Posted by: David Schneier
assess, assessment, assessments, Audit, bank, banking, CISO, CISSP, compliance, compliance officer, compliant, credit union, credit unions, CU, disaster, disaster recovery, DR, enterprise risk, enterprise risk management, ERM, exam, examination, examinations, examiner, examiners, exams, framework, governance, GRC, guidance, information security, information security office, infrastructure, ISO, oversight, policy, procedure, regulation, regulations, regulations audit, regulatory, regulatory guidance, risk assess, risk assessment, risk assessments, risk management, risk-based, risks, technology
About a decade ago a family member chastised me for having an auto repair shop do my oil changes for me. She (yeah, you’re reading that right – “she”) pointed out how ridiculously easy it was to drain the old oil, replace it with the new stuff and check a wide variety of fluid levels, connections and filters without having to pay someone else to do it. On one hand she had a valid point, it sure didn’t sound very difficult. On the other hand I immediately wondered how I would get to the plug where the oil needed to drain through in order to open it, where would I collect the old oil and how would I dispose of it once I did? And what the heck would I do if something went wrong? Plus I would need to remember to buy the new oil, perhaps a filter or two and then figure out how to check a myriad number of items to make sure the car was running right. Or I could keep going to my mechanic and pay him the $39 to take care of it for me. I’ve always had a way of considering things via the risk vs. reward formula and that was an easy one – have the professional do it. It would take me more than an hour not including shopping for the needed supplies and there was an increased risk that I would miss checking something, forget to tighten something or simply do a bad job. I’ve been earning more than $39 per hour for a long time and so I decided that I should just work an extra hour and use the proceeds to let the professionals do their job.
Which is why I don’t much care for any manner of compliance-based assessments that are self-administered.
Companies have had this crazy notion for more than a decade now that the best way to identify and address risks inherent within the infrastructure is to ask key stakeholders a somewhat generic set of questions and use their responses to figure out what’s what. Most of the time the people driving these initiatives are either information security professionals or corporate compliance people who either believe they already know where the problems are or are looking for the simplest and easiest way to satisfy some requirement. But what they often fail to grasp is that it’s almost impossible to draft a common set of questions that either apply to the vast majority or worse, will be interpreted consistently across the stakeholder population. Plus the perceived benefit of using a self-assessment approach to reduce effort and required support resources is almost always an illusion. Most of the time saved in not having someone ask the questions and record the answers is instead consumed by needing to explain the format, explain the questions or trying to clarify and clean up the responses. While supporting one such program recently each assessment required a kick-off meeting, a follow-up meeting to review the status of the assessment, a third meeting to review the initial draft of the questionnaire, a fourth meeting to review the resulting report(s) and a largely untracked number of hours to help generate all of the related support documentation. Regardless of the size of the entity being assessed each one consumed somewhere close to eight hours. While that might seem like a scary large number, the really scary part was that based on which risk analyst was responsible for the assessment and the personality/mindset of the stakeholder completing it the results looked very different from one another. It was almost impossible to generate meaningful metrics across the assessment population because a “Yes” answer for one question might mean the same as an “N/A” in another; there was no way to know that.
Another issue I’ve always had with the self-assessment approach is that while some stakeholders take it seriously and do a remarkably thorough job, others race through it with little hesitation just to fill in the blanks and get it off their desk. Sometimes you can detect which is which, sometimes you can’t. Plus the approach fails to capture much of the rich and relevant information related to each question and the underlying risk behind it. I recall conducting a team-driven risk assessment years ago where one stakeholder after the next covering a very broad sampling of the infrastructure kept lamenting on the lack of a proper disaster recovery plan. They had something to show auditors/examiners but to a person no one believed it was a truly viable plan. All but the CIO brought it up as a concern and when pressed a bit about why that was they all shared a common concern: If their main office was closed unexpectedly for twenty-four hours, regardless of the reason, they were likely out of business. A related self-assessment question would ask “Do you have a current and recently tested DR plan?” – most respondents on that engagement would simply have selected “Yes” and moved on to the next question without ever being challenged to share their concerns. Where’s the value in having a repository of questions and answers when it fails to capture the true essence or dimension of risk?
And the biggest issue I’ve always had with self-assessment questionnaires and their related templates is that they’re so often poorly designed. I can guarantee you that each of them has at least one question which makes zero sense to anyone who reads it. They either answer it based on what they think it’s asking, answer with an “N/A” or require follow-up with the people managing the process to have it explained. And you’d be amazed how many times even the author is challenged to provide a meaningful answer (including this guy). One thing’s for certain, a self-anything needs to be designed and written so that everyone understands what they need to do without having their hand held. Plus it’s rare that questionnaires are customized so that each stakeholder is only asked those questions that truly make sense. An application owner should never be asked if their anti-virus solution is current and up-to-date. A business process owner should never be asked about software change management. Yet seldom have I encountered a self-assessment process which does anything like this and so the audience is burdened with time consuming yet unnecessary questions.
Really though in the end my overriding problem with the self-assessment approach is that it fails to capture the expertise and guiding hand of true risk and assurance people. The process is often supported by analysts who don’t really have a feel for conducting assessments and are satisfied that all of the blanks are filled in. I have a nose for when there’s something beyond a simple answer and know when to scratch at the surface to bring it to light. By not allowing expert hands to guide the process potentially huge amounts of valuable and possibly critical details are being missed thus undermining any perceived value of the process. When you consider that all tolled and tallied the self-assessment approach versus the guided assessment approach doesn’t really save you much time (if any) and that it results in a weaker finished product, why would you elect to use it? One answer is that regulators push for it because perhaps it’s better than nothing (I can’t get any of those I know to comment). Another is that the people sponsoring these initiatives lack the fundamental comprehension to understand their options and chose what they perceive as the less complicated approach (again, I don’t know for sure it’s just a theory). What I do know is that when done right a risk assessment is managements best friend, a fundamental belief behind the recent spike in ERM activity.
While recently having my car serviced the mechanic discovered a nest of some sort in the engine block, he thinks it was probably squirrels. Because of this discovery he went searching for all the wired connections to make sure they weren’t chewed up and destroyed, quite a few were as it turns out (the car had been idle for several months). The bill only added the cost of the replacement wires but nothing significant for the time it took to first find which were affected and then replace them. Had I attempted the repair myself I might have noticed the nest and likely would’ve cleared it but know for certain I never would’ve thought to check the wires, where to look for them or what to look for. I was smart enough to rely on a professional with a nose for that sort of thing and it saved me time, money and best of all the aggravation of having the car break down somewhere unexpectedly. Good thing I didn’t go the self-repair route.