Attach Buttons And Understand Critical Thresholds

Why Attached Buttons Two by Two Help Understand a Benefits Realization Curve

Complex Program Benefit Realization Can Be Slow

[This article is inspired by the book: The High-Impact PMO, How Can Agile PMO Deliver Value in a Complex World I have published in October and that is available on Amazon]

Grigori was once the PMO of a company-wide program whose deployment concerned several tens of local organizations around the world. Each organization had to install a lean management system in order to improve their performance.

All organizations had to be connected to several others at one time or another one. For example, they supplied other organizations with products or services. They had interfaces with functional departments like Finance, Marketing, and Information Technology.

The program was going to last at least five years due to the time required to change in depth the organizational culture. Several organizations volunteered as pilot organizations and developed their lean management practices with the support of Master Black Belts.

Grigori was in the early years of the program. Yet, the program sponsor was expecting quick visible improvements in the global performance of the company. Alas, even after several pilot organizations had adopted the lean management principles, no real high-level gains were visible. Benefits realization was far behind the initial premise.

What could be the problem?

How 1000 Randomly Attached Buttons Help You To Discover the Phenomenon

Stuart Kauffman is a complex systems researcher who studies the origin of life on Earth. Kauffman explains the following simple thought experiment in his book At Home in the Universe, the Search for the Laws of Self-Organization and Complexity. It is about random graphs. Random graphs are a set of dots, or nodes, connected at random by a set of lines, or edges.

You can realize this toy problem at home yourself right away with buttons. Take all the buttons ripped off your shirts you can have saved, add maybe a few more, let say one hundred. Take some sewing thread. Now spread these buttons across your dining room floor and randomly attach two buttons at a time using bits of thread.

While you progress by linking buttons randomly two by two, one after the other; every time you have connected two buttons, pull one random button and observe how many buttons come together. At a certain point, you will pass a critical threshold. When lifting one button randomly, you will get most of the other buttons as well. You won’t need to have connected all buttons.

You do not need to link all buttons with a thread to pull them all together at once[i].

Well before normally expected, most buttons will have been interconnected and grouped into a single set of buttons. This threshold is represented in the above figure. It appears when around 50% of the buttons are connected.

The curve is what we call an S-shaped or a sigmoidal curve. The steepness of the curve at this critical point depends on the number of buttons. When this number is small, the steepest part of the curve is shallow. When the number increases, the steepness becomes more vertical.

This phenomenon is similar, explains Kauffman to phase transitions. During a phase transition certain properties of a medium change, often discontinuously, because of the change of some external conditions (here sewing two buttons).

Can this experiment serve to explain why benefits realization in a large complex program may not be a linear curve?

From Isolated Components to an Integrated Network of Connections

Imagine that instead of buttons, you repeat the same exercise again with people or with processes.

The properties of the isolated components do not define the complex system they form. What is connected, how it is connected, and to what it is connected, are all dimensions of great importance to understand observed system behavior. The nature and structure of the connections define the system.

In the real life of an organization and its ecosystem, these people and processes will be already more or less connected. Like in the button experiment, at some critical level of connectivity, the system stops being a set of apparently isolated components and becomes an integrated network of connections.

Here also we have this significant behavior change at a critical point.

You have surely already seen that your organization generates such transitions as the result of changes and connections you introduced. What is called transformation in an organization is nothing more than a successful transition. What is interesting for a Project Practioners and especially for PMOs is to understand as much as possible this phenomenon of transitions and how it is generated.

Tipping points are at the end of this process.

Grigory hypothesized that one of the possible reasons he did not yet see benefits in the deployment of his program was maybe that not enough organizations were connected. He did not have to pressure the system. On the opposite, he had to let the interconnections develop without creating unnecessary constraints on them. His role was only to remind the vision of the lean system they all looked for and to facilitate what needed to be.

Your remarks, own experience, and additions will be warmly appreciated.

To Your Continued Success!


Check out my new Book that explores a series of real life snapshots showing how project management practioners and especially PMOs can confront a VUCA world.  It gives valuable insights that will allow you to more successfully navigate the wave of complexity that is coming our way.

You can also contact me at

[i] Picture inspired from Stuart Kauffman, At Home in the Universe: The Search for Laws of Self-Organization and Complexity, 1995,

Leave a Comment