Researchers Finding Rewarding Careers As Software Entrepreneurs

Entrepreneurs When Steve Sands was studying for his Ph.D. in neuroscience in the late 1970s, he planned to write his doctoral thesis on experiments matching brain function with behavior. However, a lack of good software or computerized methods for correlating brain activity with behavior eventually forced Sands to drop his idea. "I ended up doing a thesis on animal behavior [alone] rather than the neuroscience project I really wanted to do," he recalls. Like many scientists who are frustrated by the lack of available software tools to aid in specific experiments, Sands took up programming while he was working on the psychology department faculty at the University of Texas, El Paso. And like a number of these computer-savvy researchers, Sands eventually was able to parlay this experience into a new career: He wrote a pair of software programs tailored to neuroscience research that he took to market. Today, he is chairman of Neurosoft Inc., the El Paso-based company he formed in 1985 to commercialize his "electric source imaging" software programs, SCAN and STIM. These products enable researchers to map the origins of the brain's electric signals to behavior and other stimuli. His programs make it possible to merge signals from electroencephalogram and magnetic resonance imaging into images that pinpoint sites of brain activity. In 1989, a subsidiary, NEUROSCAN Inc., was set up to market these and other Neurosoft products. "Six hundred labs around the world use my software to do research," says Sands. Neurosoft sponsors three meetings on neuroimaging annually, and Sands and other scientists working at the company publish the results of their own experimental work in neurophysiology in peer-reviewed journals. As a business, Sands's company has done well: "We've been profitable for every year we've been in business, although not by much in some years," he maintains. He estimates that he has earned more as a software developer than he would have if he had remained an academic scientist. But turning scientific software into a commercial success isn't always easy, according to scientists who have tried to become entrepreneurs. Finding financial backers for novel programs is nearly impossible. And the mechanics of running a business are not part of graduate science training-at least, not yet. 'EYE-OPENER': Entrepeneur Edmund Glaser found that he had a lot to learn. "The development of a software or technology company really requires a lot of skills that a university scientist doesn't have," observes Edmund Glaser, a professor of physiology at the University of Maryland School of Medicine in Baltimore. Along with his son, Jacob, Glaser established MicroBright Fields Inc., based in Colchester, Vt., to commercialize Neurolucida, software that provides quantitative measures of brain cells and tissue. They founded their company in 1987. "As soon as we became involved in commercialization, we started to see all that we didn't know," Edmund Glaser recalls. For instance, the business plan bank officers want to examine before they will talk about lending money is quite different from the scientific prospectuses he wrote for grant proposals. "Financiers are so often looking for immediate winners," he says. "I took a short course [through the university's business school], and it was an eye-opener to find out what it takes to entice investors, other than your relatives." Sands attributes much of Neurosoft's business success to his alliance with Ron Wright, a businessman and consultant he met while on sabbatical at the University of Texas Medical School in Houston in 1987. Wright joined Neurosoft that same year to provide the business expertise Sands lacked, including experience in developing export markets for technology. "We made flights to companies all around the world, whether we could afford it or not," says Sands. Today, about 70 percent of Neurosoft's sales are outside the United States, especially in Germany. The company tries to reach potential customers through advertisements in scientific publications and mailings to a company-developed database of 17,000 names. "What really helps is when prominent researchers use the software," says Sands, noting it generates a lot of favorable word-of-mouth "advertising." It also has been important-if a little unnatural at first-for Sands to promote his software in person. "You need to be willing to go to trade shows and sell your product, as well as be able to withstand criticism of it," he says. People can be "pretty blunt," he reports. Sands has learned to stay calm and try to get the critics to be specific: "I try to deduce if there is something [about the software] we should really fix." "It's very different in academics," he comments. "One spends a lot of time preparing a paper, thinking of contingencies and how to respond to this criticism or that one. In the business world, you have to stand on your feet and defend your software product." CUSTOM-MADE: Daniel Myer developed Custom Dialog database-searching software. Daniel Meyer ended up in the software business unexpectedly. Originally trained as a zoologist, he says he "fell into" a software career while working in a toxicology lab in Maryland. "I got involved in searching databases, looking for references to chemical interactions for the compounds we were testing" under contracts with the U.S. Environmental Protection Agency. Eventually, Meyer specialized in data handling, which he preferred over lab work. After a five-year stint in the chemical information division of the Philadelphia-based Institute for Scientific Information, he formed Advanced Research Technologies (ART) with headquarters in Rosemont, Pa. Meyer first helped client companies implement systems for chemical database searching. Later, he developed software-Dialog Quickstarts and Custom Dialog-designed to search databases such as Dialog and DataStar from Knight-Ridder Information Inc. When ART was ready to expand the market for its database-searching software, Meyer contacted Mountain View, Calif.-based Knight-Ridder Information to market ART's products. Knight-Ridder has sales offices located around the world, which would greatly enlarge ART's reach. And ART's database searching software was a natural fit with Knight-Ridder databases, so it was relatively easy for the two companies to work together. In January of this year, Meyer's company merged with an early ART customer, Telebase Systems of Wayne, Pa. Telebase provides access through CompuServe to several newspapers, chemical research databases, and company profiles. The merger allowed both firms to expand their information- products offerings as well as establish an organization with the expertise to develop and market database access software for the Windows platform and the World Wide Web. The new company retains the Telebase name, and Meyer serves as general manager. Meyer says scientists considering the software business need to understand the scope and interests of the product's customers. "You've got to get out there and know [if] there is a market for your product, that there are people who will pay for it," he says. Meyer adds that the serious entrepreneur must find software buyers willing to pay license fees at a level that will support a business, not just a sideline hobby. 'JUDGEMENT CALL': Steve Fritz decided to license his radiology imaging software. Potential financiers will also demand marketing data, says Glaser. "We're always asked, 'What is the market?'" But it was just beyond us to do a sophisticated marketing survey," he says. In most cases, scientists develop "niche" software with limited markets that are well-known to the creator, but hard to document in the way venture capitalists like. None of the scientists with emerging software companies interviewed for this article was able to obtain significant outside capital. Instead, they were forced to be "self-financing." Sands says he used money from the sale of his house to buy his first computer, and also took out personal loans during the early years. Other scientist-entrepreneurs attempt to keep their "day job"-often an academic position-as long as possible. Another source of development funds that Glaser and others have tapped are Small Business Innovation Research (SBIR) grants that are offered by federal agencies such as the National Institutes of Health, the National Science Foundation, the Department of Energy, and others involved in funding research. These grants support R&D efforts to advance promising technologies so that they might attract real interest in the marketplace. An additional obstacle to making a software company successful is trying to maintain two careers, one as a scientist and the other as an entrepreneur, while the company is in its early stages. Sands recalls finally deciding to leave his university faculty position when the dual life of professor and entrepreneur became too much. "I realized that either I do both jobs [professor and software developer] poorly, or one thing [well]," he says. Steve Fritz, a professor of radiology in the University of Maryland School of Medicine, says it's hard to decide when it's "safe" to trade the academic life for the business world. "Very few people do that. I think it's a judgment call. You have to be comfortable with insecurity and uncertainty," he reports. Fritz decided against making the leap into business after he developed some software that controlled specialized image-processing hardware. Instead, he licensed his software to a private company. DeJarnette Research in Towson, Md., has since incorporated features of Fritz's radiology imaging software into the controller program for the computer hardware it markets. Academic scientists who are considering a move to the entrepreneur's hot seat should have an "exit strategy" or alternative career plan in mind, says Fritz: "If you leave your academic field to develop and market a new software product, and it fails in five years, you are likely to be so far behind the current state of knowledge in your scientific specialty that you can't return to a research faculty position." Legal Steps Provide Protection For Programs ADVICE: John Hornick urges potential software entrepeneurs to pay attention to legal details. Creating a successful scientific software company requires attention to legal details that are easily overlooked. John Hornick, an intellectual-property specialist and partner with the law firm of Finnegan, Henderson, Farabow, Garrett and Dunner in Washington, D.C., points out several ways to protect programs. Copyright and Patent Protection: Software "source code," the lines of instructions that make the program work, can be protected with a copyright. Patent protection is not available for software, although it might be for the methodology the software makes possible. A computerized inventory-control system might be patentable, while the math underlying the software is not. Ownership: In most cases, a scientist's employer will own any software developed on the job. For academic scientists, the picture is less clear. Patentable work done with federal funding belongs to the university; copyrighted material may not, depending on the institution. Steve Sands, chairman of El Paso, Texas-based Neurosoft Inc. and a developer of software used in neuroscience research, took great pains to maintain ownership of his software when he accepted a faculty position at the University of Texas, El Paso. Sands made it clear that the work had been started before he joined the university, and that he would continue the work separately from his faculty duties. "I made it very explicit and got the university to sign a document acknowledging the arrangement," he says. Confidentiality: Hornick recommends getting a confidentiality agreement signed from any potential source of financial or marketing assistance to protect against the possibility that the idea will be "stolen" and developed without due compensation. Licensing: The seller and buyer need an agreement outlining the cost of the software and their relationship after the sale. In some instances, the seller simply provides the software and the source code, leaving it to the buyer to solve problems or make improvements. In other cases, there may be ongoing contact, with the seller providing the hardware that runs the software as well as updating or improving the software as needed or requested. If hardware is involved, a maintenance agreement may also be needed, says Hornick. Escrow Agreement: Since software companies come and go, a buyer may want the seller to put the source code in escrow. If the seller goes under, the buyer gains access to the source code and assumes responsibility for all further updating and development. The costs of these agreements vary widely, says Hornick. "If I can modify existing agreements, it will cost our clients much less than if I have to create them from scratch." He recalls one client who spent $100,000 for a range of agreement templates the company planned to use for some time, while another client used some existing agreements with a few modifications at a cost of about $10,000. -L.K. Lee Katterman, a writer based in Ann Arbor, Mich., is editor of Research News, a publication of the University of Michigan.

Researchers Finding Rewarding Careers As Software Entrepreneurs

Interested in reading more?

Become a Member of