Humans are O(2^N): Dealing With Unmanaged and Human Induced Complexity

Once upon a time, I found myself pitted against a pack of fire-breathing dire wolves with only a spoon and cereal bowl to defend myself. Let’s try that again. Once upon a time, I attended the kind of meeting that every single IT worker  eventually attends. It’s the usual story, some folks outside of IT desired to automate a number of workflows that were, at the time, almost entirely manual. They wanted to make them go faster. So, of course, we needed to look at their workflows, so we sat down and had a meeting. We started to walk through the workflows, seeing where we could automate. As the meeting progressed, the number of exceptions, sub-processes, changes in status due to special attributes, judgment calls, and other decision points began to overwhelm all the participants in the meeting. It was maddeningly byzantine from soup to nuts. It didn’t take long before the complexity became apparent to everyone present and the individuals that had called the meeting eventually paused, looked at each other, looked at me, and told me that they’re going to have to go back and map all of this out and come back. Yikes! The question that entered no one’s mind (sadly, including mine) was this, “Who’s doing all this work currently?” The answer is human middleware.

Mark McDonald’s blog post has really stuck with me since I’ve read it. Organizations will use people to keep themselves from doing the scary and difficult work of changing the way they do business. The world has changed; the nature of work has changed, customer expectations have changed, and yet, the organization continues unthinkingly in the way it always has. With my own example, no one ever called a time out and said, “Who is the customer? Can they navigate this system without assistance? Why is it this way in the first place? Can we change? What are the consequences of changing it? Is changing it worth investing in?” Instead, we all (including myself, I’m quite guilty in this story) just wanted to automate the current existing process. There was no impulse to make it simple. To make it repeatable. The consequences, had we proceeded, would have been to create a digital Rube Goldberg machine. Digital paper shuffling is still just paper shuffling.

Some of the symptoms Mark describes we all see (from his blog post):

• Inconsistent business processes, which create multiple versions of the truth, conflicting business rules, complex operational interfaces, and different ways to get things done. All of which create bottlenecks, backdoors, and conflicting answers requiring people to figure it all out
• Baseline budgeting, that makes resources decisions based on a prior year’s baseline that bakes in inefficiencies and creates an incentive to keep and grow the middleware and budget as a symbol of executive influence or power.
• Accretive change which is constantly asking for more, different, and additional stuff rather than favoring of keeping the organization capable and lean.

What’s the way out? Well, it will require an evolution of your organizational culture. Evolution doesn’t tend to be self-initiated, there’s usually some form of pain required. The good news is that the pain doesn’t necessarily have to come in the form of layoffs. It can come in the form of asking the tough questions we tend to shy away from. At the tactical level, good work can be done. If we go back to the example I led with, we shouldn’t have had just a simple meeting with workflow diagrams. We should’ve had some form of Kaizen event. We should’ve reframed the conversation from mere automation to true improvement. We should’ve asked the, “5 Whys”. We should’ve led with customer value instead of looking at merely improving efficiency, especially as some reductive coefficient to the current workload.

Complexity is inevitable but not all complexity is created equal. We should strive for the complexity in our organizations to be the end-result of simple rules being applied over and over again not as a result of unmanaged and informal processes. When we can place powerful computational resources at our backs, we can move people from being cogs in the machine to providing real value. They can start doing  things computers fail miserably at such as delivering empathetic and creative value. Fact of the matter is, computers suck at giving advice. Customers will need advice. They will need human contact. Is your organization using humans as humans or as fuzzy logic gates?

Note about the title: The title corresponds to the computer science concept of big O notation; that is, the understanding of the limits of algorithms as their set sizes approach a particular value. I have no empirical evidence that people working together are O(2^N), also known as EXPTIME. There’s probably a study out there. I just haven’t found it yet.

Getting Punched in the Hall

Last night I took a call from a friend who’s been tasked to architect an update to an aging infrastructure for a large organization. The state of their infrastructure was in such a state of disrepair that daily crises, lack of available parts, and lapsing warranties prompted the organization to migrate to new hardware.

His call was prompted by the fact that, in the preliminary work that he had done, the services he had moved consumed significantly more resources than was anticipated and by a wide margin. The vast disparity called into the question what it would take to bring overall performance in line with industry standards. The cost estimates were so far outside what was budgeted that the only recourse would be to prioritize services and proceed from there.

As I asked about simple stuff such as who has an understanding of what services are critical and which ones aren’t, he said something that really struck me, “You know, every day I walk down the hall and every day someone jumps out and punches me in the gut with a new issue. I can’t tell which issues should have priority. No one knows what the business critical services are outside of, ‘All of them.'” There’s the heart of it. There’s no organization agreement on what has priority in the place of business and without that agreement, decisions are temporal and political. In fact, I would theorize the vast state of neglect is symptomatic of a long history of failure to agree.

So you wake up and find yourself in an organization that has doesn’t have strong leadership. You’re tasked with improving conditions and have some budgetary support within your own business unit. After that, things drop off quickly. Petty bickering and in-fighting dominate meetings and decisions are put on hold. Nothing calls up professional fight or flight like the above situation. You’re ready to label the organization a ship of fools and jump ship to a more attractive offering.

Not all organizations are created equal. There’s nothing wrong with leaving but make sure you leave it in a place better than you found it. In this particular situation, I called upon the advise a Gartner analyst once told me,  “Don’t be afraid to be a little wrong. Few things prompt immediate engagement like correcting someone.” My advise was for him to create an informal catalogue of the services his organization provides (however incomplete), document the supporting infrastructure elements and, this is key, for him to order them by what he thought was the most critical. Then he should call a series of cross-group work sessions to refine the list and gain consensus, starting at the operational level. He would need to make sure to recruit individuals who are likely to be engaged in the work and who are, “Dedicated to winning the World Series, not to getting a better contract.”

Build informal groups that are dedicated to excellence within the organization, engage your peers, and build pockets of high-performance. Create a conspiracy of success with the peers you partner with to deliver service to your customers.

When dysfunction dominates an organization’s culture and puts a stranglehold on decision making, the answer isn’t to look up for the solution. The answer is to provide leadership yourself. It’s not uncommon for executives to become defensive when the business isn’t performing as it should and to point fingers at each other. Instead of mirroring that behavior at the operational level, it’s important to fill that void with strong positive leadership. By doing so, you’re organically changing the culture of the organization. When you are successful and bring about an improvement that each of the representative groups can take credit for, you stand a good chance of improving the probability that those that represent you on the board of directors will work together in the future. Success always has many parents. Positive leadership is remarkably resilient and contagious. All it takes is one person to convince a few others to commit to it (and the grit to remain committed).