STRATEGY · HBS tackles informatics' basic question: Where's the money?

By John Russell

March 8, 2005 | To a hammer, almost everything else looks like a nail. That perspective was apparent during a panel discussion* about business models for bioinformatics, held during Harvard Business School's 10th annual Cyberposium. Most panelists were bruised traveling bioinformatics' rutted road, and their prescriptions for success resembled the businesses they currently run. What a shock.

Nailing down a single formula for successfully selling bioinformatics products or services has proven elusive. Failures outnumber successes, and many of the companies still in the game can't dump the "bioinformatics label" fast enough.

"We like to call ourselves a biotechnology company," said Keith Elliston, CEO of Genstruct, whose predictive modeling platform is helping fellow panelist Enoch Huang, director of molecular informatics at Pfizer Research Technology Center, identify mechanisms of action for compounds.

Peter Barrett, senior partner at Atlas Venture and former chief business officer at Celera Genomics, agreed. He noted a client company that "some people would look at as a bioinformatics company. Frankly, we call it a cancer diagnostics company."

Not lost on the panel was that pharma picks up the bioinformatics tab. Huang wryly noted: "We subscribe to the Ingenuity Pathways Analysis platforms. We were subscribers to Celera Discovery systems. We subscribe to Bio·IT World magazine. We still have a collaboration with Gene Logic. In many respects, we are the industry and represent the way the industry thinks about bioinformatics. I feel like I have a big bull's-eye on my chest right now."

Ramon Felciano, CTO of Ingenuity Systems, said the subscription business model is working well for his six-year-old company, although it took about four years before the product was ready. "I'm not sure I'd have sat on this panel a couple of years ago," he said.

"We are now starting to see more companies become interested in enterprise-type solutions because some companies don't want their scientists sending their data over the Web to our servers, so they want to grab a snapshot of that entire thing and put it inside their firewall. We just did a deployment at Merck at the end of last year," Felciano said.

Two problems often cited with subscription models are the low price points commanded by software and the small size of the life science market. "It's very hard to get good value for what you've created from the company because it's only 'software,'" Barrett said.

A few companies, notably Spotfire (visualization tools), Phase Forward (clinical trials software), and VizX Labs (GeneSifter), are succeeding, but, Barrett warned, "You have to pick a big enough niche, and you really have to understand the customer."

Elliston took pains to divide the bioinformatics world in two: IT-centric systems best suited for organizing and presenting science data versus heavy-duty computational applications that create results. There's a need for both, he argued, but the market dynamics are different.

Barrett pressed Pfizer's Huang on the point. "[Attrition] is the thing that costs you hundreds of millions of dollars, yet when it comes to the poor companies that are providing maybe the value there, you don't want to give them royalties. You want to give them a subscription-based model, and that's where I think the innovation stops, because unless it's a fair balance of innovation to value, no one's going to invest in that in the field."

This is hardly a new complaint. However, many bioinformatics technologies have not delivered as promised, and the proliferation of unproven tools has presented pharma with a smorgasbord that is difficult to digest.