Insights & Perspectives

    Full Pipeline.
    Wrong Bets.

    Organizations spent three decades getting faster at building and shipping software. They never addressed whether they are building the right thing. Every improvement in speed and delivery was real. The bets filling those pipelines were never validated.

    Will Evans / Fugue Strategy Advisors
    March 2026
    14 min read
    The Scale of Strategic Failure · Enterprise Decision-Quality Research
    88%
    of 1955’s Fortune 500 are gone
    AEI · De Geus (1997)
    67%
    of well-formulated strategies fail
    Sull et al. · HBR 2015 · N=7,600
    <3%
    of leadership time on the future
    Hamel & Prahalad · HBR 1994
    30%
    higher returns from active reallocation
    McKinsey Quarterly · 15-year study

    Your roadmap is full. Your pipeline is humming. Your teams are shipping. And the bets that fill every one of those roadmaps were placed before anyone observed what is actually happening in the environment they were supposed to address.

    Every system has a constraint. A single point that governs throughput for the entire value stream. That constraint never disappears. When you address it at one node, it migrates to the next weakest point in the system. The history of enterprise IT over the past thirty years is the history of that migration.

    Through the 1990s, the binding constraint was engineering capability. Building software was slow, fragile, and expensive. Waterfall processes stretched timelines to years. A series of lightweight methodologies emerged to address this bottleneck: Rapid Application Development in 1991, Scrum in 1995, Extreme Programming in 1996. In February 2001, seventeen practitioners codified what had been forming in practice for a decade. The Agile Manifesto. Small accountable teams, timeboxed increments, fast feedback loops. Agile addressed engineering as the binding constraint. Teams could build. But they still could not test, deliver, measure, or monitor what they built. The constraint moved.

    By the early 2010s, the bottleneck had migrated to delivery. Teams could produce working software faster than ever, but they could not get it into production reliably. Testing was manual and slow. Configuration was fragile. Release management was a chokepoint. Operations and monitoring were disconnected from development. In 2013, Gene Kim, Kevin Behr, and George Spafford published The Phoenix Project, explicitly modeling it after Goldratt's The Goal. The book applied Lean principles, Theory of Constraints, and sociotechnical systems theory to the delivery problem. The DevOps movement that followed addressed testing, configuration, release management, operations, and monitoring as a unified value stream. CI/CD pipelines, infrastructure as code, accountable stable teams, and automated feedback loops alleviated the delivery bottleneck. For organizations that invested in it, the delivery constraint is addressed, or addressable. The system can move value from concept to production with speed and reliability. The constraint moved again.

    Around 2018, the bottleneck migrated upstream to product. "Projects to Products." Product-driven organizations. The books were written. The movements launched. Except most organizations never fully built the decision layer that product leadership requires. What they call product teams are often order-taking feature factories for the business, with little connection to organizational strategy, KPIs, or value realization. They did not build a decision layer. They built a pass-through.

    The intent was there. But the timing was brutal. Just as enterprise IT was beginning to address the product constraint, the pandemic hit. Markets evaporated overnight. Organizations scrambled to shift to remote work, then back to the office, then to hybrid models that satisfied nobody. Budgets were frozen, then redirected, then frozen again. Some organizations managed to build real product capability through that disruption. Many did the best they could with what they had, which often meant rebranding Business Analysts and Project Managers as Product Managers without providing the clarity of role definition, the competency through training in economic trade-off decision-making, or the capacity to learn new ways of working while simultaneously keeping the lights on. The function that was supposed to make outcome-driven decisions about what to build, what problems to solve, and where enterprise value lives was handed to people who were never given the tools or the time to learn what that function actually requires. The decision layer never fully formed. And ProductOps, the discipline designed to operationalize that layer, arrived to scale a capability that is still nascent in most organizations.

    And the constraint moved past product, too.

    The bets filling every roadmap are committed before they ever reach a product leader. Strategic commitments arrive pre-formed. Planning assumptions arrive untested. The highest-paid opinion in the room sets direction for every team, every budget, every roadmap. And those commitments are built the same way they were built twenty years ago: internal data, leadership experience, periodic customer surveys, and assumptions that were never validated against what is actually happening in the environment the organization exists to serve.

    The constraint is now in the intelligence feeding the entire system. And every improvement below it is increasing the throughput of unvalidated bets.

    I
    The Constraint Migrates

    The Bottleneck at the Top
    of the Bottle.

    Goldratt made the principle plain:

    "Any improvement not made at the constraint is an illusion."
    Eliyahu Goldratt · The Goal · 1984

    The constraint in a system of value delivery is never static. It sits wherever the system is weakest. Address it at that node and it moves to the next one. The system's total throughput is always governed by wherever the constraint currently sits, regardless of how much capacity exists everywhere else. An hour saved at a non-bottleneck is a mirage.

    This is the lens through which the entire Ops lineage needs to be read. Agile did not solve engineering. It addressed engineering as the binding constraint, and the constraint moved to delivery. DevOps did not solve delivery. It addressed delivery as the binding constraint, and the constraint moved to product. Product management was supposed to be the next discipline to address the next bottleneck. It stalled. Most organizations never built the decision layer. They built feature factories.

    And the constraint moved past product to the intelligence feeding the entire system. Where it is sitting now. Unaddressed.

    I saw this at the DevOps Enterprise Summit in San Francisco in 2016, presenting a case study on the Lean DevOps transformation we ran inside Johnson and Johnson. The world's largest healthcare company. Highly siloed, matrixed IT organization. The kind of place where people say "DevOps just won't work here." We used enterprise architecture to identify constraints, run experiments, and decrease lead-time across all of IT rather than optimizing specific functions or products. We focused on system throughput. And through those experiments, we uncovered something that mattered more than velocity: the enterprise IT organizations that survive are the ones that learn to continuously re-align IT with business intent. Not deliver faster. Deliver the right thing. The constraint was not in the engineering or the delivery. The constraint was in knowing what to build and why.

    That question has only become more urgent. Roadmaps are filled to capacity with commitments to build things that may or may not ever deliver value, because those commitments are based on assumptions that were never tested against observed reality. The system is optimized for throughput at every node below the constraint. And the constraint itself, the intelligence that determines what gets built and why, has never been operationalized.

    The intellectual foundations of strategic planning were dismantled thirty years ago. In 1994, the structural reason planning kept failing was identified with precision:

    "Strategy cannot be planned because planning is about analysis and strategy is about synthesis. That is why the process has failed so often and so dramatically."
    Henry Mintzberg · The Rise and Fall of Strategic Planning · 1994

    Analysis disaggregates. Synthesis integrates. The formal planning process was exceptionally good at the former and structurally incapable of the latter. The harder organizations tried to be rigorous in their planning, the further they moved from the emergent, adaptive intelligence that effective strategy actually requires. Planning produced commitment to a predicted future, and that commitment made organizations less capable of perceiving when the actual future was different from the predicted one.

    Two years later, the blade was sharpened further. The essential problem, as "Strategy as Revolution" laid bare, is the failure to distinguish planning from strategizing. Planning is ritualistic, reductionist, extrapolative, and elitist. It works from today forward, not from the future back. And then the line that should be carved into the wall of every boardroom in the Fortune 500:

    "The bottleneck is at the top of the bottle."

    Where are you likely to find people with the least diversity of experience, the largest investment in the past, and the greatest reverence for how things have always been done? At the top. The people holding the strategy pen are the people least equipped to see what the strategy needs to address.

    The research on how leadership actually spends its time confirmed what the theory predicted. Senior managers devote less than 3% of their time to building a shared perspective on the future. The urgent drives out the important. Every quarter. And organizations respond to the resulting drift not by investing in foresight but by restructuring. Cutting denominators rather than growing numerators. Optimizing what exists rather than sensing what is emerging.

    The Constraint Migrates Upstream
    1990sEngineeringCapabilityADDRESSED BY AGILE2010sDelivery SpeedRelease & OpsADDRESSED BY DEVOPS2016+Product StrategyWhat to BuildUNRESOLVEDNOWStrategicIntelligenceThe Right BetsFORESIGHTOPSThe binding constraint

    This is the constraint that has been sitting at the top of the strategic funnel, unaddressed. Enterprise IT did not spend thirty years failing to build and deliver software. Agile addressed the engineering bottleneck. DevOps addressed the delivery bottleneck. Both effectively. The system can now move value from concept to production faster than at any point in history. But the constraint is no longer in the code or the pipeline. The constraint is in the intelligence. Roadmaps are filled to capacity. The pipeline is full. The teams are shipping. But the bets that fill those roadmaps were placed before anyone observed what was actually happening in the environment those bets were supposed to address. The system is optimized for throughput of commitments that were never validated.

    The planning process cannot produce strategy. The people running it are the least likely to see what needs to change. And the constraint will not wait to be noticed. It simply moves to wherever the system is weakest.

    It moved to the top of the funnel. And nobody built the infrastructure to address where it landed.

    II
    The Sensing Gap

    The Missing
    Discipline.

    The Sensing Gap
    WHAT IS ACTUALLY HAPPENINGBehavioral signalWeak signals of changeDisconfirming evidenceEmerging needsPre-categorical patternsThe Sensing GapStructural distanceFILTERED: QUALITATIVE SIGNALFILTERED: DISCONFIRMING EVIDENCEWHAT REACHES DECISION-MAKERSQuantitative dataConfirming evidenceStakeholder opinionLast quarter's resultsInherited assumptionsForesightOps closes this gap with continuous behavioral infrastructure

    The sensing gap. The structural distance between what is actually happening in an organization's environment and what reaches decision-makers in a form they can act on. This gap is not visible from inside the organization. Organizations that have it are not aware of what they are not seeing. That is the defining characteristic of the problem. They are planning, executing, and adapting. Doing strategy the way strategy has always been done. What they do not have is a way to know whether the information available is the information that matters. Whether the signals reaching the decision-making level are the signals that would actually change the decisions being made. Whether the planning process is beginning from an accurate read of what is happening in the communities and markets the organization exists to serve, or from an accumulated set of internal assumptions that have never been systematically tested against external behavioral reality.

    The discipline that closes this gap is what I am calling ForesightOps.

    ForesightOps is the operational infrastructure for continuous behavioral research. It transforms what organizations observe about real human behavior into strategic intelligence that improves decision quality before commitments are made. Not after outcomes have revealed what was missing.

    It is not a research methodology. It is not a technology platform. It is not a consulting engagement with a terminal state. It is the discipline of building the systems that make genuine organizational listening continuous, scalable, and connected to the decisions that matter. It persists. It compounds. It operates whether or not an external advisor is in the room. The advantage widens with every cycle.

    The Strategic Stack · Where ForesightOps Sits
    STRATEGY & DISCOVERYForesightOpsDiscover what is actually happening. Test assumptions. Inform strategy.DOES NOT EXIST IN MOST ORGSThe binding constraint in the system.The intelligence to place the right bets. INFORMS WHAT TO BUILD AND HOW TO SELL PRODUCT MANAGEMENTProductOpsInherits strategic constraints from aboveREVENUE & GROWTHRevOpsInherits strategic constraints from above INFORMS DELIVERY DELIVERY & QUALITYDevOps
    The Structural Problem
    Most organizations have operationalized delivery, product management, and revenue. They have not operationalized discovery. ResearchOps scales the mechanics of research, but it typically sits too far down in the organization, informing feature-level trade-offs rather than strategic ones. The research happens. The findings get filed. The strategy proceeds on the assumptions that were already in the room before anyone went looking. ForesightOps fills the top of the funnel with evidence instead of opinion. It addresses the binding constraint in the system: the intelligence to place the right bets.
    III
    Five Core Practices

    What ForesightOps
    Practitioners Do.

    Informed by John Boyd's OODA Loop (Observe-Orient-Decide-Act) and Snowden & Boone, "A Leader's Framework for Decision Making," HBR 2007

    These are not project phases. They are continuous, interconnected practices that define the discipline. An organization doing ForesightOps is doing all five, all the time, at whatever scale fits its maturity. Each practice generates the input the next one requires.

    01
    Observe
    Continuous Discovery
    Ongoing behavioral observation at community scale
    Structured observation of what people actually do in natural environments. Not periodic surveys. Not focus groups. Not stakeholder opinion treated as customer evidence. Continuous, embedded, behavioral observation using trained community researchers with access and trust that external consultants cannot replicate. The goal is pre-categorical signal: what is happening before anyone has decided what category it belongs to.
    Contextual Inquiry
    Narrative Interviews
    Cultural Probes
    Mobile Ethnography
    Participant Observation
    02
    Synthesize
    Signal Synthesis
    Abductive sensemaking from observed behavioral signal
    This is not translation. Translation implies a one-to-one mapping from input to output. Synthesis is an abductive act of creation. It produces something new that did not exist in any single observation. Raw behavioral signal from continuous discovery becomes Behavioral Impact Personas grounded in observed motivation rather than demographic categories, weak signal registries that detect directional change early, and living intelligence layers that update as new signal arrives. Discovery produces the raw material. Synthesis is the designerly act of making meaning from it.
    Behavioral Impact Personas
    Weak Signal Registry
    JTBD Maps
    Pattern Clustering
    03
    Test
    Assumption Testing
    Validating planning assumptions against behavioral evidence
    Every strategy has load-bearing assumptions: things the plan requires to be true that could plausibly turn out false. Most organizations cannot name theirs. ForesightOps practitioners map assumptions explicitly, prioritize them by strategic risk, and test them against observed behavioral evidence. This is where the most valuable disconfirming evidence surfaces.
    Assumption Map
    Assumption Audit
    Disconfirming Evidence Log
    Decision Quality Record
    04
    Navigate
    Futures Navigation
    Scenario planning grounded in observed signal, not internal extrapolation
    Building plausible futures from what has actually been observed, not from what the leadership team extrapolates from last quarter’s results. Four plausible scenarios constructed using the Oxford Scenario Planning Approach, each stress-tested against the current strategy. Strategies that perform well across all four are robust bets. Strategies that require a specific future to occur are vulnerable ones.
    Scenario Sets
    Strategy Stress-Tests
    Robust Strategy Maps
    Bet Vulnerability Assessment
    05
    Decide
    Strategic Activation
    Connecting intelligence to decisions with explicit evidence traceability
    Closing the loop between observation and commitment. Every recommendation in a ForesightOps-informed roadmap traces back to specific behavioral observations, specific patterns, specific community voices. If a strategic initiative cannot trace its rationale to evidence, it is an assumption. Assumptions should be tested, not funded.
    Evidence-Traced Roadmaps
    Innovation Portfolios
    Capability Handoff
    Foresight Playbooks
    These five practices are continuous and cyclical, not sequential project phases. Strategic Activation feeds back into Continuous Discovery as the organization learns what new questions its roadmap commitments have created. The cycle does not have a terminal state. That is what makes it infrastructure rather than a project.
    The Continuous Cycle of Strategic Action
    After Boyd's OODA Loop (1976) and Snowden & Boone, "A Leader's Framework for Decision Making," HBR, November 2007
    Infrastructurenot project01OBSERVE02SYNTHESIZE03TEST04NAVIGATE05DECIDE
    IV
    Principles

    What ForesightOps
    Practitioners Believe.

    Assumptions age. Evidence compounds.
    Every planning assumption degrades as conditions change. Every behavioral observation adds to a growing intelligence layer. The advantage widens with every cycle the infrastructure operates.
    Decision quality is the only lever you control.
    Good process with bad luck beats bad process with good luck over time. ForesightOps improves the expected value of every strategic decision across the entire portfolio. The improvement is systematic, and it compounds.
    Observe behavior. Not attitudes.
    What people do and what they say they do are not the same thing. The methodology is built on behavioral signal because behavior is what actually drives outcomes. Watch the hands, not the survey responses.
    The most valuable finding contradicts what you believe.
    Research designed to confirm produces noise dressed as signal. Every method is structured to surface disconfirming evidence and weak signals. Confirmation is cheap. Disconfirmation is where the strategic value lives.
    Build capability, not dependency.
    Every engagement ends with the organization more capable of independent foresight practice than when it began. Researchers are trained. Protocols are documented. Methodology is transferred. The goal is self-sufficiency.
    Strategy should be traceable to evidence.
    If a strategic initiative cannot trace its rationale to behavioral observation, it is an assumption in disguise. Evidence traceability is what separates strategy from opinion.
    ·

    This is the gap in the Ops lineage. Not a missing feature. A missing discipline. The constraint moved, and nobody built the infrastructure to address where it landed.

    The organizations that hedged well in every previous disruption share one characteristic. They did not predict the future. They built the infrastructure to see it forming. They placed better bets because they had better intelligence feeding the decision. Not better analysts. Better systems.

    You cannot react your way to alpha. You have to build toward it.

    References
    1 Mintzberg, H. (1994). The Rise and Fall of Strategic Planning. Free Press.
    2 Hamel, G. (1996). "Strategy as Revolution." Harvard Business Review, July-August 1996. Reprint 96405.
    3 Hamel, G. and Prahalad, C.K. (1994). "Competing for the Future." Harvard Business Review, July-August 1994. Reprint 94403.
    4 Goldratt, E.M. (1984). The Goal: A Process of Ongoing Improvement. North River Press.
    5 Evans, W. and Landy, M. (2016). "The Need for Speed: Enabling DevOps through Enterprise Architecture." DevOps Enterprise Summit, San Francisco. Video
    6 Sull, D., Homkes, R. and Sull, C. (2015). "Why Strategy Execution Unravels." HBR, March 2015. N=7,600.
    7 Hall, S., Lovallo, D. and Musters, R. (2012). "How to Put Your Money Where Your Strategy Is." McKinsey Quarterly, March 2012.
    8 Perry, M.J. (2022). "Only 52 US Companies on Fortune 500 Since 1955." American Enterprise Institute.
    9 De Geus, A. (1997). "The Living Company." HBR, 75(2):51-59.
    Will Evans
    Chief Strategy Officer at Fugue Strategy Advisors and founder of ForesightOps. Thirty years at the intersection of User Experience Design, Human-Centered Design, Ethnography, and Strategic Foresight. Fortune 50 engagements across fintech, pharma, healthcare, and e-commerce. Founder of LeanWX NYC. Lecturer at NYU Stern School of Business.