For a venture-backed technology startup like my company, Iodine, the clock starts ticking on day one--and every day after that is a race to grow fast enough to buy some more time.

The best way to achieve that fast growth is to build something people love so much they tell everyone they know about it--turning users into disciples. This sounds obvious, but it's a high bar. To breed disciples, your new thing can't be merely better than the status quo; it needs to be so great that it is sought out over the old, familiar thing. To do that, your new thing must be more than provocative; it must be dangerous.

Dangerous doesn't mean you need to build a risky product, as in risky for potential customers to use. Rather, it means being transgressive, upending the way people were living their lives previously. It means delivering an experience in a way that more cautious folk would consider verboten.

Many successful companies have thrived by crossing into dangerous territory. Facebook has time and again pushed the boundaries on privacy, in the interest of building more connections for its users (and more revenue opportunities for itself). And Uber, everybody's favorite example of a no-holds-barred company these days, has willfully flouted local laws and regulations in its haste to conquer new cities and countries.

At Iodine, we set our sights on doing something dangerous. This is partly because we're working in health care, an industry riven with inefficiencies. Up to one-third of health care spending is wasted, in overutilization, fraud, and errors. In a $3 trillion industry, that's a big incentive to ignore conventional wisdom and do something radical.

But doing something dangerous in medicine is fraught, because there are rules and regulators at every turn, like HIPAA and the FDA. The strictures are there for good reason: Lives are at stake, after all, and medicine is nothing if not complicated. It'd be plenty easy to rush into health care with the fine intention of fixing a broken system, but with no sense of how difficult it is to understand what you're trying to fix. We've seen other startups rush out promising cloud-based products for hospitals or insurers without realizing that health care companies are loath to let their data migrate onto a startup's cloud platform. There's a fine line between doing something dangerous and doing something stupid--and we're trying to stay away from stupid.

Nonetheless, Iodine is doing a couple of things that could make some in the industry uncomfortable. We're soliciting thousands of reviews from patients about their experiences with medications and treatments, bringing true consumerism to medicine. That's unsettling to those whose products or services might be judged. We're also letting users personalize their treatment regimens, making it easier for people to find what works best for them--stopping short of the regulatory ban on giving medical advice without a license. (There's no ban on providing medical information, even if it's highly personalized.)

The new mobile product we're building might be the most brassy thing we've done yet; I can't say too much, but it turns people's experience with a medication into an algorithm that helps them decide (with their doctor) if it's working. We're itching to see whether they find it transcendent, or too much.

Of course, the reason we're willing to take this risk is that there's little to lose. The clock is ticking, and we have to do something big enough to merit more investment. You could call our strategy embracing danger, cautiously. After all, doing something dangerous is easy. Doing something dangerous and smart is hard.