and they expected Calera to make a huge return. Small markets were not the road to 20x returns.
“We got into an argument,” recalls Fruchterman. “I’m saying, ‘But blind people need it.’ They’re saying, ‘Yeah, but we create jobs and we’re investing money for pension funds, and we have to make money to fill our social goals. To work on a million-dollar-a-year product would divert us from making all the money that we promised to make in exchange for our investment,’ which is a pretty fair statement. I could not really disagree, but I also kind of went, ‘Wow. So good technology won’t get built.’ That really irritated me.”
Fruchterman quit. He had been with Calera seven years, held several jobs, from CFO to marketing VP, and was ready for something new. He was also determined to build the reading machine, so he made a proposal to the board that he hoped would make the reading machine possible without losing his shirt.
In exchange for certain non-competes, Calera agreed to license him the OCR technology at barely above cost so that Fruchterman could create two companies: one a for-profit to use the OCR for postal applications (which was a market Calera had chosen not to pursue) and second a nonprofit to build the reading machine for the blind. Fruchterman created RAF and a 501c3 called Arkenstone, respectively, and became CEO of both.
“I told my wife that I would only do the 501c3 for a year or so, and then go full time back to regular tech. But I ended up doing both for six years before I realized I had to choose one way or the other.”
Arkenstone performed better than Fruchterman hoped. He shipped the Calera OCR technology, which was a PC co-processor card listing for $6,000, to blind customers for less than half that price and threw in DECtalk voice synthesizer. The market Fruchterman anticipated quickly took off. After three years, Arkenstone was very slightly profitable on $5 million in revenue and 501c3 status, a strange beast in the eyes of the IRS, which associated 501c3 status with hospitals and philanthropies in those days. That prompted an IRS audit, which Arkenstone cleared without issue.
It was pretty clear to me that I liked doing tech for good more than I liked doing regular for-profit.
After several years running both RAF and Arkenstone, Fruchterman realized that he was not headed back to a regular tech job as he had promised his wife, and he wanted to focus entirely on Arkenstone. The financial opening to do so came in classic Silicon Valley fashion: In 1995 his former employer, Caleram was acquired by its competitor, the legendary Robert Noyce’s Caere, and Fruchterman’s shares were in the money.
The modest (by today’s standards) $600,000 windfall was enough to expand his two-bedroom, one-bath ranch house on Middlefield Road in Palo Alto and help send his three children to college.
“My wife and I said, ‘Gee, do we want to try to make millions more?’ I realized that I might make more many years out from RAF, but I really felt the tech for good was me, as opposed to buying into the Silicon Valley serial entrepreneur rat race.” He resigned from the for-profit RAF, which is still an operating company, to focus on Arkenstone.
“My friends were shocked,” he says, “but it was pretty clear to me that I liked doing tech for good more than I liked doing regular for-profit. They’re both challenging, and frankly, I saw a fallow field. Everyone else is crowding around the for-profit field, always trying to convince VCs to invest more money, while over here, there’s no one in tech for good, and you can have a big impact. What really changes are the financial set points. I went from, how do you get to a highly profitable venture, to how do we get to a break-even venture that does good. [The answer was to] give up on making money, just do good tech and find a way to break even.”
“Why would anyone do that?” is, of course, the obvious question. Doesn’t everyone in Silicon Valley dream of getting rich?
ruchterman comes at the question from two directions: family and school. He grew up outside Chicago in a Catholic family with six kids, and his father was a lawyer at International Harvester, so the idea of service was part of his family’s ethos. Technology was a big topic in family conversation, but “To us, entrepreneurship meant owning the dry cleaner or a Chinese restaurant, and we sort of thought, ‘Why would you do that?’ On the other hand, there was always this expectation in our family that you would do good. That was real.”
As a high schooler, Fruchterman says he was also deeply impressed by the example at St. Viator, his Catholic all-boys high school, where the teachers were “living a life of service and giving back. They loved computers and math as much as I did. They were really good role models who could have had more lucrative careers, but they chose education instead.”
The head of the math department turned him loose on calculus sophomore year and steered him to Caltech, amusing the young Fruchterman by reciting the Caltech “cheer,” which was also on the office wall:
“Cosine, Tangent, Hyperbolic Sine,
Three point one four one five nine
E to the X DY DX
Tech, Tech, Tech!”
“I am a tech geek first,” Fruchterman likes to say, “not a social activist,” which is about as prideful as the Caltech grad gets, but there’s no question that social problems own his imagination and purpose in life.
Even after he was working full time at Arkenstone on technology products for the blind, he was restlessly looking for new issues to address. Human rights violations came up high on the list, especially after revelations in the early 1990s about the cover-up of the 1981 El Mozote massacre in El Salvador, in which Salvadoran soldiers killed at least 800 civilians.
He wondered how technology could have prevented such a travesty, and he took long walks with a friend on the Stanford dish trail — where many Silicon Valley ideas take shape — to brainstorm highly improbable high-tech solutions, including force fields to protect vulnerable communities. He concluded that the real opportunity was to deliver timely, actionable information about abuses. El Mozote’s perpetrators were disguised as much by the passage of time as anything else.
Fruchterman had ideas about how technology could help, and he took them to groups like Amnesty International and Human Rights Watch. He was surprised to discover “that their technical levels were really low, and they wanted help but had no money.” Fruchterman wanted to help, but Arkenstone did not have the money either.
In early 2000, Fruchterman had a surprise break that, like the Caere acquisition of Calera, would create an opportunity to tackle more projects. Out of the blue, a businessman named Dick Chandler offered to buy Arkenstone for $5 million. Chandler also promised to continue producing reading machines. Fruchterman took the offer and used the funds to stake a new 501c3 called Benetech, which borrowed everything but the profit motive from Silicon Valley.
“Essentially we took the standard venture capital model and tweaked every one of the investment criteria for social good. What Benetech does is commands the same set of talents and dynamics. It’s just the financial set points are different. We were looking to create break-even ventures that do good.”
ruchterman took advantage of the 1999 dot-com bust to lease the outfit’s current office at a great rent (including cubicles!) and set out on two tracks: one to raise so-called risk capital — in effect restricted and unrestricted grants from donors — because $5 million would not last long, and the second to get projects under way.
One project was the next stage in his work to help the blind read by digitizing virtually all reading materials, from digital audio books (formerly on tape) to digital Braille files that drove an electro-mechanical Braille reader (formerly Braille books), to iPad-based, listen-and-follow systems for dyslexic readers.
The Skoll Foundation, the Omidyar Network and the Social Profit Network kicked in nearly $4 million to develop the technologies, and Benetech won its first Department of Education contract for BookShare in 2007.
One BookShare beneficiary is Brian Meersma, a senior at Cornell who is studying industrial and labor relations and who has dyslexia. Meersma has been a BookShare client since elementary school, where his disability was so severe that basic comprehension was a challenge.
“To get the words, I had to read so slowly that I had no idea about the meaning,” he says. Today, Meersma gets his text books, as well as pleasure reading, via BookShare, which has a mode for iPads that combines sped-up audio with text that highlights in sync with the audio. Meersma frequently sends BookShare team ideas to improve the service and requests new books be added to the library.
“In a lot of ways,” says Meersma. “Bookshare really gave me the opportunity to go to college. The ability to tap into a library that I could read on my own really opened up possibilities.”
Human rights was another early ambition at Benetech, and Fruchterman recruited Patrick Ball, a leading expert on human rights data, in 2002 to bring his Human Rights Data Analysis Group (HRDAG) under Benetech’s umbrella, where they worked for 12 years on data problems — like estimating civilian death tolls in inaccessible war zones — and developed Martus, a self-encrypting and self-replicating information collection and management tool to help human rights investigators keep their data safe from prying eyes.
Martus represented the kind of tool that Fruchterman thought might have helped human rights investigators break the El Mozote story sooner and perhaps bring the Salvadoran civil war to a faster conclusion. Hundreds of human rights groups around the world tried Martus, and a few dozen put the tool into regular use, notably in places like Guatemala, Burma, and countries in central Africa.
“Martus has never had a security-related bug, which is a credit to Benetech’s engineers,” says Ball. At the same time, he adds, Martus was not as widely adopted as he initially hoped. “What we discovered is that it’s hard to get people to use tools they are unfamiliar with,” says Ball, “and that new tools really raise anxiety. Many would just prefer to use familiar tools like YouTube, WhatsApp, or just paper.”
In addition, over the past decade, mobile platforms also became more secure by default just as the nature of threats changed significantly. The original design for Martus worked well for “smash-and-grab” scenarios, like secret police grabbing a computer or hacking a server, but today spear phishing attacks that can drop malware on a device to read data in machine memory are threats outside of Martus’ scope.
The Martus experience serves as a sort of cautionary tale in Ball’s eyes. Even the best tech and best intentions sometimes won’t deliver the intended results. Ball takes that observation a step further, in a challenge to his evangelical technologist friend and former colleague Jim Fruchterman, to argue that some of the technical innovation coming at the social sector may actually be dangerous to some communities.
For example, data and machine learning about marginal groups, like the homeless or residents of poorer neighborhoods, may end up in the hands of police or informing “predictive crime” algorithms. “Taking risks to get from here to there is what Silicon Valley loves,” says Ball, “but (in many projects) the risks are not being born by people doing the disrupting, they are born by the communities being disrupted. That is what worries me.”
Fruchterman describes the Benetech’s projects over the years much the same way a VC would their portfolio: any number of failures, a few singles and doubles, and a very small number of home runs, BookShare being more like a grand slam.
Fruchterman travels a lot these days hoping to find donors ready to stake new projects and learning from other social entrepreneurs and tech-for-good initiatives. His dream is to raise more risk capital so that his Benetech team can develop more prototype solutions that might lead to the next BookShare.
Is it possible, for example, to create a smart search interface for social services? Today, people in need of services can’t easily discover what’s available to them. Google has already mastered that for day-to-day, consumer services, and Fruchterman is working with players in the field to try to unify their work and datasets. Or perhaps homelessness can be addressed if only there were better ways to track client populations, and the causes of their homelessness, and mark progress. That’s another one high on Fruchterman’s list.
If Fruchterman knows one thing, it’s that high-impact projects take lots of time and money to bring about lasting meaningful change. Bookshare was the result of 20 years of work for Fruchterman when it launched in 2007 with the Department of Education. His former colleague, Ball, notes that Benetech’s success has as much to do with time and trial as it does with tech.
“These problems don’t get solved at overnight hackathons,” says Ball. “Benetech’s great strength has been its persistence, which allowed for engagement with a lot of people, and lots of iteration.”
Those long cycles suggest slow progress, unless more Benetechs take the field. The good news is that Benetech is no longer alone. “For the first 10 years,” says Fruchterman, “we thought we were the only ones. Now other teams are showing up, and the quality of what I see coming out of Stanford, Berkeley and Santa Clara has gone way up.”
Skoll’s Sally Osberg sees the changing playing field as well.
“Incredible potential can be unleashed by someone like Jim,” she says, “and the rising generation sees this potential. We see incredible talent that’s just as drawn to the nonprofits as they are to the commercial sectors.”
Fruchterman’s example, more than his next single or home run, may be the most important platform he leaves the tech world.