Complexity reduction - When less is more

You may have seen the meme with a guy in front of a board trying to connect the dots into a sophisticated scheme. Although this scene from Always Sunny In Philadelphia is more about a conspiracy, it vividly shows how tough it is for our minds to grasp complex situations.

So, let’s start from the basics. A two-element system has a single 1:1 relationship. Just Alice and Bob, and that’s it. But the more elements you add, the more 1:1 relationships you have to deal with. Eight elements can have a total of 28 pair relationships, which leads to 40,320 permutations.

The more moving parts you have in a system, the more potential points of failure you have to address. And, unsurprisingly, it’s more difficult to manage. Add to that increased operating costs, inefficiencies and reduced productivity, and it’s a no-brainer: simpler is better. 

But achieving simplification is easier said than done. 

When left unchecked, complexity grows over time. This is completely natural as a business grows and its technology evolves. 

That’s why reducing complexity must be an enterprise-wide, continuous effort. 

Complexity reduction is a strategic undertaking to simplify a company’s organizational structure, business processes, products/services, information systems, infrastructure and/or software systems. While complexity is inherent to any organization, especially as it grows, reducing it means eliminating unwanted complexity. 

Complexity, in general, can be a competitive advantage. For example, an increase in the number of Stock Keeping Units (SKUs) correlates with a rise in EBITDA (Earnings Before Interest and Taxes Depreciation Amortization) — up to a point. Algorithm complexity can also allow the system to outperform its competitors; just look at Google’s ranking algorithms. 

Unwanted complexity, in turn, refers to the unnecessary, wasteful connections and dependencies between components and elements. These dependencies don’t add value to the business. Instead, they slow down processes, increase operating costs and serve as additional points of failure. 

Regarding the technology estate, its complexity also inevitably increases with time. Bespoke systems, designed to accomplish one thing, are modified time and time again to keep up with evolving business needs. New operations require additional solutions, which in turn must be integrated with existing systems. Advancements in technology add applications to the software stack without replacing outdated solutions. 

On and on it goes. 

Without simplification, a technology estate can reach a point of complexity where it stifles innovation instead of enabling it. Unwanted complexity also results in systems requiring more staffing — and more financial resources — to simply keep things running as they are. 

It’s ultimately a value drain. For example, had The Weather Company not switched from 13 data centers to one, it likely would not have been able to cut down the ratio of time dedicated to maintenance from 80% to 30%.   

To make matters worse, overly complex technology estates increase risks, decrease business agility and prevent organizations from identifying and acting on new business opportunities. 

Eliminating unwanted complexity allows you to: 

• Reduce operating costs 
• Improve customer and employee experience 
• Improve the organization’s ability to innovate and adapt 
• Accelerate time-to-market 
• Reduce risks 
• Quickly identify the root causes of errors and fix them 
• Ensure alignment between teams and processes 
• Improve accuracy and consistency in data analysis 

Divide and conquer was Caesar’s approach. But you have even more tools at your disposal. Here are seven methods for your toolkit to reduce complexity. It would be unfair to say one is better than the others, however. Which methods willwork best depend on the condition and complexity of your current technology estate, as well as your business needs. 

Retire 

The simplest way to reduce complexity in a given system is to reduce its number of elements. When it comes to the technology estate, it means identifying the systems that don’t add value to the business or have excessive opportunity costs — and retiring them. 

Before retiring a legacy system, you must make sure you preserve all the data you may need in the future. You may also need to migrate it to a new system before retirement. 

Divide

This approach involves dividing a system into multiple smaller systems — segments. Segmentation usually allows for the improvement of each element’s efficiency for a specific scope of tasks. Think about market or target audience segmentation: the underlying principle is the same. 

When applied to the technology estate, network segmentation is one example of this approach in action. Network segmentation means dividing a network into several subnets that each act as a self-contained smaller network. It reduces congestion, improves network performance and enhances network security. 

Modularize

This approach means switching to a modular architecture in your infrastructure and software solutions. In this type of architecture, any given system consists of interconnected yet independent components — modules. They can be changed without bringing the whole system down. 

The modular approach has multiple benefits, including: 

• Component reusability 
• Facilitating testing 
• Enhancing scalability 
• Improving maintainability 
• Streamlining change deployment 

Standardize

Process and data standardization are among the efficiency-improving tools with the highest ROI (Return on Investment). 

Process standardization is key to ensuring a consistent output quality across the organization. Well-documented, standardized processes are also a time-saver for employees — and a powerful driver of operational efficiency. 

Process standardization is an equally valuable opportunity to analyze the current workflows and identify and remove steps that increase unwanted complexity. Plus, documentation will also make processes reviewable, enabling regular reassessments and adjustments. 

Data standardization ensures that data can be used across systems to deliver better insights into the organization’s operations or to enable automation. 

Consolidate

If two or more systems accomplish similar or identical goals, replacing them with a single solution is another way to reduce complexity. For example, multinational companies can benefit from consolidating local systems into a single cross-jurisdictional solution. 

NBC Universal, for instance, used 19 separate ad-selling applications at one point. The company’s CIO (Chief Information Officer) led the effort to combine all of them into a single solution, improving inventory visibility and accuracy. 

Under this approach, you can also create functions or teams working on complementary activities, or consolidate similar services into one. 

Just keep in mind that consolidation should enable synergy, be it in terms of costs, capital or revenue. Otherwise, it may only add to complexity instead of reducing it. 

Reorganize

Sometimes, the order of processes or the way system elements are organized is the root cause of waste. In this case, rearrange them to facilitate and simplify the process. For example, this can work for the data flow between systems or governance processes. 

Automate

Take stock of all the manual processes still in place. Chances are high that some of them — if not all — can be automated. 

Automation doesn’t just reduce complexity. Take testing as an example: QA (Quality Assurance) testing automation speeds up time-to-production, ensures consistency, eliminates the risk of human error and enables continuous testing. And, of course, it reduces QA costs. What’s not to like? 

Here’s a word of warning: processes must be simplified before you automate them. Otherwise, you’re automating complexity, not reducing it.