Carving-up the policy pie
Today being Pi day 2019, think of the organization's suite of policies as a delicious pie with numerous ingredients, maybe a crunchy crust and toppings. Whether it's an award winning blue cheese and steak pie from my local baker, or a pecan pie with whipped cream and honey, the issue I'm circling around is how to slice up the pie. Are we going for symmetric segments, chords or layers? OK, enough of the pi-puns already, today I'm heading off at a tangent, prompted by an ongoing discussion around policies on the ISO27k Forum - specifically a thread about policy compliance.
Last month I blogged about policy management. Today I'll explore the policy management process and governance in more depth in the context of information risk and security or cybersecurity if you will.
In my experience, managers who are reluctant or unable to understand the [scary cyber] policy content stick to the bits they can do i.e. the formalities of 'policy approval' ... and that's about it. They leave the experts to write the guts of the policy, and even take their lead on whether there ought to be a policy at all, plus what the actual policy position should be. I rather suspect some don't even properly read and understand the policies they are asked to approve, not that they'd ever admit it!
The experts, in turn, naturally concentrate on the bits they are most comfortable with, namely writing that [cyber] content. Competent and experienced policy authors are well aware of the potential implications of [cyber] policies in their areas of specialty, so a lot of their effort goes into the fine details, crafting the specific wording to achieve [their view of] the intended effect with the least amount of collateral damage: they are busy down in the weeds of the standards and procedures, thinking especially about implementation issues and practicalities rather than true policies. For some of them anyway, everything else is dismissed as 'mere formalities'.
Incompetent and inexperienced policy authors - well, they just kind of have a go at it in the hope that either it's good enough or maybe someone else will sort it out. Mostly they don't even appreciate the issues I'm discussing. Those dreadful policies written in pseudo-legal language are a bit of a giveaway, plus the ones that are literally unworkable, half-baked, sometimes unreadable and usually unhelpful. Occasionally worse than useless.
Many experts and managers address each policy independently as if it exists in a vacuum, potentially leading to serious issues down the road such as direct conflicts with other policies and directives, perhaps even laws, regulations, strategies, contractual commitments, statements of intent, corporate values and so forth. Pity the poor worker instructed to comply with everything! The underlying issue is that the policies, procedures, directives, laws etc. form a complex and dynamic multidimensional matrix including but stretching far beyond the specific subject area of any one: they should all support and complement each other with few overlaps and no conflicts or gaps but good luck to anyone trying to achieve that in practice! Simply locating and mapping them all would be a job in itself, let alone consistently managing the entire suite as a coherent whole.
So, in practice, organizations normally structure their policies into clusters around business departments such as finance, IT and HR. If we're lucky, the policies use templates making them are reasonably consistent in style and tone, look and feel, across all areas, and hopefully consistent in content within each area ... but that enterprise-wide consistency and integration of the entire suite is almost as rare as trustworthy politicians.
That, to me, smells very much like a governance issue. Where is the high-level oversight, vision and direction? What kind of pie is it and how should it be sliced? Should 'cyber' policies (whatever that means) be part of the IT domain, or risk, or [information or IT] security, or assurance ... or should they form another distinct cluster? Who is going to deal with all those boundaries and interfaces, potential conflicts and overlaps? And how, in fact?
But wait, there's more! Re the process, have you ever seen one of those, in practice - an actual, designed, documented and operational Policy Management Process? They do exist but I suspect only in mature, strongly ISO 9000-driven quality assurance cultures such as aerospace, or compliance-driven cultures such as finance, or highly bureaucratic organizations such as governments. Most organizations just seem to muddle through, essentially making things up as they go along. As auditors, we consider ourselves fortunate to find the basics such as identified policy owners and issue/approval status with a date! Refinements such as version numbers, defined review cycles, and systematic review processes, are sheer luxuries. As to proactively managing the entirety of the policy lifecycle from concept through to retirement, nah forgeddabahtit!
Compliance is an example of something that ought to addressed in the policy management process, ideally leading to the compliance aspects being designed and then documented in the policies themselves and at implementation time being supported by associated awareness and training, metrics and activities to both enforce and reinforce compliance. Again, in practice, we're lucky if there is any real effort to 'implement' new policies: it's often an afterthought.
Finally, there's the time dimension: I just mentioned process maturity and policy lifecycle, but that's not all. The requirements and the organizational context are also dynamic. Laws, regs, contractual terms, standards and societal norms frequently change, sometimes quite sharply and dramatically (GDPR for a recent example) but usually more subtly. Statutes are relatively stable but the way they are interpreted and used in practice ('case law') evolves, especially early and late in their lifecycles - a bathtub curve. Various implementation challenges and incidents within the organization quite often lead to calls to 'update the policies and procedures', whether that's amending or drafting (seldom explicitly withdrawing or retiring failed or superseded policies!), plus there's the constant ebb and flow of new/amended policies (and strategies and objectives and ...) throughout the business - a version of the butterfly effect from chaos theory. And of course the people change. We come and go. We each have our interests and concerns, our blind spots and hot buttons.
Bottom line: it's a mess because of those complications and dynamics. You may feel I'm over-complicating matters and yes maybe I am for the purposes of drawing attention to the issues ... but then I've been doing this stuff for decades, often stumbling across and trying to deal with similar issues in various organizations along the way. I see patterns. YMMV.
I'm not sure these issues are even solvable but I believe that, as professionals, we could and should do better. This is the kind of thing that ISO27k could get further into, providing succinct, generic advice based on (I guess) ISO 9000 and governance practices.
There's still more to say on this - another time. Meanwhile, I must press on with the awareness and training materials on 'spotting incidents'.