What is the optimal business process for delivering a disruptive product that has never been made before? That question is like asking Ikea to write assembly instructions for an item of furniture that hasn’t yet been invented.
In the past when the world was simple, when phones just made phone calls, when business models lasted for decades not days, when offering to digitally disrupt someone was a lewd act, the life of an organization designer was easy. Almost every business on the planet could be described by the steps needed to produce the product. Want to make a car? Here’s the material, machinery, people and processes you need, its written down, its been tested, proven and documented hundreds of times over.
Now we live in a world where we are delivering services that were invented weeks, if not days ago. There’s no proven business model, no defined processes, no organizational structure blueprint to start from. You could say we have no idea how to deliver the outcome, the only thing we do know is what we want the outcome to be.
How do you build and run an organization to deliver an outcome when you have no idea how to best do it? That’s not to say you have no idea how to produce the outcome, you must have had some idea or you wouldn’t have decided to do it in the first place. But what you don’t know is how to consistently and efficiently produce the outcome.
When Jeff Bezos started Amazon.com he wanted to produce an outcome, selling books online. He had some idea how to do it, but as no one else had ever done it at scale there was no best practice to follow, no proven business model, and no organization structure to copy. He knew they needed to take orders, pack orders, and ship them. But the process didn’t exist and neither did the metrics of how many people in each role you needed to cater for certain volumes. He had no idea what characteristics of service would delight customers, and which would frustrate them.
How many of our businesses are in a similar situation today? We know there is a need in the market for a new business model, a disruptive one that delivers a new outcome, and we know that we have the basic building blocks to deliver it. But how do we write the business processes to deliver it? How do we create the Standard Operating Procedures to give our people clarity in their work if we have never done it before, if no one has ever done it before?
The end of the Designed Organization
In a world where the outcome has never been delivered before, the old model of designing an organization in which roles, processes and structure are designed ahead of time to optimally deliver an outcome is now impossible. Optimally being the keyword, for an organization with some base knowledge it wouldn’t be difficult to map out a process for delivering an outcome even if no one had ever done it before. But businesses don’t tend to thrive because they manage to deliver an outcome, they thrive because they optimally deliver it. The organization that does it best wins, not the one that just manages to do it.
The challenge is that once a structure and set of processes have been put in place it becomes very hard to change them. Organizational processes set faster than quick drying concrete on a summers day, and trying to radically change them, or even tweak them once underway is harder than solving world hunger. If you do manage to design an organization that delivers the outcome you are seeking, you will be stuck with it. “It works so why change it?”, and “Don’t fix something that’s not broken” will be the mantras that put the handbrake on your path to Optimisation. Fortunately, you won’t be stuck with it forever, because very soon someone else will come along and deliver the same outcome, faster, cheaper, or better and you’ll be out of business.
Okay, this isn’t anything you didn’t already know, the question is what can we do about it? If the “designed organization” doesn’t work in today’s world of rapidly changing outcomes, what does?
Introducing the Neural Organization
Before I explain the concept of the Neural Organization I need to take you on a small trip over to the world of neural networks and machine learning. Don’t panic, it’s a short trip and we’ll be back here in a moment.
Neural Networks – Figuring out answers when we don’t know the question.
Have a read of my an article on neural networks to understand them in more detail, but for now just know they are a form of Artificial Intelligence, more specifically a type of machine learning. What they are brilliant at is learning from repetition what works and what doesn’t, and they do it without needing to be told why.
If I asked you to list a set of rules to identify a cat in a photo you would probably come up with some obvious ones such as:
- They have fur, a tail, pointy ears, about the size of a small dog, and four legs.
We’ve all seen a cat and we can all do a pretty decent job of writing a business process for identifying if a photo has a cat in it. Just like we can write a process for producing the outcome we want our business to deliver. Or can we?
What if the cat is a Manx cat, they don’t have tails, what if its missing a leg, what if you can only see half of the cat in the photo? Do your business rules still work? No. A diligent employee that follows your business process of looking for a small four-legged animal with a tail and fur would have to throw those photos in the no-cat pile.
The problem is that we don’t actually know how to describe or write a business process for identifying a cat in a photo because its actually really, really complicated. Just when we think we’ve got it, a photo that we weren’t expecting comes along with a hairless Manx cat sitting upside down in a vase. Just like we don’t know how to write a business process to optimally deliver an outcome because things will always happen that we didn’t expect.
This is where neural networks come in. We don’t need to tell a neural net how to spot a cat, we just tell it when it has spotted one. Then every time it gets it right it learns and updates itself (changes its network connections) to make it more likely to get it right again, and before you know it, the neural net can spot a cat in a photo with almost perfect accuracy. And here’s the key thing, you never even had to tell it what a cat looked like.
Now imagine that in an organizational context. Every time you produce an outcome that makes your customers happy you learn from it. Your organization changes slightly every time you deliver an outcome, if it was good you reinforce the way it happened, if it was bad you tweak something.
The Neural Organisation
Today, large organisations are bound up in processes and policies, these have been written to optimally produce a consistent outcome in a known set of circumstances. But if we change the circumstances do the processes and policies still work, or do they start to get in the way? How fast can you change your processes and policies to adapt to the new circumstances? That was a rhetorical question, there is no way on earth that a policy change will happen in time to adapt to the pace of change happening around you.
Intelligent beings or warm cogs?
A thing that we often forget is that an organization is a collection of highly intelligent sentient beings, that are intrinsically motivated to the do the best they can with the resources they have. We often treat our intelligent self-aware resources (our people) as dumb process followers, and we place more importance on process compliance than on intelligent decision making. We take the vast capability of the collective intelligence of our people, and we dumb it down to using them as warm cogs in a machine.
Inside a neural network is an array of self-learning nodes that improve with every outcome, learning what produced the right result, and reinforcing it, or changing if the result wasn’t right. This is what makes neural networks work, we let each individual node learn and change the way it works, independent of any overall process or policy.
We have invented artificially intelligent neural networks that consist of self-learning nodes, and we trust those nodes to learn and change, in fact its not just trusting them to change, we want them to change, that’s what makes a neural network work. Let’s be clear, those nodes that we trust to make their own decisions and independently change the way they work, they aren’t actually intelligent, in fact they are really dumb.
Yet we take an organization of highly intelligent nodes (people) that are far better equipped to learn and adapt than a dumb neural node, and we force them to blindly follow policies and processes, removing their ability to think and adapt. To a large degree we still run organizations in a century old Taylorist model, controlled by processes and policies intended to last for decades, not days.
The next evolution of organization – the neural organization
What if your organization didn’t have any processes or policies? What if your people were entrusted to constantly adapt the way they do things? What if instead of trying to assimilate your pool of intelligent beings through command and control, you set them free to learn, think and act independently? What if you could harness the full intellect of every person in your organization to constantly adapt and improve? Do you think you’d have a more powerful capability than a bunch of warm cogs following a process?
Unleash the power of the many, discard the process of the few.
Ah you say, but that would be chaos, everyone acting individually, no order, no process, just chaos. Not quite. In removing processes and policies you replace them with purpose and pride. If all your people have a common sense of purpose and a pride in what they do you won’t get chaos, you will get focus and collaboration.
Machines need process. People need purpose.
Purpose is something startups are very good at, they have a strong sense of why they exist that everyone in the organization understands, and a sense of pride in striving to achieve it. This means they can prosper without processes and policies, you don’t need to tell people what to do if they inherently have a sense for what needs to be done. They may be producing a product or service that has never existed before, so there is no process for it, instead there is a collective purpose and an array of individual intelligence all constantly learning and thinking about what it takes to achieve the purpose.
The concept of a neural organization fits perfectly in a startup model. For organizations familiar with rapid prototyping this should come naturally, except instead of rapidly prototyping the product, a neural organization rapidly prototypes the structure, roles, and processes.
Rapid prototyping improves the “what”. Neural organizations improve the “how”.
The concept of a neural organization gets a bit harder when applied to a large existing organization, particularly where a culture of process has overtaken purpose. The majority of large organizations would struggle to find a united purpose across all of their people. There may be groups that have a visible or unspoken purpose (such as a customer service team, an R&D department or sales) but it is highly unlikely that across the entire organization there is a single unified purpose that all members of the organization are driven by. In absence of a unified purpose, processes and polices are used to coordinate activities to produce a predefined outcome.
The time has come to accept that the outcomes we deliver, and the way we best deliver those outcomes is in a constant state of flux. Holding fast to our polices and processes is the surest way to hold back the progress of an organization. Unleashing our people to act individually as the intelligent beings that they are, and giving them a common sense of purpose and pride is the key to future organizational success. It is time to become a neural organization.
[…] Neural Organizations: The Future of Organizational Intelligence? […]