Katrina’s total system disruption

How many recovery plans anticipate what actual human beings might do?

I recently attended a CIO/CFO roundtable that, among other things, focused on the difficulty of demonstrating the business value of IT. Exasperated, one CIO said half-jokingly: “Want to see if IT matters? I’ll just go unplug all the servers and we’ll see how long the business runs.” He added that of course IT could never make that threat anymore, because it would be too disruptive -- unacceptable downtime is measured in seconds these days, not hours or weeks.

Yet total systems disruption -- water, power, and communication, to name only a few -- is exactly what happened in New Orleans last week, and the immediate aftermath offers a view of the value of IT as well as our vulnerability to disruption in a technology-mediated world.

Although few analyst firms tried to parse Katrina from an IT perspective last week, Gartner did set up a Katrina blog for their analysts, including one who volunteered his time to the rescue effort and others who looked at the disaster from a technology perspective.

One analyst lauded the power of social networking for quickly identifying missing persons and reuniting lost families. Another focused on the importance of backup power systems, noting that fewer than half of all mobile cell sites have backup generators and most have only a few hours of reserve power. (Working cell phones would have helped greatly last week). And others focused on issues that will become more critical in the wake of Katrina-driven fuel-cost increases, such as telecommuting and supply-chain optimization.

And then there was Gartner Research Vice President French Caldwell, who focused on better and more holistic uses of predictive technologies. “No one can be fast enough if relief starts once the storm is over,” Caldwell asserted. “Perhaps we need a paradigm shift that recognizes the tremendous predictive power that computing improvements have brought in recent years.”

Caldwell noted that although hurricane prediction models have continued to get better, what was needed in New Orleans was a more holistic, social-dynamics model that took into account human behavior as well as the weather. How many people and in what locations would be unable to evacuate? How many would refuse to evacuate because they were determined to stay with their house and belongings?

In a world where IT leaders are increasingly focused on human behavior along with organizational and cultural issues -- not just on hardware and code -- Caldwell’s observation seems right on the mark. Most financially sound enterprises have risk management and disaster recovery plans that incorporate data, assets, and critical business processes. But how many of those plans take into account what the actual human beings involved might do? How many take into account the “soft” issues that always seem to foul up the best-laid plans?