Failing an audit sounds like the last thing any company wants to happen. But that's because audits are seen by many as the goal of a security program. In reality, audits are only the means of testing whether enforcement of security matches the policies. In the broader context, though, an audit is a means to avoid a breach by learning the lesson in a "friendly" exercise rather than in the real world. If the audit is a stress-test of your environment that helps you find the weaknesses before a real attack, you should be failing audit every now and then. After all, if you're not failing any audits there are two possible explanations:
1) You have perfect security.
2) You're not trying hard enough.
I've never met a security person who will claim they have perfect security. Nemertes research further illustrates this issue. In our most recent benchmark we found that in the past three years 36% of companies had suffered a breach and yet only 15% had failed an audit. I can't emphasize this enough: Those numbers are *backward*.
Companies should be failing audits, whether internal or external, far more often than they suffer breaches. The fact that few companies are failing any audits should be cause for concern, not celebration. I would celebrate if there were no companies suffering from actual security breaches because then we could assume that the audits were working: uncovering problems to fix them before they became breaches. But unfortunately, it seems that audits are not thorough enough, consistent enough or "hard" enough.
If you accept that the purpose of internal or external audits is not just to "prove" security but to "improve" security, then the audit should subject the company to enough pressure to validate that it can withstand a security breach. Otherwise, what's the point? Proving that the company can withstand the auditors?
If you don't fail audits, you're not trying hard enough. Statistically speaking, on average a company will have some form of security event at least once a year and suffer a serious breach once every three years. From a statistical perspective, it should also follow that companies should fail at least an internal audit every three years or more often.
Arguably, if you pass the audit and fail the breach you're doing worse than a company that failed an audit and suffered a breach. Clearly the second company's security sucks and it has a lot of work to do -- but at least it knows it. The first company is worse off: Its security sucks but it doesn't know it because its audit sucked too. Or worse, the company does know it and it's fudging the audit, as if fooling an auditor is a worthwhile result.
So, go ahead, fail an audit. If you're not failing any, ask yourself: Is my security really that good, or are we just going easy on ourselves? Are we confusing the means for the goal, succeeding in audits only to eventually fail in real life?