CHAPTER 5

New Fundamentals of Leadership and Management for the Digital Age

I would say to the House, as I said to those who have joined this government: “I have nothing to offer but blood, toil, tears and sweat.” …You ask, what is our aim? I can answer in one word: It is Victory, victory at all costs, victory in spite of all terror, victory, however long and hard the road may be; for without victory, there is no survival.

—Winston Churchill

If complexity requires new ways to collaborate, as we stated at the beginning of the last chapter, it also brings new considerations for leadership and management.

What is Leadership? Let’s be specific. In the context of organizations, it is about empowering people to pursue a shared vision.

A leader must therefore be capable of designing and impressing upon his/her people

1. The image of a desirable future
2. Why and how it is possible to achieve it

In his famous 1940 speech to the House of Commons that preceded the rescue operation in Dunkirk, Churchill does precisely this: highlights the Vision (“Victory”), the why (“without Victory there is no survival”), and the how (“Blood, Toil, Sweat and Tears”).

In our daily work, we may not face the same dramatic challenges of a war and we may not need the inflammatory rhetoric required for those dark days. But, if we want to pride ourselves with some sort of leadership skills, we must make sure that some of the basics are covered. Indeed, in what follows, we are concerned with leadership for organizations.

First, a leader owns a Theory, a set of well-tested assumptions within a well-defined realm of validity. Without Theory, Management—the activities we do to achieve our vision—becomes a “whack-a-mole” game and finger pointing becomes the rule.

Second, a leader is capable of communicating effectively inside and outside the organization why the Theory will produce the desired results; in other words, the role of the leader is to create predictability of outcome for the efforts he/she requires from their people as well as constancy of purpose among them.

Third, a leader is selfless and relentless in elevating everyone’s abilities through continuous teaching and mentoring.

How is all this applicable in the digital age? An analogy may help.

In the early 1980s, American car manufacturers had dramatically lost competitiveness and “Quality” came onto the scene. Quality professionals had a major opportunity to ride the high wave of innovation and systems-based management underpinned by the work of Dr. W. Edwards Deming. Instead, they decided to pummel themselves into irrelevance by hiding behind nonsensical ISO-like quality assurance schemes and surrendering to the statistical hallucinations of six-sigma and Lean.

Their major sin was to think of “continuous improvement,” the mantra of Quality Management, as just a set of techniques with no direct connection to business results besides some cost savings.

People involved in leading the development and implementation of digital technologies have a choice: follow the path of the Quality professionals and walk into their own professional cage, or arm themselves with cause-and-effect reasoning and stand up to the challenge.

Software technologies have always been considered, at best, a necessary evil. No Top Executive of a non-tech organization comes from ­Technology. This is changing rapidly. Digital Strategists now sit at the C-suite table and are called upon to deliver the redesign of their organizations in light of what they can enable business-wise. Digital must no longer ­simply “support” the business; it must “be” the business because in today’s world, every company is a technology company.

The inherent challenge for digital professionals is to “see” the business and be breadwinners. They have a unique opportunity to impact the information and material flows that shape the success of their organization. Digital technology has the possibility to overcome the fallacies of budget-driven, functional silos and to alleviate the predicaments generated by a hierarchical view of the organization.

The starting point to becoming a real leader in the digital age is ­understanding how to manage the flow in an organizational system. As I mentioned in the previous chapter, we need to introduce some new language for a new solution. To manage flow, leaders must grasp some fundamentals: understanding variation and why we need a constraint.

The Need for a Constraint

Although the word constraint may sound negative, in practice it is not. It turns out that we actually need a constraint and that is very useful. This is true for two precise reasons:

1. We need a constraint to facilitate the synchronization of the different processes in the system (the constraint is the element that dictates the pace at which value is generated).
2. We need a constraint because an “unbalanced” system is more solid and it is more easily managed when the impact of variation is taken into account.

In light of the above, we can in fact choose the constraint; we decide where it is possible to conveniently (strategically) position the constraint in order to manage it.

Let’s see in more detail why an unbalanced system is simpler to manage than a balanced one. The simplest way to exemplify the power that stems from focusing only on one constraint in an organizational system is a production flow. However, this is only one example. The production line represented in the following example could just as well represent all the processes of an organization, or even the various elements in an entire supply chain. It would be a huge disservice to the Theory of Constraints to consider it as something that is restricted to manufacturing and logistics. Indeed, the following line of reasoning can and has been applied for decades to a myriad of business environments.

A balanced plant or system is a production flow where the process stages (for the sake of simplicity, machines) that generate the finished products all have the same nominal production capacity. The goal of a balanced system (Figure 5.1) is to match the capacity of the various stages so that there is no excess capacity in any part of the system.

Figure 5.1 A balanced system

Unfortunately, this scenario is quite unrealistic because we are overlooking the impact of variation on the different stages of the chain. At each stage, the amount of production will be affected by variation resulting in the impossibility of the system being “balanced.”

Let’s look at a simple example.

Let’s imagine we have a market demand of 10 units of our product and a production process with 5 stages, where every stage has the same production average and the same amount of variation as the others. On average, every step of the process can produce, and make available to the next step, 10 semifinished units per hour, with a performance that varies within a range from 6 to 14 units.

Let’s simulate the first production cycle.

The first stage passes on to the second stage something between 6 and 14 units (let’s say 11), the second stage will be able to produce any result between 6 and 14 (let’s say 14); similarly, the third stage will generate 8, the fourth stage 12, and the last stage 7.

Given this combination of output, which is perfectly compatible with the machines chosen (by average and variation), how much finished product really comes out at the end of the first production cycle (Figure 5.2)?

Figure 5.2 First production cycle

The first stage will deliver 11 units to the second stage. The second stage is able to process all 11 units. The third stage will receive the 11 units, but, due to variation, it will only process 8 of them and pass them on to the fourth stage. The fourth stage has an instantaneous capacity to process 12 units but, as it only received 8, it will only process those and pass them on to the next stage. The last stage has an instantaneous capacity to process only 7 units; consequently, 1 unit will get left behind. The final result is that, after an input of 11 units, only 7 are produced (Figure 5.3).

It is clear that 4 units are somehow “lost” inside the production chain as work in progress (Figure 5.4). In this case, 3 units are stuck between machines 2 and 3 and the other unit between machines 4 and 5:

Figure 5.3 Result of first production cycle

Figure 5.4 Units “lost” inside production chain

Figure 5.5 Second production cycle

To summarize, even if the market demand was 10 units, equivalent to our “average capacity,” we would not be able to satisfy demand because interdependencies and variation make it impossible to “balance” the production.

Under the initial assumptions, let’s see what happens as time goes by. Because variation over time is the same (the range is from 6 to 14), and all the stages have the same average value, the second production cycle will be able to use the semifinished units that are still in the system.

Let’s look at a simulation of the second cycle:

The first stage will deliver 9 units to the second stage (Figure 5.5). The second stage will be able to process all 9 units. The third stage will receive the 9 units plus the 3 semifinished already present in the system and will process all of them and pass them on to the fourth stage. The fourth stage is able to process 9 units, leaving behind 3 semifinished units. The last stage will use these 9 units, plus the 1 already present as WIP, ending the cycle with 10 units produced and 3 left in front of the penultimate processing stage.

As the system is made up of a series of interdependent bottlenecks, in every single cycle, the instantaneous production capacity of the entire production flow will be equal to the capacity of the stage that has produced the minimum capacity. Moreover, as every machine has a capacity that varies above and below 10 units per cycle, it is almost certain that at least one of the machines has an instantaneous capacity that is less than the average value. Therefore, the real capacity of the entire system will always be something less than 10 units per cycle, at least as long as the stages are interdependent. We are in a paradoxical situation where we produce WIP without managing to satisfy market demand (10 units/cycle).

As production continues (the number of cycles increases), there is an increase in the semifinished material that accumulates between one production stage and another; this is due to the effect of variation in each stage and goes on until the production stages become decoupled. Now no stage is forced to remain inactive and each stage can operate independently from the previous one, with a global capacity that is close to its own average value.

The price of decoupling is an increase in WIP (Inventory); this means keeping money frozen within the system (Figure 5.6).

Figure 5.6 Increase in semifinished material

With a balanced process, we can deliver 100 percent of market demand only if these two conditions are met at the same time:

• We accept to size the process so that the market demand for every cycle is considerably less than the average value of the production capacity for each stage.
• We wait for the condition of decoupling of the production stages to happen.

These conditions allow us to deliver 100 percent of the demand, ­without generating too much WIP, only by having a much greater production capacity in every stage compared to actual demand. In our example, the system would be able to deliver 100 percent of market demand if this demand was 8–9 units/cycle (considerably lower than the average of 10 units/cycle).

A different approach is to increase the capacity of some of the stages of the production process so that there is always only one limiting stage (constraint) in the flow of material (Figure 5.7).

Figure 5.7 The constraint in the flow of material

This choice allows us to eliminate most of the WIP, as long as the capacity of the constraint (and therefore of the other stages) is sized on the basis of market demand (market demand must be equal to the average production capacity of the constraint). Also, the flow of the material has to be managed around the constraint; this strategy also foresees an oversizing of production capacity, but it also has a number of advantages.

Instead of balancing the system, and then trying to improve it by balancing it and reducing variation in all the production stages, we apply the Theory of Constraints (TOC) solution. After identifying (or choosing) the constraint, we manage the system around the constraint itself. With this solution, we “unbalance” the system because it is simpler and cheaper to manage.

In an unbalanced system, everything revolves around the constraint phase and there are many advantages to this. A detailed plan is made for this phase only. This schedule allows us to manage the whole production system. Also, reducing (global) variation in an unbalanced system means concentrating on and investing in the constraint phase only, not in every single part of the production process. Consequently, increasing the productivity and improving the performance of an unbalanced plant is considerably cheaper and less wasteful in terms of time and energy.

The algorithm that embodies this approach to the synchronized scheduling of finite capacity, which has been largely used by a multiplicity of industries all around the world for more than 30 years, is called “Drum–Buffer–Rope” (DBR); it is a three-step process:

1. We identify/choose the constraint.
2. We exploit the constraint.
3. We subordinate the system to the constraint.

To be more specific, first we identify the constraint in the cycle, and then we decide to manage it effectively. Managing the constraint effectively means managing it so it never stops: in other words, it must never be “starved” of input. In TOC, this is called exploiting the constraint.

For this exploitation to happen, every other stage/process in the system has to subordinate. This means that there must be enough production capacity upstream to always produce semi-finished units to feed the constraint continuously and enough production capacity downstream to guarantee that what the constraint produces can be processed. Statistically speaking, this means that the “ranges of variation” upstream and downstream of the constraint phase must have a lower limit of oscillation that is greater than the upper limit of the constraint phase.

As far as WIP is concerned, whereas in the case of the balanced system it must be available in front of all of the production stages so that uncoupling can occur, in the case of the system with a constraint, WIP must only accumulate in front of the constraint stage to keep it continuously fed. What is more, if the material is released to the whole plant at the pace at which the constraint can process it, the amount of WIP needed to achieve optimal performance can also be kept to the minimum.

Protecting the Constraint: The “Buffer”

The algorithm we have just described has the aim of maximizing the performance of the system and it is totally dependent on the correct and continuous functioning of the constraint. Managing the constraint effectively implies that it must never “starve” and that is why we must have some kind of protection mechanism in place. In TOC, the protection mechanism is called “buffer management.” In a system that is unbalanced around the constraint and subordinated to it, the buffer protects the constraint from variation. The unit of measurement of the buffer is time ­(Figure 5.8).

Figure 5.8 Buffer time

Once we know the amount of variation the constraint is affected by (both in terms of input it receives and its processing time), we can define how much time in advance the material has to be ready in front of it. The final, and most important, step is to manage the buffer. Essentially, we have to have an operational procedure that supports our decision-making process; in order to be sure that the constraint works all the time, we have to guarantee that the buffer is never reduced to zero.

Since we believe that statistical predictability is the foundation of management, we must monitor the consumption of the buffer statistically and act accordingly. If the consumption is “statistically stable,” we do not intervene; vice versa, if the consumption is statistically unstable, then we must take action.

Summary of Chapter 5

• Leadership is about empowering people to pursue a shared vision. A leader must therefore be capable of designing and impressing upon his/her people (a) the image of a desirable future and (b) why and how it is possible to achieve it.
• A leader is someone who
  1. owns a Theory, a set of well-tested assumptions within a well-defined realm of validity. Without Theory, Management—that activity we do to achieve our vision—becomes a “whack-a-mole” game and finger pointing becomes the rule.
  2. is capable of communicating effectively inside and outside the organization why the Theory will produce the desired results; in other words, the role of the leader is to create predictability of outcome for the efforts he/she requires from their people as well as constancy of purpose among them.
  3. is selfless and relentless in elevating everyone’s abilities through continuous teaching and mentoring.
• The starting point to become a real leader in the ­digital age is understanding how to manage the flow in an ­organizational system.
• Some management fundamentals for managing flow are understanding variation and why we need a constraint.
• Variation affects all human processes and it needs to be ­understood and managed.
• Statistical predictability is the foundation of management.
• A constraint is a positive thing because we can use it to facilitate the synchronization of the different processes in the system (the constraint is the element that dictates the pace at which value is generated).
• We need a constraint because an “unbalanced” system is more solid and it is more easily managed when the impact of ­variation is taken into account.
• It makes no sense to run a “balanced” plant because at each stage, the amount of production will be affected by variation, resulting in the impossibility of the system being “balanced.”
• In a system that is unbalanced around the constraint and subordinated to it, we use a buffer to protect the constraint from variation. The unit of measurement of the buffer is time.