If you had to protect a million dollars in gold, would you spend $3 million to do it? In certain circumstances, you might.
If you wanted to steal a million dollars in gold, would you spend $3 million to do it? Many would answer, "of course not."
But, again, in certain circumstances they would be wrong. I’ll explain in a moment, but first some background. What I’m talking about is a discipline called IT risk management. Risk management began in the financial markets where the risk was about stocks and securities. By extension, it was applied to an organization’s digital resources. That aspect falls under the purview of the Chief Information Security Officer (CISO) for your enterprise.
What’s that you say? You don’t have a CISO? Yes you do. Like all CxO titles, the position exists even without the title. Someone is, as part of whatever position they do hold, in charge of information security. So, therefore, they’re in charge of risk management.
The concept of risk management isn’t that difficult to grasp. It involves something called a "risk matrix," which is used to produce a number called a "risk metric," which, in turn, is used to choose a course of action. The risk metric is defined as the combination of the probability of occurrence of harm and the severity of that harm. Frequently the severity is monetized as the value either of the resource or the cost of mitigating the harm to the resource.
Let me give you an example. In my writings on authentication and authorization I’ve strongly advocated a method called Risk-Based Access Control (Risk-BAC). When properly implemented, the authentication system will construct a "risk matrix" based on the context of the authorization ceremony. That is, knowing who is logging in, what method they are using, where they’re located, which platform they are using, what the time is, etc. can all be distilled into an authorization risk metric. We then take into account the "what" -- what resource the person is attempting to use and calculate a value.
Based on a locally controlled formula combining the risk metric with the attribute value yields a calculated figure, which the authorization engine can use to determine one of four states:
State 1: The authentication is valid, the connection is safe, the access is allowed.
State 2: The authentication is questionable, further authentication factors should be asked for.
State 3: The authentication is valid, the connection is questionable, the authorization level should be reduced.
State 4: The authentication is questionable, the connection is questionable, the access is denied.
The risk management process can be distilled to four steps:
Step 1: Identify the context.
Step 2: Gather the data.
Step 3: Assess the risk.
Step 4: Treat the risk,
Now let’s look in particular at the last step: Treat the risk. Again there are four possibilities:
Risk avoidance: Remove the risk from the resource.
Risk reduction: Lessen the impact of the risk
Risk retention: Be aware of the risk and compensate for it.
Risk transfer: Let someone else foot the bill (e.g., insurance)
Going back to the million dollar resource we wish to protect, the fallacy here is to treat the value of the resource ($1 million) as the sole factor making up the severity of the harm. A bank, for example, touting the benefits of its strong room would lose far more than $1 million if the gold was stolen. The harm to their reputation could be valued in the billions. So spending $3 million to protect the gold is not really farfetched. Likewise, the thief might value the theft at far more than the gold’s value because it would enhance his reputation as a thief, so spending $3 million might be (for him) a good investment.
Consider all aspects of the risk when calculating a risk metric. As James Lam, former chief risk officer of GE Capital, once said: "Over the longer term, the only alternative to risk management is crisis management, and crisis management is much more embarrassing, expensive and time consuming."