|
Blog
Why Most Decisions Die in Translation, and the A3 Method That Prevents It
The Failure Point Most Leaders Miss
Most decisions do not fail because they are wrong. They fail because they do not survive translation.
A leadership team aligns around a strategy. The logic is sound. The direction is clear. But as that decision moves across functions, layers, and incentives, it begins to degrade. Priorities blur. Assumptions shift. Execution fragments.
What started as a coherent decision becomes a series of interpretations.
This is not a failure of strategy. It is a failure of clarity: the inability to preserve a decision’s logic as it moves from conception to execution.
In practice, this breakdown is both common and costly. Sales teams communicate different versions of the same value proposition. Functional leaders pursue competing priorities while believing they are aligned. Capital narratives shift depending on the audience, eroding credibility with investors.
Each issue appears isolated. The underlying failure is not.
A3 Is Not a Document. It Is a Discipline.
Toyota developed a mechanism designed to address this exact failure point. Known as A3, it is commonly described as a one-page report. That description is directionally accurate but fundamentally incomplete.
A3 is not a document. It is a discipline that forces clarity before a decision ever leaves the room.
At its core, A3 imposes a simple constraint: the entire problem, analysis, decision, and plan must fit on a single sheet of paper. This constraint is not about brevity for its own sake. It is about forcing precision. Leaders are required to define the problem in concrete terms, ground their understanding in observable conditions, identify root causes rather than symptoms, and articulate countermeasures that logically connect to those causes.
The sequence matters. The logic must hold. There is no space for ambiguity or excess.
If you cannot explain the decision on one page, you do not yet understand it well enough to execute it.
Sam Palazzolo
Why Decisions Break Down in Practice
Most organizations do not lack intelligence or effort. They lack a shared, disciplined method for converting ideas into clear, transferable logic.
As a result, alignment becomes superficial. Teams may agree in conversation but do not operate from a common understanding. Each function fills in gaps independently, introducing variation at every handoff. Over time, these small deviations compound into material execution failure.
This pattern is particularly visible in high-stakes environments. In growth-stage companies, leadership teams often believe they are aligned on priorities, yet execution reveals competing interpretations. In capital markets, founders present narratives that shift across meetings, signaling a lack of underlying coherence. Investors and operators respond not to the stated strategy, but to the inconsistency behind it.
These are not communication issues in the conventional sense. They are failures of narrative integrity. The underlying logic of the business is not consistent enough to carry across audiences without distortion.
The Role of Constraint in Forcing Clarity
A3 addresses this problem by standardizing how thinking is structured and communicated.
A well-constructed A3 does not simply describe a decision. It makes the reasoning behind that decision explicit and testable. The problem is clearly defined. The current condition is grounded in data and direct observation. Root causes are identified through structured analysis. Target outcomes are specified. Countermeasures are directly linked to those causes. An execution plan assigns ownership and timing.
Because all of this is captured in a single, coherent view, the decision becomes portable. It can move across teams and levels without being reinterpreted at each step. The integrity of the logic holds.
Constraint is what enables this. By limiting space, A3 eliminates the ability to hide behind complexity or defer clarity. It forces leaders to resolve ambiguity at the point of decision rather than allowing it to surface during execution.
Most execution failures are not operational. They are failures of clarity that compound over time.
Sam Palazzolo
PDCA Is the Engine Inside A3
A3 does not produce clarity by accident. It produces clarity because PDCA is built into its structure.
Plan, Do, Check, Act is the thinking sequence that governs how a well-constructed A3 moves from left to right. The left side of the page is the Plan phase: problem definition, current condition grounded in direct observation, root cause analysis, target condition, and proposed countermeasures. This is where the discipline is most demanding, and where most organizations cut corners by moving to action before the thinking is complete.
Do is the execution plan: specific actions, clear ownership, defined timing.
Check is where most organizations fail. A3 requires a follow-up review: did the countermeasures produce the expected result? Without this step, execution becomes a one-way door. There is no mechanism to learn, no feedback loop to close.
Act is the final phase: if the countermeasures worked, standardize them. If they did not, return to the Plan phase with new information and a sharper hypothesis.
This is why A3 functions as a translation tool rather than simply a reporting format. PDCA enforces a complete thinking cycle. The single-page constraint makes that cycle visible and auditable. Every reader of the A3 can see exactly where the logic holds and where it does not. Gaps cannot be hidden behind slides, narrative, or volume.
Most organizations complete Plan and Do, then move on. A3 treats Check and Act as non-negotiable. That is where institutional learning lives, and it is where most execution disciplines fail to close the loop.
From Factory Floor to Boardroom
Although A3 originated within manufacturing, its relevance today extends well beyond the factory floor.
Across SaaS organizations scaling from $5 million to over $500 million in revenue, through private equity-backed transformations, and in capital raise processes, the pattern is consistent. Where A3 discipline is present, decisions move faster, alignment is more durable, and execution is more consistent. Where it is absent, organizations compensate with more meetings, more documentation, and more oversight. None of those measures address the root issue.
This is not about Lean as a philosophy. It is about clarity as a competitive advantage. In environments where speed and precision matter, the ability to maintain a consistent, defensible narrative across stakeholders is a differentiator.
Why Leaders Resist It
Despite its effectiveness, A3 is often resisted, particularly by experienced leaders.
The discipline removes the ability to rely on abstraction, to substitute volume for clarity, or to defer thinking to later stages. It exposes gaps in understanding quickly and publicly. For leaders accustomed to operating through discussion rather than structured reasoning, this can feel constraining.
That constraint is precisely the point. By forcing clarity early, A3 prevents misalignment from compounding later, when the cost of correction is significantly higher.
The Test of a Decision
Most organizations do not struggle to generate ideas. They struggle to preserve them.
A decision may be sound at the point of origin. But if its logic cannot survive movement across the organization, it will degrade into interpretation. And interpretation is where execution breaks down.
This is the problem A3 was designed to solve. By forcing clarity at the source, it ensures that decisions can move without losing their integrity.
Because in any organization of scale, the test of a decision is not whether it was right when it was made.
It is whether it survives translation.
Sam Palazzolo, Managing Director, Tip of the Spear Ventures | Founder, The Javelin Institute
12+ years ago I led a Tech (SaaS) startup to PE exit. Since, I have scaled 15+ organizations from $5M to $500M (2x $1B+).
References
- Sobek II, D. K., & Smalley, A. (2008). Toyota’s Secret: The A3 Report. MIT Sloan Management Review, 50(1), 17–24. https://sloanreview.mit.edu/article/toyotas-secret-the-a3-report/
- Sobek II, D. K., & Smalley, A. (2011). Understanding A3 Thinking: A Critical Component of Toyota’s PDCA Management System. Lean Enterprise Institute. https://www.lean.org/Bookstore/ProductDetails.cfm?SelectedProductId=349
- Lean Enterprise Institute. (n.d.). A3 Thinking and Problem Solving. https://www.lean.org/explore-lean/a3-thinking/
- Liker, J. K. (2004). The Toyota Way: 14 Management Principles from the World’s Greatest Manufacturer. McGraw-Hill.
- Sutton, R. I., & Rao, H. (2014). Scaling Up Excellence: Getting to More Without Settling for Less. Crown Business. (See also: “Why Great Innovations Fail to Scale,” Harvard Business Review.)
- Rumelt, R. P. (2011). Good Strategy/Bad Strategy: The Difference and Why It Matters. Crown Business.
What the NFL Draft Actually Teaches Leaders About Capital and Decisions
The 2026 NFL Draft opened in Pittsburgh not with consensus, but with conviction, disagreement, and immediate second-guessing. A non-obvious quarterback went first overall. Teams traded aggressively up the board and down. Franchises reached for need over value. One organization even stumbled operationally, contacting a player they would never have the opportunity to select. More than 300,000 fans watched in person. Millions more across broadcast platforms watched the chaos unfold in real time.
This is not a selection ceremony. It is a market.
And like any market, it exposes something leaders would prefer to ignore: even with shared information, aligned incentives, and billions at stake, decision quality varies widely. Not because the rules are unclear. Because decision-making is hard, and the draft does not let you pretend otherwise.
“The draft rewards teams that treat optionality as a strategic asset. Most companies treat it as indecision. These are not the same thing.”
Sam Palazzolo
Capital Allocation Is the Strategy
Strip away the spectacle and the draft is a capital allocation exercise. Each pick is a finite asset. Each trade is a reallocation of that asset across time horizons. Teams are not simply selecting players. They are constructing portfolios, balancing risk, upside, and time to return on a compressed, public timeline.
The organizations that consistently outperform are not the ones that “pick well” in isolation. They understand relative value. They know when to trade up and when to trade down, when to accumulate more shots on goal, and when to convert uncertainty into optionality. The discipline this requires is not natural. It has to be built.
Most businesses do not build it. Hiring decisions are treated as discrete events. Capital deployment is reactive rather than structural. The draft forces explicitness because every move carries a visible, immediate cost. In business, that cost is usually hidden. Hidden costs do not discipline organizations. They enable them to avoid the conversation altogether.
The Illusion of Consensus
Every team enters the draft with access to similar data. Game film, combine metrics, interviews, analytics. The inputs are broadly shared. The outputs are not.
Pittsburgh reinforced this gap. Teams looked at the same board and reached fundamentally different conclusions. Some prioritized positional value. Others prioritized immediate need. Some bet on upside. Others on certainty. This is not incompetence. It is interpretation, and that distinction matters.
Business leaders routinely assume that better data will produce alignment. It rarely does. Data reduces uncertainty. It does not eliminate judgment. Judgment is where teams diverge, where strategies separate, and where leaders either earn their seat at the table or reveal they were never ready for it. Strategy is not about having the right information. It is about making consequential decisions in the presence of incomplete information, competing interpretations, and real stakes.
“Data reduces uncertainty. It does not eliminate judgment. Judgment is where teams diverge, where strategies separate, and where leaders either earn their seat at the table or reveal they were never ready for it.”
Sam Palazzolo
Trades Matter More Than Picks
The most sophisticated teams in Pittsburgh were not just evaluating players. They were managing position.
Trades defined the early rounds. Some organizations moved up to secure specific targets. Others moved back to accumulate additional capital for future decisions. The real advantage was not in who they selected. It was in how they positioned themselves to select.
This is where the business analogy tends to break down. Most organizations focus relentlessly on outcomes: the hire, the acquisition, the product launch. They underinvest in option creation. Expanding the pipeline before committing. Structuring deals to preserve flexibility. Maintaining the capacity to act as new information emerges. The draft rewards teams that treat optionality as a strategic asset. Most companies treat it as indecision. These are not the same thing, and conflating them costs organizations more than any single bad hire ever will.
Execution Risk Never Goes Away
Even in a system engineered for precision, execution failures happen.
The Steelers’ misstep, engaging a player before they were on the clock, circulated quickly as a footnote and a punchline. It should be treated as a case study. Operational breakdowns occur at the worst possible moment, under the brightest lights, in the most consequential circumstances. This is not a football problem. Boardroom decisions, M&A processes, and go-to-market launches fail for the same reason. Not because the strategy was flawed, but because execution was not tight enough under pressure.
Strategy sets direction. Execution determines outcome. And execution degrades fastest precisely when the stakes are highest. Any leader who has not stress-tested their team’s operational discipline against a high-pressure scenario has not actually prepared for one.
Market Narratives vs. Structural Reality
Pre-draft coverage focused heavily on quarterbacks and skill players. The early rounds told a different story. Teams invested in offensive linemen and foundational positions, the least glamorous assets in the building.
This is a pattern that repeats. Markets reward visibility. Systems reward durability. In business, this shows up as chronic overinvestment in customer acquisition over retention, top-line growth over margin quality, product features over infrastructure. Organizations chase what generates attention and underinvest in what generates results.
The best franchises in professional football understand this and act accordingly. The best businesses do too, though fewer of them are willing to say it out loud when the board is asking about growth metrics.
The Draft Is Now a Media and Revenue Engine
The modern draft is not purely a football operation. It is a commercial platform. Hundreds of thousands of attendees. Multi-network broadcasts. Three days of continuous digital engagement. The event has become a content engine that drives fan acquisition, advertising revenue, and brand expansion.
This matters because it changes the conditions under which decisions are made. Choices are no longer internal and sequential. They are public, monetized, and subject to immediate narrative formation. Strategy is no longer just executed. It is performed, in real time, in front of an audience with an economic stake in the story.
Businesses face exactly this shift. Earnings calls, product launches, investor narratives, and public leadership moments are all environments where decision-making and storytelling have merged. The line between the two has blurred past the point of retrieval. The draft simply compresses that reality into three days and makes it impossible to ignore.
What the Draft Actually Teaches
The NFL Draft is typically framed as a lesson in talent evaluation. That is the least interesting part of the system.
What it actually represents is a compressed, high-stakes model of how organizations allocate capital, interpret information, manage risk, and execute under pressure. Some teams will emerge from Pittsburgh with classes that hold up. Others will not, and that outcome will be debated for years while the organizations involved continue making the same structural decisions.
The more immediate takeaway is this. In a system where information is widely available, incentives are aligned, and the stakes are impossible to ignore, performance still diverges. Not because the rules are unclear. Because decision quality is not a function of data access or stated commitment. It is a function of discipline, structural thinking, and the willingness to act on judgment when judgment is all you have.
The draft does not solve that problem for the teams that struggle with it. It exposes them.
That is the point worth paying attention to.
Sam Palazzolo is Managing Director of Tip of the Spear Ventures and Founder of The Javelin Institute. He works with VC, PE, and family office-backed companies to scale revenue, build leadership capacity, and execute at the intersection of growth and capital.
References
- Massey, C., & Thaler, R. (2013). The Loser’s Curse: Decision Making and Market Efficiency in the National Football League Draft. Management Science. Wharton School, University of Pennsylvania. https://faculty.wharton.upenn.edu/wp-content/uploads/2013/08/massey—thaler—losers-curse—management-science-july-2013.pdf
- Harvard Sports Analysis Collective. (2021). NFL Draft Report: Behavioral Bias and Draft Strategy. Harvard University. https://harvardsportsanalysis.org/wp-content/uploads/2021/04/HSAC-NFL-Draft-Report.html
- Anonymous. (2025). Optimizing NFL Draft Strategy: Trade Value, Risk, and Decision Modeling. arXiv. https://arxiv.org/abs/2504.07291
Why 90% of AI Initiatives Stall Before Scale
Most executives do not have an AI problem. They have a scaling problem.
According to McKinsey Global Survey data, while AI adoption is widespread, most organizations struggle to translate initiatives into measurable financial impact, with roughly 80% of companies failing to see meaningful bottom-line results and the vast majority of efforts remaining stuck in pilot phases.1,2 Other industry analyses push that figure further, suggesting that as many as 90% of AI efforts stall before enterprise-scale deployment.6 These are not fringe estimates. They are the consensus.
What makes this pattern so stubborn is that the failure point is almost never the technology. The models work. The demos impress. The pilots check out. The gap between a successful proof-of-concept and a functioning enterprise system is not a gap in model capability. It is a gap in system design, and most organizations are not asking the right questions when they try to cross it.
The Real Constraint: Architecture, Not Algorithms
The prevailing instinct in most organizations is to treat AI as a layer, a feature to be added on top of an existing operating model. Deploy a copilot here. Automate a fragment of a workflow there. Test an isolated use case and monitor the results. This approach generates compelling early data and frustrating long-term outcomes in roughly equal measure.
The reason is structural. AI systems that cannot orchestrate across workflows, access unified data, or operate within governed environments will not scale. They remain trapped in pilot mode regardless of how sophisticated the underlying models become. The constraint is not the reasoning capability sitting on top. It is the architecture sitting below.
This distinction matters because it changes where investment and attention should go. The organizations closing the gap between pilot and platform are not the ones with better models. They are the ones that redesigned how work gets done before they deployed AI into it.
AI does not fail because it is immature. It fails because it is deployed into systems that were never designed to support it.
Sam Palazzolo
The Shift to Agentic Architecture
The architecture that supports real scale is not single-use AI tools operating in isolation. It is agentic systems: networks of specialized AI agents that collaborate across tasks, data, and decision layers to execute end-to-end workflows.8 The shift from isolated tools to agentic platforms is not a product upgrade. It is a structural redesign, and it requires rethinking four dimensions simultaneously.
The first is orchestration. Single-agent deployments create incremental value at best. They automate a task, reduce a cycle time, or surface a recommendation. Multi-agent orchestration creates operating leverage, because it coordinates entire workflows rather than fragments of them. The value is not in any individual agent. It is in what happens when agents can hand off work, share context, and execute sequentially across a business process.
The second is data interoperability. Agents depend on shared context to function. A system in which data is fragmented across business units, tools, or legacy platforms does not just create inefficiency; it actively degrades AI performance, because agents operating on inconsistent or incomplete inputs produce inconsistent and incomplete outputs. A unified, accessible data layer is not a nice-to-have for agentic architecture. It is the substrate on which the entire system runs.
The third is modularity. Most organizations build AI capabilities the way they built enterprise software in the 1990s: each use case gets its own implementation, its own integrations, and its own dependencies. This approach creates technical debt at scale. Decoupling reasoning, memory, orchestration, and interfaces allows systems to evolve without being rebuilt from scratch. More importantly, it enables reuse, and reuse is what produces compounding returns rather than compounding costs.
The fourth is embedded governance. Organizations that bolt governance on after deployment discover, predictably, that the system resists it. Real-time monitoring, traceability, and policy enforcement are not features to be added after a system proves itself. They are design requirements that determine whether a system can be trusted at scale. Governance that arrives late rarely catches up.
Why Most AI Initiatives Stall
The failure pattern is consistent enough across industries that it deserves to be called a pattern rather than a series of unfortunate events.3,5 AI gets deployed into fragmented systems, where data remains siloed and inconsistent across the functions that need to use it. Workflows are not redesigned for automation; instead, AI gets layered onto processes built around human handoffs and manual coordination. Governance arrives after the fact, when the cost of retrofitting it is far higher than building it in would have been. And each new use case gets built from scratch, without reuse, so the organization accumulates a portfolio of disconnected experiments rather than a coherent capability.
The result is not technical failure. It is economic failure. The organization cannot scale what it has not standardized, and it cannot standardize what it has not architected. The pilots succeed. The P&L does not move.
Most AI pilots succeed technically. They fail operationally. That is a more expensive kind of failure.
Sam Palazzolo
From Pilot to Platform
Scaling AI requires a shift in orientation, from experimentation to system design. These are not incompatible; experimentation is necessary to generate learning. But experimentation without a path to platform is expensive R&D with no return.7 The leading organizations are not running more pilots. They are building infrastructure on which many use cases can run.
What that infrastructure looks like in practice is an agentic platform: a reusable agent library, a shared orchestration layer, persistent context and memory across deployments, continuous evaluation frameworks, and vendor-agnostic integration that prevents the platform from becoming hostage to any single technology provider. These are not speculative capabilities. They are the architectural choices that separate organizations generating real AI ROI from those still presenting slide decks about it.
The economics of this approach are fundamentally different from the pilot-by-pilot model. Each new use case built on existing infrastructure has a lower marginal cost and a shorter deployment cycle than the one before it. The platform compounds. The alternative, rebuilding from scratch each time, does not.
There is also an operational shift embedded in this architectural one. The traditional model is humans executing workflows with AI assistance. The platform model inverts that: AI systems execute workflows with human oversight. That distinction is not cosmetic. It determines how teams are structured, how decisions are made, and how the productivity gains from AI actually flow through to outcomes.
The Operating Model Has to Move Too
Technology alone does not solve this problem. This point is worth stating plainly, because most AI transformation efforts are structured as technology deployments rather than operating model redesigns.4 The technology gets deployed. The teams do not change. The workflows do not change. The decision rights do not change. And then leadership is puzzled when a well-architected system underperforms.
Agentic systems require AI-native workflows, smaller and more outcome-oriented teams, and humans positioned above the execution loop rather than inside every step of it. These are organizational design questions, not engineering questions. They require the same executive attention that the technology investment receives, and they rarely get it. The organizations that close the gap between AI capability and AI impact are the ones that treat the operating model redesign as a first-class deliverable, not an afterthought.
Fix the System, Not the Statistic
The 90% failure narrative is directionally correct and strategically misleading in equal measure. It is correct that most AI initiatives fail to reach scale. It is misleading because it implies the problem is with AI. It is not. The problem is with the systems AI is being asked to run in.
The organizations that close this gap will not win because they found a better model or a smarter vendor. They will win because they redesigned their architecture, workflows, and operating models before they deployed at scale. They built for composability, built for orchestration, and built governance in from the start.
The question worth asking is not whether the technology is ready. The question is whether your system is.
Sam Palazzolo
Fractional CRO | Growth Architect | Capital Strategist
References
- McKinsey & Company. The State of AI in 2023: Generative AI’s Breakout Year. McKinsey Global Survey on AI. https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai-in-2023-generative-ais-breakout-year
- McKinsey & Company. The Economic Potential of Generative AI: The Next Productivity Frontier (2023). https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-economic-potential-of-generative-ai-the-next-productivity-frontier
- McKinsey & Company. Scaling AI: From Experimentation to Impact. McKinsey Digital & QuantumBlack Insights. https://www.mckinsey.com/capabilities/quantumblack
- McKinsey & Company. Rewired: The McKinsey Guide to Outcompeting in the Age of Digital and AI (2023).
- Gartner. AI in Organizations: Adoption and Maturity Trends. Various reports, 2022-2024.
- NTT DATA. Global GenAI Report: Why Many AI Initiatives Fail to Scale (2024).
- Massachusetts Institute of Technology, Industrial Performance Center / MIT Sloan Management Review. Research on AI adoption and value realization.
- QuantumBlack. Creating a Future-Proof Enterprise Agentic Platform Architecture (2025). https://medium.com/quantumblack/creating-a-future-proof-enterprise-agentic-platform-architecture-c21fc48406a5
Your Value Is Specific. Is Your Price?
|