Here’s how a corporate death spiral begins. The first act begins with the product team. Its goal is to keep offerings attractive in the market. But in its daily work, many of its processes are manual, complex, poorly documented, and incredibly fragile. There are many workarounds. Managers are relying on quick fixes. Time is never enough.
Act two. A competitor or a new startup dreams up a new offering. Someone high up in the company makes a bold promise that dazzles investors. Now the product team is tasked with an urgent project. They need to deliver a set of new product features.
Employees then discover that the new web page can’t be fully integrated with the inventory system. This, in turn, impacts the existing workflow built by the Quality Assurance department. Failing that step of checking violates another rule set by the Finance department. There are numerous conflicting demands. People are bypassing standard operating procedures in order to “get things done.” This sets the stage for the third and final act.
A steering committee steps in to arbitrate conflicts. The product team is told to freeze its requirements so that IT and manufacturing can move into production. Everyone works overtime to communicate, coordinate, and get approvals. The press release is being drafted for the imminent product launch.
The product finally hits the shelves. But the market responds with a yawn. The hype fades away as quickly as it arrived.
There are many reasons for the lukewarm response. There are some missing displays in the retail stores. There is an IT outage during the peak hours of website traffic. There is some confusion over the pricing promotion created by the sales team. Then there’s the suspicion from consumers that the product packaging has not been well designed.
In the “after-action review,” the management team concludes that it’s an execution problem. We need to have better alignment, they say. The market, meanwhile, is moving on to the next big thing. The product team has a new idea…
The question is obvious: How can you get out of this cycle?
Why large companies struggle with growth
The scenario above is commonplace. It’s especially common within large, traditional companies. Managers are often working within tightly coupled, monolithic business units. Organizational handoffs—from R&D to manufacturing to marketing to customer support—rely on face-to-face coordination and email chains. They require multiple approvals. The product development cycle is long. The production ramp-up is longer still.
These managers would have a hard time precisely mapping out the end-to-end workflow. Even top executives would have trouble “connecting the dots” across the company. No one knows exactly what’s happening daily. All they have is a vague idea. Operating manuals scatter. And if people spend a month or two drawing up a comprehensive map, it probably looks like this.
The organization is complex with complicated decision-making. It confuses everyone. An organization of this nature can only execute the most routine tasks. But even doing that is stretching the company’s limits because the system is error-prone. Everyone seems to be in a constant state of exhaustion. People find it almost impossible to find time to think. Innovation is a luxury.
You may think tech companies are immune to such complexity. Some may think companies that move “digital bits” without delivering “physical atoms” always have a simpler existence. But that’s simply untrue.
Google, Facebook, LinkedIn, Netflix, eBay, Twitter, and Etsy all went through near-death experiences. Virtually all tech startups began with a tightly coupled system in a single code base. Even Amazon started in 1996 as a single application. It ran through a web server that talked to a database on the back end. Over time, these startups all sprawled into monolithic architectures comprising millions of lines of code.
You can’t grow further with these enormous architectures where all the functionalities are sitting within one application. So these companies had to break it up. They turned the entire IT infrastructure into tens of thousands of microservices. These microservices could then exist as independent, decentralized modules that had standard interfaces with other modules. Let me make this more concrete.
Breaking down a complex IT system into simpler parts
Historically, software developers have been responsible for app features. Meanwhile, IT operations have been responsible for testing sites and the actual production environment on the computer servers. It’s natural that app developers would coordinate with IT operations. They would coordinate the testing schedule, feature launches, and troubleshooting. And all the intricate processes and dependencies have lived in people’s heads.
The survivors in the IT sectors are the ones who rebuild the organization from the ground up. There is no escape. They simply bite the bullet before time runs out. They do so by keeping all the pre-existing IT functions unchanged. But the input and output of each function get standardized. Communications between different teams are made through APIs (application programming interfaces). This means less ad hoc email and fewer manual Excel spreadsheets.
Now, the format of the requests between app developers and IT operations becomes consistent. The next step is to automate those routine coordinations. All of a sudden, app developers can run production-like environments on their own laptops. People can create, on-demand, the testing environment that developers need via self-services. The role of IT operations changes too. IT is not responsible for manual coordination. They no longer push code manually to the actual production. Instead, they focus on designing a better, faster production environment. In other words, app development and IT operations are decoupled.
Now hold that concept and port it over to a manufacturing setting.
Manufacturing organizations with agility
Consider a traditional manufacturer. Its centralized manufacturing plant shares a radical idea with the other five different product groups. There will be no more product launch meetings. Instead, all the factories will make their production schedules transparent, daily. Product groups can upload their requirements online. The manufacturing plants will then display when the required capacity is available. The online system will show whether they can meet the requirements. And if it is impossible to ramp up production fast enough, any product group is free to negotiate with other external contract manufacturers. The system even shows detailed contact information of these external parties.
No more email. Everything is done via a service platform that looks like Apple App Store.
Now you ask, am I dreaming? This can’t possibly be done, some may say. But there are companies that have already done so or have already moved very far in that direction. And not just digital natives, like Netflix. It’s not just Amazon, either.
Haier Group—one of the world’s biggest home appliance manufacturers—is one example of a company that has taken this approach. For years, the China-based Haier has organized itself as a swarm of self-managing business units. Haier has some 4,000 microenterprises (MEs). Each comprises just 10 to 15 employees. Instead of being centrally orchestrated, these MEs independently transact with one another. They are given full autonomy to deliver the final product to consumers. Certain MEs manufacture specific component parts; others provide services like HR or product design. Coordination is managed through internal platforms, making the idea comparable to an app store.
What Haier has pursued is a modular organization. Its architecture is loosely coupled with well-encapsulated units. No more monolithic businesses. Since the interfaces across these MEs are highly standardized, automating information exchanges becomes possible.
In such settings, managers can request services without face-to-face meetings. No email. No phone call. Just one click away. There is no need for project planning sketching beyond the next three months. You can even get rid of project escalations because no one is holding you back except yourself.
Each team is small and self-contained. If there’s a failure, it won’t cause global disruption. That’s system resilience. Remember, a big organization is brittle unless you break the monolithic system into tiny units.
How to jumpstart simplicity
All these benefits may seem like something from the distant future for most of us. You and I are probably not working at Netflix or Haier. But the seed of a movement is always in your hand. Think, for example:
- In your immediate team’s regular work, can you identify standard, routine work versus work that is nonstandard? The Pareto principle suggests that 80% of your department’s time is spent doing routine tasks, while 20% is not. For that 80%, do you have a clear picture of your department’s information flow? Think about the inputs and outputs.
- Once you have a full picture of how these routines work, identity opportunities for simplification. Can you make information transparent to other departments and transmit it through a self-service? Are there process steps that make no sense? To what extent can these routine tasks be outsourced or even automated? Your aim is to eliminate those annoying emails and telephone calls that get in your way every day.
What you are doing is no mere housecleaning. You are disciplining yourself to pay down the coordination debt of your organization. If you don’t, the workflow complexity will explode like compound interest as your company’s scope of work increases.
Most of us are not working at Haier or Netflix. But the benefit of using these techniques, however imperfectly, is to preserve our own sanity. High quality and large quantities of work will always be demanded from your team. But everyone deserves to have a life outside of work. To achieve that, you need to pay down the coordination debt.
P.S., Does your company struggle with complexity? Have you observed good practices that overcome such struggles? Let me know what you think. Join our discussion below.