Contents

---

Introduction

As discussed in our other white papers^[1], a successful corporate initiative must be grounded in a systems approach—one that identifies a high-leverage intervention point, where a small change can produce disproportionate results.

However, adopting a systems approach and identifying the right leverage point is only the beginning.

Implementation brings its own challenges.

This white paper offers a high-level overview of the key factors that must be managed during implementation. Fittingly for its subject, this paper is more concise than others. It does not attempt a full theoretical analysis; nor (being a methodology white paper) does it attempt to provide a roadmap for application. Both of these issues (the theory and the bridge to application) are briefly touched upon in the final sections.

---

The Map Is Not the Territory

In analyzing an organization as a system and identifying a leverage point, a model—explicit or implicit—is always created. Such models fail in two predictable ways:

• The model, by virtue of being a model, is always a simplification and therefore incorrect. The map is not the territory. What we will seek is a model that is good enough.
• Even if a perfect model were possible, it would still not fully capture reality. (This is addressed in more depth below.)

Flawed Maps and the Scientific Method

To address the inherent limitations of any model, we apply the scientific method combined with rapid iteration.

The scientific method, simplified, begins with a hypothesis and seeks to falsify it through experimentation. Note the key principles:

1. Learn through experiments
2. Seek to disprove—not confirm—the hypothesis.

In a corporate setting, we call such an experiment a pilot. Instead of jumping straight to implementation, we begin with a pilot that tests the hypothesis.

Entropy

However, flawed models are not the only issue. Even a perfect understanding of the system—if such a thing were possible—would not eliminate entropy—that is, friction, resistance, irreducible unpredictability.^[2] These must be managed.

(This parallels the foundational distinction, in philosophy, between form and matter: form is structure, matter is what is formed into that structure. See the Theory section.)

A full discussion of entropy management lies beyond the scope of this paper. For more, refer to our dedicated white paper on this topic.

---

Mini Case Study: of Entropy and Maps

The Situation

Company X was unable to serve all potential customers, severely reducing its revenue. In an industry where a wait of hours was too long, its wait times stretched to two weeks.

The Analysis
Our analysis revealed that a certain step of its fulfillment process was a bottleneck. Theoretically, this step should have been fully automated. In practice, due to special cases, it ended up having to be handled manually.

Revenue worth a third of the company’s current revenue stream was being lost because of this.

The pilot
Accordingly, an implementation pilot focused on fixing this step for the company’s main product line (~60% of its revenue), by fully automating it—a task for the IT department.

Entropy strikes
However, the pilot kept getting delayed. All sorts of matters kept cropping up that required assistance from the IT department. Thus, the latter did not have time to work on the pilot project. In the end, the CEO realized that sacrifices would have to be made. In coordination with the IT department, we set an IT development plan that was aligned with the strategic priorities of the organization and which prioritized the pilot. The CEO protected the IT department from pressure to pause the plan in order to work on other things. Even when requests came from Heads of Department, the CEO and the Head of IT held firm to the plan.

The map was not the territory
Once the plan was set, a fix was indeed rolled out that automated the majority of cases. However, the impact on revenue was unexpectedly quite modest. The following was learned from the pilot:

• The fulfillment process was full of special cases. Thus, a fix that was expected to result in a one-third increase on 60% of the revenue (i.e. a 20% increase overall) turned out to only apply to half of transactions, i.e. 10% overall. Worse: even among those half, a further half constituted special cases, thus reducing the impact to a meager 5% overall increase in revenue
• The complexity of these special cases was being managed by the IT department, who were doing their best with limited resources to keep the company’s proprietary software abreast of the ever-changing special cases. This had been the unrecognized case all along.
• The true bottleneck was thus the capacity of the IT department. Indeed, this was observed in multiple ways:
  • The roll-out of the pilot was slowed down by lack of capacity of the IT department
  • The handling of special cases was not happening as fast as special cases cropped up, also due to insufficient capacity of the IT department
  • Almost all the processes of the company were being slowed down in some way or other due to the IT part having issues, and the IT department being very much behind in resolving these issues
  • The outcome of the pilot showed that fixing a process bottleneck required sophisticated exception handling, which in turn required the IT department to have sufficient capacity

Course correction
The correct root cause of the organization’s issues having now been identified, we moved quickly to address it by:

• Rolling out AI solutions to increase the productivity of existing staff by >20%.
• Hiring more IT staff

This organization was not at all agile. It was in the red when the initiative began, and its inability to rapidly iterate pilots and then rapidly apply the lessons learned painfully slowed implementation. During all this time, it continued to lose money.

Indeed, implementing pilots and addressing entropy both require agility. We will discuss the need for agility below.

---

The Leadership Team Is Flawed

The leadership and implementation teams—being human—are flawed.

Leadership is a crucible that reveals a person’s weaknesses. As the saying goes: If you want to know what someone is really like, ask their subordinates.

---

Infobox: How leadership flaws are elicited by circumstances

One pathway through which the leadership’s flaws will come out in implementation is via dilemmas. In a context of wicked and imponderable problems (see white paper 2), implementation surfaces a myriad of dilemmas, large and small. These dilemmas interact with the flaws of leadership. For example, entropy creates friction and opportunities, both of which upset the original plan. In response, the leader must choose when to change course, and when to hold the course.

Leaders, being flawed, often choose based on their flaws. A scattered leader may allow such distractions to derail the plan. A rigid, overly procedural leader may ignore emerging realities and be blindsided.

---

Taking a step back, the situation can be fairly described as follows:

• A flawed team…
• …is attempting an initiative that is more likely to fail than succeed

This is not a promising combination.

One of the most critical success factors, therefore, is honest self-assessment. Leadership must take a hard look in the mirror, examining how their own flaws affect the initiative.

In our experience, leadership flaws—directly or indirectly—are the root cause of most implementation failures.

Assessing the impact of the leadership’s flaws must be done case by case, though certain themes are recurring:

• Failure to give full backing
• Distraction mid-initiative
• Hypocrisy: failing to walk the talk
• Over-focus on strategy, neglecting details
• Over-focus on details, missing the big picture

---

Mini Case Study: a Tale of Two CEOs

Example 1
Company X was executing an ambitious growth plan. The Chairman, however, liked to micro-manage. He maintained informal ties with senior managers, from whom he heard a biased perspective on what was going on in the company. As a result, he would meddle in operations without informing the CEO. Initially minor, the issue snowballed.

We devised a plan, in coordination with the CEO:

• Move the Chairman’s contacts to another department, not involved in the main initiative
• Thus, he would receive information on this secondary department, and his impulsive interventions would be channeled into this secondary department

Unfortunately, the CEO resigned before this could be implemented.

Example 2
Company Y’s CEO was implementing a change of strategy for his organization. Being self-aware, he realized that he was a big-picture thinker who disliked managing the details. He therefore hired a VP as second in command who was good at managing details. Unfortunately, as time went on, the relationship between them became more and more stressed. The VP eventually left.

Commentary
The examples in this section are deliberately failures. The issue of leadership self-management is particularly hard. It reflects a broader governance dilemma: quis custodiet ipsos custodes—who will guard the guards themselves?)

Leaders are responsible not only for initiatives, but for managing the flaws of their teams. Who, then, will manage the leaders’ own flaws?

For more on this topic, see our white paper on Managing leadership flaws.

---

The Enterprise Is a System

The enterprise is a system. This insight must guide not only design but also implementation.

While this idea is developed in depth elsewhere (especially white paper 2), we remind the reader of key system-derived implementation challenges:

1. Homeostasis: Systems resist change.
2. Subcomponents with divergent goals: Departments have competing goals. This gives rise to office politics.
3. People are the most important parts of the system: In social systems, people are the most important parts.

Traditional approaches recognize these aspects, but focus on change management, communication, and process design. We suggest instead (as detailed in the relevant white paper):

• Use system features (e.g., homeostasis, bottlenecks, fault lines) to anticipate how the system will react.
• Pre-empt and respond. This requires being more agile i.e. having a faster observationa and decision-making cycle^[3]
• Internalize the changes through culture, mentoring, and handover.

---

Mini Case Study: Budget Wars

The situation

• Company X was in the midst of switching from a top-heavy to an agile organizational structure.
• The next step in the transformation involved the purchase and roll-out of certain IT tools
• We anticipated:
  • Homeostasis: the system would resist this next steo
  • Fault lines: the resistance would occur at the interface between two departments, and likely would be phrased in terms of budgets. (“Who will pay?” is the easiest topic for departments to disagree about)

Of course, the budget disagreement would be a pretext, the real issue being resistance to change. From a purely budgetary perspective, the amounts involved would be so inconsequential as to not be worth fighting over.

How we prepared
Having anticipated this issue, we prepared for it as follows:

• We planned to trigger this issue very early, while the pilot was still on-going
• We scheduled certain events (and specifically preparations for the next budget allocation cycle) at times that suited our planned response

How we managed the conflict
Rather than confront the resistance head-on, we let them run with their objections. We gave them rope.

We allowed the budget stalemate to play out in full view. No escalation. No public pushback. We let them think they were holding the line—while knowing that the timing was on our side.

(Remember that we could afford to do this without in any way delaying the project. After all, we had triggered this issue early, at the time of our choosing, well before the end of the pilot testing phase.)

Soon, as we had planned would happen, the time came when they themselves needed cooperation from other parties. Preparations for the next fiscal cycle began. (Remember that we had scheduled this at a time that would suit us.) Budget planning meetings were convened. And suddenly, both departments realized that the unresolved tooling issue was a strategic liability. With leadership watching closely, they were being seen as blockers to progress—and it was starting to affect their standing in broader budget negotiations.

Privately, the narrative had begun to shift: “Why are we still discussing basic tooling?”

Now the cost of continued resistance began to weigh heavily. Recall that the original issue was a pretext, and from a budget perspective was not even worth fighting for.

At this point, in order to make sure that no one lost face, we stepped in “on their side” and “in their defence” against the criticism from other department heads. We arranged with them a resolution that allowed them to back down without appearing to have backed down.

---

Surprises Await

Reality keeps moving during implementation. Surprises—internal and external, predictable and unexpected—are inevitable.

If the initiative begins while the firm has surplus resources, these shocks can be absorbed. Hence the importance of beginning initiatives while the firm still has excess resources, and not in desperation.^[4]

If the firm does not have surplus resources, solutions must be found on a case-by-case basis.

Interestingly, because the firm is a system embedded in larger systems, it is usually possible to find a solution—this is provably true.^[5]

Mini Case Study: Beaten by the Competition

Company X had successfully resolved a key bottleneck, freeing up 20% of Operations staff. This had only modest bottom-line impact:

• Freeing up manpower, in itself, does not impact net income
• However, there was some net income improvement resulting from how the improved process generated less waste

Company X could have parlayed this improvement into cost savings via redundancies; however, since the staff in question was qualified and good, they preferred to use them to develop a new product.

Unfortunately, while they were doing this, a competitor anticipated them and released their own, even more advanced product. Company X’s development efforts were obsolete even before they had resulted in anything tangible.

What a catastrophe!—or not? By fixing their original bottleneck, Company X had released tremendous potential that they could flexibly allocate in whatsoever direction they wished. (This is typical of corporate initiatives that are based on a system view of the organization.) Thus, they simply shifted their efforts to the next generation product. They also took advantage of the competitor’s premature launch to learn what not to do and save time and effort.

Their competitor did not have 20% slack capacity in the Operations Department, and was not able to keep up with them.

---

Opportunities or Problems?

It is natural to see the above-described factors as problems.

It is a step in the right direction to regard them not just as problems, but as potential problems or opportunities.

But in truth, both perspectives are wrong and harmful.

These factors simply are.

We must see them without value judgment. The labeling of something as a problem or opportunity comes from the beholder, and blinds the beholder.

Although this sounds philosophical, it has practical implications. Consider two contrasting case studies.

Mini Case Study: Blinded

Company X had a new leadership team that correctly recognized its operations were unsustainable. They conducted sound analysis, acknowledged the company as a system, and launched a solid transformation initiative.

However, they approached challenges—like homeostasis and resistance to change—as problems to be defeated. They put great effort into stakeholder engagement and change management, but entrenched interests still blocked implementation. A few years later, the company became insolvent.

Company Y, facing an even worse situation, took a different approach. Though they also faced entrenched interests, internal disorder, and external shocks, their leadership dared to see these, not as problems, but as neutral facts.

They did not try to fight or solve every problem. On the contrary, they let the pot stew. The unaddressed issues generated more and more disruption. As a result, homeostasis kicked in: the entrenched interests felt more and more threatened by the disruption, and focused their efforts on managing the entropy-generated issues and external pressures and surprises. As a result, they were unable to effectively resist the implementation. The change initiative rolled out smoothly.

The key difference? Company X viewed challenges through a value-laden lens, as problems. Company Y viewed them neutrally—as facts to be managed. This shift in perspective allowed better, more focused execution.

---

Agility

Agility is required to:

• Implement pilots^[6]
• Anticipate and respond to entropy
• Anticipate and respond surprises
• Manage the above instead of reacting

Without agility, most of the strategies in this paper become inapplicable.

Even the principle outlined in the previous section, of viewing these facts as facts rather than problems becomes impossible to apply. The brute facts, unmanaged, indeed turn into unstoppable problems.

We define agility as a fast observation and decision-making cycle—a compressed OODA loop (Observe–Orient–Decide–Act). See our white paper on agility for more.

---

Mini Case Study: Chinese Automakers

As of this writing, Chinese automakers like BYD are outmaneuvering global rivals (GM, VW, Tesla) by slashing their car development time by more than half, and cutting the time-on-market by even more: a Chinese-make car model is on the market for ~1.6 years on average, compared to an average of 5.4 years for foreign brands.

They are compressing their OODA loop (Observe-Orient-Decide-Act), turning speed into a decisive advantage:

1. Observe: They rapidly gather market feedback (e.g., Chery adjusted its SUV for European roads in weeks after identifying ride-quality issues).
2. Orient: Flat hierarchies and a “fail fast” culture let engineers bypass bureaucracy—BYD’s CEO approves designs in days, not months.
3. Decide: AI and simulations replace slow physical prototyping. Zeekr tests parts digitally in hours, cutting development time.
4. Act: Vertical integration (BYD makes 75% of parts in-house) and massive workforces enable 24/7 execution, launching models in as little as 18 months (vs. 5+ years for Western brands).

Furthermore, whereas Western car makers typically iterate the OODA loop once per model, Chinese automakers iterate it multiple times:

• During the development process, they iterate multiple times. Even when the car is close to production, they have no hesitation in making fundamental design changes that would scare their Western counterparts.
• Rather than seek a perfect product, they allow the model onto the market earlier than Western counterparts, collecting feedback and rapidly iterating out the wrinkles.

Result: Chinese EVs like BYD’s $7,789 Seagull hit markets faster and cheaper, while legacy automakers struggle to adapt.^[7]

Interestingly, the applicability of such a strategy to the auto industry was raised by Richards (2004) decades ago.[^1]

---

The theory

In this section (which can be skipped), we briefly review the theory underpinning the above.

Partitioning the problem

Consider the overall system composed of:

• The leadership and implementation team that is conducting the corporate initative
• The enterprise (except for the leadership and implementation team when they are in that role)
• The environment

The above forms a partition (i.e. a mutually exclusive and collectively exhaustive framework).

As is, the partition captures all the aspects of implementation, but does not yet capture the distinction between the consequences of flawed maps and of entropy. In order to capture this, we will further partition the enterprise into form and matter.

Form Versus Matter

The distinction between form and matter is foundational to philosophy. As mentioned, form is structure; matter is what occupies that structure.

When we understand something, we are usually grasping its form. Even if we had a perfect model—a perfect map—we’d still fall short of grasping its matter. That irreducible element escapes full understanding.

For instance, even if we knew every detail of an atom, some aspect of its behavior would remain beyond comprehension and irreducibly unpredictable. This is confirmed by modern physics, and it is impressive that ancient philosophers such as Thomas Aquinas and Aristotle should have been able to demonstrate this result.^[8]

Applied to the enterprise: even a full understanding of the system, were such a thing possible, would not eliminate entropy—that is, friction, resistance, irreducible unpredictability.^[9] This must be managed.

A full treatment of this subject is beyond this white paper. See our white paper on the Principles for implementing corporate initiatives.

Consequences for implementation

The above results in the following partition:

• The leadership and implementation team that is conducting the corporate initative
• The enterprise (except for the leadership and implementation team when they are in that role)
  • Form
  • Matter
• The environment

Consequently, an effective implementation must address each of the above:

Valence

When derived in this manner, the different headings of our implementation framework appear clearly for what they are: they are not problems to be solved (although they appear to be). They are simply components of the situation. As such, aside from what the viewer projects on them, they are neither positive nor negative. They have no valence; they simply are.

Coloring them with valence introduces information (the valence) that is false into the formulation of the problem. This in turn reduces our ability to formulate correct solutions.

---

Beyond this White Paper: Application

The curse of an author excited about his subject is that he eventually has to stop. This paper being a methodology white paper, it does not attempt to provide a roadmap for application.

However, the below examples can provide a flavor of some of the implementation practicalities that would have been treated, if this were a full-length book.

Scale changes everything.
Example: a Fortune 500 company has an additional tool in its toolbox: selling the problem-business. It also has an additional handicap: greater remove from the operational level.
Why: the management a Fortune 500 company is, in many ways, more akin to portfolio and cash management than it is to the management of a mid-market company. It is a very different world. This increased remove from the operations makes it harder for a Fortune 500 to buck the effects of homeostasis.

Industry changes everything.
Example: a commercial bank faces additional challenges relative to many sectors, such as the inertia of its huge balance sheet, the friction of regulations, and sensitivity to a whole range of possible external shocks which do not impact non-financial companies as much.
Why: the core business of a commercial bank is maturity transformation: get shorter-term deposits from a large number of people, and repackage into longer-term loans. This is risky: the bank has to be able to fulfill short term obligations, while its assets are mostly tied up long term. The riskiness has resulted in a lot of regulation. As a result, the banking business is all about management of risk (which, among other things, requires a big enough balance sheet to absorb statistical anomalies), and management of regulations. Hence, while implementing a business transformation, a bank is more likely to hit a regulatory wall, or to be hit by an external shock such as rising interest rates, or by a regulatory-induced external shock such as an interest rate movement that decimates the (paper) accounting value of some of its assets.

Geography changes everything. A multinational faces additional factors when implementing, ranging from tax considerations to regulatory considerations to the increased complexity of implementing, to fake diversification.
Why: most of this is self-evident, except for the “fake diversification” element. “Fake diversification” arises when the corporation believes that it has diversified (and therefore reduced) its risk by being present in multiple countries, when in fact it has increased the rislk. This mainly arises in two main ways:

• Regulations nullifying diversification. For instance, in the case of multinational banking group, each national regulated entity is ring-fenced (excess capital / liquidity cannot be used to offset lack thereof in other countries) and regulations often limit intra-group support.
• Misunderstanding risk: businusspeople have generally learned that diversification reduces risk. This is false in the real world.
  • This false idea arose from outdated models of financial risk that did not take into account time. Once time is taken into account, diversification turns out to be potentially very risky.
  • Intuitively, this can easily be understood. Imagine putting all your eggs in two baskets rather than one. This definitely reduces your risk of a single catastrophic loss. But imagine if you need to manage the allocation of eggs to baskets long term. Along the way, good things (e.g. a generous hen adding eggs to a basket) and bad things (e.g. a truck driving over a basket) will happen to the baskets, many times. It should be intuitively obvious that in this new scenario the superiority of spreading your eggs is no longer obvious.

Proper planning matters. This includes:

• Measurement of success
• Planning of resource needs – what does this actually cost in time, people, budget?
  Example: there exist companies that do an annual budget, but never check the budget against what actually ended up happening. Need we asy more?

Hopefully, the examples provided, if not worth a million words, have at least been insightful and useful. However, in order to apply these concepts in practice, you will need a practitionner. Indeed, when sick, do not read a medical treatise. Call the doctor.

---

References

Aristotle. Physics, multiple editions.

Boyd, John, R. (28 June 1995). “The Essence of Winning and Losing”. danford.net. A supposed five-slide set by Boyd.

Aquinas, T. De Ente et Essentia (On Being and Essence), multiple editions.

Richards, C. (2004). Certain to Win: The Strategy of John Boyd, Applied to Business. Xlibris US.

---

Notes

---

1. Notably White Paper 2 (link) . ↩︎
2. In physics, entropy is the amount of information that we cannot know about a system, and is calculated as the Shannon information of all the possible states of the system, given its known macrostate, times a constant (the Boltzmann constant). In this white paper, we use the term in the same sense. Indeed, the irreducible unpredictability mentionned in the text is precisely the amount of information that we cannot know about the system. We further highlight that there is a component of information that is not only unknown, but unknowable even in principle. ↩︎
3. C.f. Boyd (1995) and Richards (2004) ↩︎
4. Recall from White Paper 2 (link) that this was one of the insights identified by BCG. ↩︎
5. The demonstration is beyond the scope of this paper; see the relevant white paper. ↩︎
6. Indeed, since the map is not the territory, several iterations of the pilot project are almost always required. Therefore, the organization must be agile enough to swiftly conduct and learn from these iterations until the intervention can be rolled out at scale with confidence. This is why rapid iterations are a key element of the solution. ↩︎
7. Carey, N., Shirouzu, N., (2025, July 3). How China’s new auto giants left GM, VW and Tesla in the dust. Reuters. https://www.reuters.com/investigations/how-chinas-new-auto-giants-left-gm-vw-tesla-dust-2025-07-03/ ↩︎
8. C.f. relevant references here. ↩︎
9. See above note 2 on the definition of entropy. ↩︎