The newest “best practice” is an old banned practice.

Guess what’s now “best practice.”

A few years back it meant locking PCs down tight, preventing anyone outside IT from doing anything creative with information technology unless they could get it done with Excel.

“Shadow IT” … business departments implementing information technology without any IT involvement? An information security nightmare!

Nightmare? Look at the number of recent, massive data breaches whose targets have been … what shall we call the opposite of shadow IT … sunlit IT?

The actual evidence seems to show that shadow IT is, at worst, no less secure than sunlit IT.

The exact same cast of characters that advocated tight lockdown just a few years ago are now, without even a trace of contrition, writing favorably about how smart CIOs are supporting and leveraging shadow IT rather than trying to stamp it out.

Research current thinking on this and you’ll find the IT punditocracy has discovered IT has to handle some tasks more quickly than others. McKinsey, for example, calls it the “two-speed IT architecture.”

Welcome to the party, folks. You’re late. Did you at least bring some decent champagne?

To be fair: Just because these folks are latecomers, it doesn’t mean their insights are worthless. The McKinsey piece, while long on what the outcomes should look like and light on how to achieve them, is a fairly decent analysis.

Decent, that is, other than ignoring the significant role shadow IT will and should play in most two-speed IT architectures.

And, decent with this possible exception: “… companies need to improve their capabilities in automating operations and digitizing business processes. This is important because it enables quicker response times to customers while cutting operating waste and costs.”

If, by “quicker,” McKinsey’s analysts mean reducing cycle times for fulfilling orders, then fair enough. While the notion that information technology can be helpful in reducing process cycle times is hardly a fresh one, it’s no less valid today than it was when first introduced in the mid-1970s or thereabouts.

But “quicker” might also mean implementing new strategies or responding to marketplace changes. If that’s what McKinsey means by “quicker,” it’s way off, because optimized processes are one of the biggest impediments to rapid change. That’s because optimized processes need supporting infrastructure, and infrastructure encourages stasis, with “infrastructure” including:

• Business process design. As these things go, this is the easy one. Designing an efficient process just isn’t all that difficult compared to what it takes to implement one. Visio is the most adaptive component of process infrastructure.

• Design of business process controls. Beyond the process itself are management practices and process metrics. Designing them isn’t all that hard. Designing them so they don’t do more harm by getting in the way than they provide in benefits? Trickier. A lot trickier.

• Design and build-out of the physical plant needed to support the business process. This could be as trivial as setting up a new cube farm, or as complex and expensive as designing and building a factory. Ignoring all other aspects of this task, whatever gets built will have to last long enough to be relevant until the company has paid off the design and build-out costs.

• Design (or selection), implementation, and integration of supporting information technology. Presumably, KJR’s audience knows a thing or three about this topic. To make sure it doesn’t get lost: The integration part is, in most situations, the most expensive and difficult technical dimension of the task.

• Employee training. The need to train employees is obvious. That their education is part of your business infrastructure is less obvious.

• Investment in continuous improvement cycles until process optimization has reached the point of diminishing returns. No matter what the new process, it’s the organizational equivalent of a skill. Skills take time, and money, and effort to acquire and perfect.

There’s one more factor to take into account to understand why investments in business infrastructure put the brakes on business agility, and that’s the project failure rate.In most companies, more projects fail than succeed. That being the case, once a company has a process and its supporting infrastructure up, running, and optimized, the fear that implementing its replacement will fail isn’t merely very real. It’s realistic.

And when the odds of success are low, it’s perfectly natural for a company’s decision-makers to take what looks like the safer bet — sticking with that has worked, even if, in the long run, the result will be an obsolete company that sells and delivers obsolete products and services.

Especially because, in so many cases, by the time it’s impossible to ignore their obsolescence, it will be someone else’s problem.

* * *

Next week: Shadow IT’s role in a two-speed business (not just IT) architecture.