Five Ways To Get Rational About Risk

Seat of the pants is no way to prioritize security spending and set project precedence. But that's exactly how some CISOs are doing business.

It's no secret that some companies excel at information security while others run around stamping out fires and never get ahead. What's the secret of first-rate IT risk managers? To find out, we interviewed a half-dozen CISOs from various industries. We didn't talk about specific types of threats. Rather, we wanted to understand exactly how these successful security leaders prioritize money and manpower.

One thing we all have in common is data overload. Infosec leaders have their go-to sources for cyberintelligence, like vendors, newsgroups, the National Institute of Standards and Technology, and regulatory bodies. But at some point, we all find ourselves overwhelmed. Call it the "needle in a needle stack" conundrum: You know there's a ton of threats out there, many of them potentially damaging to your company, but if you pick the wrong needle at the wrong time, the stack may just fall and cause death by a thousand cuts.

And there are plenty of reasons a CISO might select the wrong needle.

We'll admit that the media doesn't always help. Early last year, my phone was ringing off the hook after the Google attacks, with information security pros asking whether APT--advanced persistent threat--was the most immediate danger to their companies. Now, data loss prevention and distributed denial of service are back in the spotlight, courtesy of WikiLeaks.

We're also all over the map with risk assessments. Every company we reviewed had some type of risk management framework, but the devil is in the details. We saw no uniform best practices. Our recent InformationWeek Analytics IT Risk Management Survey, available later this month, shows that the most popular way to measure risk, by far, is qualitative categorization of high, medium, or low. In our experience, some companies have rudimentary internal risk assessment systems, supplemented by an external vendor or third party, such as Gartner. On the other end of the spectrum are companies that deploy extensive, 50-plus-question surveys and use a stringent, quantitative approach where every response has a weight; the overall tally denotes a project's risk. Fewer than half of our survey respondents, all of whom play roles in assessing risk at their companies, use such a quantitative method.

What's interesting is that the CISOs we spoke with agree that neither a quantitative nor a qualitative approach is much help with prioritization. Quantitative risk analysis is not the be-all and end-all--just because a risk is scored at 98 out of 100 doesn't mean it will be remediated. For one thing, the business significantly influences whether to spend money. And most surprising to us, in the end, many CISOs say they ignore vendor input, media reports, pundit white papers, even all their own data and make gut decisions.

Let's be clear: Gut decisions aren't useful. Very often they're based on a confirmation bias--the tendency for people to favor information that confirms their preconceptions or hypotheses, regardless of whether the information is true. If you have a confirmation bias and think laptop theft is the largest concern, whether it is or not, you'll find a way to get encryption to be the highest-priority project.

Beyond Gut Instinct

Five Ways to Get Rational About Risk

Become an InformationWeek Analytics subscriber and get our full report on reducing security data overload. This report includes 13 pages of action-oriented analysis. What you'll find:

• Why you need an emergency risk assessment process
• Exclusive sneak peek at results of our risk management poll

Get This And All Our Reports

Michael A. Davis has been privileged to help shape and educate the globalcommunity on the evolution of IT security. His portfolio of clients includes international corporations such as AT&T, Sears, and Exelon as well as the U.S. Department of Defense. Davis's early embrace of ... View Full Bio

Newest First  |  Oldest First  |  Threaded View