The Framework
One loop, five modalities.
The framework is small enough to hold in your head and specific enough to score two hundred real cases against. An organisation observes, thinks, and acts — continuously. Five modalities determine how well each phase runs. Everything else in the book, including the toolkit, builds on this.
The framework at a glance
Twelve ideas to remember
- 01
Agency
The capacity to stay aligned with reality, make sense of what it implies, and intervene with intent ahead of events. Observe, Think, Act is the operating system of agency.
- 02
The OTA loop
Observe, Think, Act — three phases run continuously. The flywheel of Chapter 5 is the metaphor for how repeated loops compound learning; it is the same loop, not a second framework.
- 03
Paranoid optimism
Paranoid enough to look for everything that can go wrong; optimistic enough to believe that knowing it puts you ahead of those who refuse to look.
- 04
The OTA 200
The empirical foundation: 200 companies, 100 failures and 100 successes, scored against the same 15-cell grid.
- 05
Five Modalities / OTA × modalities grid
Direction, Structure, Processes, Capability, and Culture, crossed with the three phases into a 15-cell diagnostic map of where failure and success concentrate.
- 06
Six Pathogens
Hubris, Complacency, Normalisation of Deviance, Groupthink, Misleading Incentives, and Inertia — the diseases that arrive with success and infect the loop from within.
- 07
The Toolkit
The leader's working repertoire for training Observe, Think, and Act. In Part IV the same tools gain a new engine; they are not replaced by a second toolkit.
- 08
Scenario Mapping
Keeps several plausible futures alive, with leading indicators for each branch. With AI, the same practice becomes continuous.
- 09
Decision Record→ Decision Ledger
One document opened before a commitment and closed after it — which, with AI, becomes a shared, queryable record every other tool reads from and writes to.
- 10
Reality Instrumentation→ Reality Lattice
Pointing the organisation's instruments at strategically relevant reality — which, with AI, becomes a continuous observation layer beyond human processing capacity.
- 11
Red Team→ Adversarial Engine
A designated challenger building the strongest counter-case — which, with AI, becomes continuous challenge with no status, ego, or career risk.
- 12
Decision Architecture→ Autonomy Ladder
The map of which decisions live at which level. The Autonomy Ladder is the governance mechanism within it through which AI earns, holds, and loses decision authority.
The OTA loop
Three phases, continuously
Most organizational failure looks, from the inside, like decisive leadership: meetings held, decisions signed, contracts executed. The book calls this the illusion of action — motion mistaken for progress. Speed is not the same as velocity. Speed means moving fast; velocity means moving fast in a direction that stays correct as reality changes. The Observe-Think-Act loop is the mechanism that keeps velocity honest, and it runs as three overlapping, continuously-informing phases rather than a start-to-finish sequence.
Observe is an active discipline, not passive awareness — building sensing systems that catch weak signals while they’re still deniable, not just consuming more information. Kodak’s leadership saw the digital-imaging transition coming in 1975 and forecast its timeline correctly in 1979; Observe was never where Kodak’s failure lived.
Think is where signals become an interpretive model — and models, once formed, become invisible: indistinguishable from “the way things are.” This is where Kodak actually failed: it saw digital coming and reasoned its way into believing the transition would stay slow enough to manage on film’s own timetable.
Act is the discipline of committing — resources, reputation, time — to a chosen course, and accepting that reality will prove you right or wrong. Most organisations mistake constant motion (meetings, pilots, reorganisations) for action; real action is a falsifiable bet.
None of the three works alone. Failure in any one collapses the agency the other two were building. The rest of the book — and the toolkit — exists to strengthen the loop that runs all three together, continuously.
Observe
Does the organisation see reality clearly and early — and does the signal reach the people who reason about it?
Think
Given the signal, does the organisation reason toward the right interpretive frame — or preserve the frame it already has?
Act
Given the reasoning, does the organisation execute — decisively, competently, and in time to matter?
One loop, two hundred organisations
The framework above isn’t built from one story. It’s scored against two hundred real cases — Nokia’s smartphone platform collapse, Boeing’s 737 MAX, Theranos, Enron, GE’s Alstom bet, Yahoo’s twice-missed Google acquisition, and many others, alongside a hundred organisations that got it right. Kodak is the worked example below because it isolates one phase so cleanly — but it’s one case among two hundred, not the exception the framework was built to explain. Browse the full corpus →
The loop, applied to one caseKodak, 1975–2012
In 1972, Roger Van Heyningen, director of the Physics Division at Kodak Research Laboratories, set up a small laboratory to explore electronic imaging. In 1979, Kodak's Larry Matteson forecast that imaging would shift from film to digital — first in government, then professional applications, and eventually the mass market — with the transition essentially complete by 2010. He was remarkably close. Kodak did not fail to see the future. It saw it decades in advance. The harder failure was acting on what it saw: the economics, incentives and decision processes of a company built around film repeatedly made the emerging digital business look less attractive than the business it threatened to replace.
Kodak is the worked example because one phase so clearly carried the outcome — and it is the one people usually get backwards: the company did not fail to see the digital transition coming.
In the study, Kodak’s Observe phase is scored Hard-Correct and carries 0% of the causal weight — it performed admirably and did not cause the outcome. Think carries 100%, classified Easy-Wrong: the interpretive frame needed was available and was not applied. The primary modality is Culture.
Read the full scored case →
- 1975ObserveSteven Sasson builds the first self-contained digital still camera in Kodak's Rochester laboratories.
- 1981ObserveThe Barabba market-intelligence study forecasts digital displacing film, with roughly a ten-year runway. It reaches senior leadership.
- 1994ThinkFisher creates the Digital and Applied Imaging division — formally separated, but left dependent on the film organisation's goodwill.
- 1996ThinkAPS/Advantix launches: more than $500m committed to a hybrid film-plus-digital-index format, extending film rather than replacing it.
- 2003ThinkPerez anchors the successor thesis on a consumer-inkjet razor-and-blade model, reasoning from an HP analogy.
- 2007ActSmartphones absorb the consumer camera category; digital-camera margins compress.
- 2012ActChapter 11 on 19 January: $6.75bn of debt against $5.1bn of assets. The core digital-imaging patent portfolio sells for $525m.
The five modalities
How each phase runs
- Direction
- The strategic choices an organisation makes — where it decides to go, and the dated, attributable decisions that commit it.
- Structure
- The architecture of authority and reporting — who is positioned to act, with what independence and what resources.
- Processes
- The machinery — planning cycles, coordination procedures, capital allocation — that turns intent into repeatable action.
- Capability
- The skills, assets, and competences the organisation carries — and whether they match what the situation requires.
- Culture
- The collective norms and reflexes that outlast any individual leader — and quietly decide what the other four are allowed to do.
The OTA × modalities grid
Three phases × five modalities
| Direction | Structure | Processes | Capability | Culture | |
|---|---|---|---|---|---|
| Observe | Observe × Direction | Observe × Structure | Observe × Processes | Observe × Capability | Observe × Culture |
| Think | Think × Direction | Think × Structure | Think × Processes | Think × Capability | Think × Culture |
| Act | Act × Direction | Act × Structure | Act × Processes | Act × Capability | Act × Culture |
The book walks the grid cell by cell. The study shows what two hundred organisations look like when you actually fill it in.
What the grid keeps showing
Three concentrations, one asymmetry
Where the data concentrates
The Think × Direction concentration
The most consequential cell in the entire grid is Think × Direction — where the quality of strategic reasoning meets the question of where the organisation is going. It carries about 20% of non-fraudulent failure weight and 18% of success weight, the peak cell in both cohorts. Same cell, different content. In failure, it is a leadership team with a confident, sophisticated model of how the industry works, applied where it no longer fits (JCPenney under Ron Johnson: a brilliant model of retail psychology, deployed in an environment that rejected it). In success, it is a leadership team whose model proves a more accurate read of the industry’s evolution than its competitors’ (Netflix holding two incompatible business models through the 2007–2012 transition; TSMC keeping its foundry-neutrality model against industry consensus).
The Direction column
The full Direction column, across all three OTA phases, is the grid’s central empirical fact: it carries 28% of non-fraudulent failure weight and 39% of success weight. The same column that makes an organisation succeed can make it fail when the disciplines that keep direction accurate are abandoned. Polaroid’s engineers saw digital photography coming — the failure wasn’t observation, it was a model (“digital will complement prints, not replace them”) that observation could not change. Amazon’s direction strength came from the opposite practice: returning to the same question — does our direction still serve the model? — as each new bet arrived, from third-party sellers to AWS to Prime. Direction strength is not a single correct insight but a model kept under examination as evidence changes.
Culture in the OTA 200
Culture is the most asymmetric modality in the grid. It is the primary modality in 90% of fraudulent failure cases, versus 8% of success cases; in non-fraudulent failures it drops to 31%, behind Direction at 41%. Culture kills but doesn’t save. Boeing’s engineers eventually accepted the 737 MAX plan, Volkswagen’s engineers understood what the defeat device was doing, Wells Fargo’s branch staff knew the account-opening targets were generating fraud — in every case the information existed, and the people who had it made a rational calculation that the cost of speaking exceeded the cost of silence. The useful question is not “is our culture a competitive advantage?” but “are cultural conditions putting us at disproportionate risk?”
Core finding
The asymmetry between failure and success
Success and failure are not mirror images, and the study scores them with two different questions because of it. Failure cases ask what caused the outcome to go wrong — weight follows causation backward. Success cases ask what distinguished this organisation from a reasonably-resourced peer — weight follows the contribution that made the difference. The same correctly-executed phase that carries zero in a failure case can carry the majority of the weight in a success case. Treating the two symmetrically is itself a reasoning error, and it is why the self-assessment below runs two separate tracks.
Concretely: Capability carries roughly 20% of success weight but only about 2% of failure weight — it saves but rarely kills — while Culture is close to the mirror image, dominating roughly 49% of failure cases against just 8% of success cases. Prevention and success-building are structurally different work; you cannot fully outsource one to the other.
Grand Unified Theory (Almost)
Why another model of organisational excellence?
Over the years, I have studied many models of organisational excellence. Several offer valuable insights, and some are impressively comprehensive. Yet I have often found myself frustrated by their practical application. Some are complicated enough that remembering and applying them becomes an exercise in itself. Others are relatively straightforward to understand but do not make it intuitively obvious what a leader should actually do to improve an organisation. Some suffer from both problems.
I wanted something simpler. A model that would help me understand how an organisation works, identify where improvements are needed, and, most importantly, know where to start.
The comparison of two hundred organisational successes and failures, examined through the OTA × modalities grid, led me towards such a model.
The result is what I somewhat jokingly call the Grand Unified Theory (Almost).
The basic idea is remarkably simple. An organisation needs a strong Direction. Structure, Processes, Capability and Culture must support that Direction, but they must also be aligned with one another. Even if each modality appears to support the strategy independently, the organisation may not function well if the supporting modalities work against each other.
These conditions cannot be designed once and then left alone. The organisation must continuously improve its Direction, the effectiveness of each modality and the alignment between them. The mechanism for doing this is the Observe–Think–Act loop running within each modality. Each loop must examine not only how the modality itself is performing, but also how well it works with the others.
AI gives us the opportunity to operate this management machinery with a quality, continuity and speed that previously would have been impossible or impractical. Humans retain accountability and responsibility for key decisions.
The Theory does not claim to capture every aspect of management. Nor does it replace the many useful insights of existing organisational models. Its ambition is more practical: to provide a sufficiently comprehensive picture of an organisation, in a form that makes it relatively straightforward to begin improving it.
Five components of the Grand Unified Theory (Almost)
- 1
A strong Direction
An organisation needs a sound strategic Direction: a clear understanding of where it is going, why that is the right destination and how it intends to succeed.
Direction is not simply a strategy document or a set of targets. It is a set of choices and assumptions about the future that must remain valid as circumstances change.
The OTA-200 research highlights the exceptional importance of Direction, particularly Think × Direction. This makes Direction the natural starting point for organisational improvement.
- 2
Supporting modalities aligned with Direction
Structure, Processes, Capability and Culture must all support the organisation’s Direction.
Each modality should be examined against the same strategic question: Does the way this part of the organisation operates help us achieve our Direction, or does it work against it?
The four supporting modalities must be designed and continuously adjusted to reinforce Direction.
- 3
Supporting modalities aligned with one another
Alignment with Direction is necessary, but not sufficient.
Structure, Processes, Capability and Culture must also work coherently together.
For example, an organisation might give teams significant decision-making authority through its Structure while its Processes still require central approval for those same decisions. Both modalities may have been designed with Direction in mind, yet together they create friction and inefficiency.
Similarly, an organisation may build advanced capabilities that its processes cannot effectively deploy, or introduce processes that its culture quietly resists.
A well-functioning organisation requires mutual alignment among all four supporting modalities.
- 4
OTA loops create and maintain effectiveness and alignment
How can an organisation establish and continuously maintain a strong Direction, effective modalities and alignment between them?
By continuously running the Observe–Think–Act loop within each modality.
Each loop must examine three things: how effectively the modality itself is functioning, how well it supports Direction and how well it is aligned with the other modalities.
Observation reveals weaknesses and opportunities. Thinking identifies possible improvements. Action implements those improvements and generates new information for the next cycle.
Every modality needs an accountable owner responsible for making its loop work and continually improving it.
The objective is to make these loops both better and faster.
- 5
AI operates and optimises the machinery, with humans accountable
AI makes it possible to operate and continuously improve this management machinery far more effectively.
It can help observe relevant developments, identify emerging misalignments, challenge assumptions, generate alternatives, recommend improvements, execute authorised actions and monitor outcomes.
Instead of applying AI only to isolated tasks, the ambition is to redesign how the entire management system operates.
AI should be used actively within every OTA phase and increasingly through agentic architectures that automate appropriate parts of the system.
Humans remain accountable, retain responsibility for key strategic and consequential decisions, and determine the authority delegated to AI.
Why the Theory begins with Direction
- Think phase
- 62.2%Mean failure-phase attribution
- Think × Direction
- ~20% / ~18%Non-fraudulent failure weight / success weight
- Entire Direction column
- 28% / 39%Non-fraudulent failure weight / success weight
These findings help explain why the Theory begins with Direction and why tools supporting better strategic thinking deserve particular emphasis. The first figure covers the full failure cohort; the other two use the narrower non-fraudulent failure subset, so the three are not interchangeable measures.
The Grand Unified Theory (Almost), in five views
1. A strong Direction
A strong Direction is the foundation. The organisation must continuously test its strategic assumptions, challenge its thinking and adjust its choices as reality changes. The OTA-200 research places particular emphasis on Think × Direction.
Scroll sideways for all five modalities
2. Supporting modalities aligned with Direction
Structure, Processes, Capability and Culture must each support the organisation’s Direction. All four need to reinforce the chosen strategy rather than undermine it.
Scroll sideways for all five modalities
3. Supporting modalities aligned with one another
Alignment with Direction is not enough. Structure, Processes, Capability and Culture must also fit together. Contradictions between them create friction, weaken execution and can undermine even a sound strategy.
Scroll sideways for all five modalities
4. OTA loops create and maintain alignment
Each modality continuously observes its own effectiveness and relationships, thinks about improvements and acts to implement them. The five OTA loops create and maintain both internal effectiveness and alignment across the organisation. Every loop needs an accountable owner.
Scroll sideways for all five modalities
5. AI-native operation
AI can strengthen and increasingly operate the observation, thinking and action activities that maintain Direction, organisational effectiveness and alignment. Agentic systems can automate continuous monitoring and other authorised workflows. Humans remain accountable and retain responsibility for key decisions.
Scroll sideways for all five modalities
Where should a leader start on Monday?
The Grand Unified Theory (Almost) tells us what a well-functioning organisation requires. But how do we get there?
The answer follows the same simple logic as everything else in this book:
Observe. Think. Act. Then repeat — better and faster.
And use AI to help at every step. Ask it questions, make it investigate, challenge its answers, verify its evidence and iterate until you have something you can act on.
1 · Observe
Understand where you stand
Begin by creating an evidence-based picture of your organisation across all five components of the Theory and all five modalities.
How strong is your Direction? How well do the other modalities support it? How well are they aligned with one another? How effectively are the OTA loops running within each modality? How well is AI being used, and are human accountability and decision rights clear?
Start by asking three questions:
How can we best observe our current situation?
What do we need to know? What information is already available? What are the most important gaps in our understanding?
What are the best KPIs?
Which leading and lagging indicators would most clearly reveal how well the five components of the Theory are functioning? Which would expose deterioration, misalignment or missed opportunities before they become serious problems?
What can we learn from others?
Which organisations have developed particularly effective practices? What are the best approaches we can find outside our own organisation, and how could we adapt them to our circumstances?
Use AI extensively in this phase.
Ask AI to propose the most useful KPIs, identify blind spots, research external best practices and suggest better ways of observing your organisation.
Challenge its proposals. Ask why a KPI matters, what it fails to measure and whether a better indicator exists. Require sources for external claims. Compare alternative recommendations and iterate.
Go beyond periodic assessments. Consider using agentic AI architectures to automate Reality Instrumentation: continuously gather approved internal and external information, monitor relevant KPIs, detect anomalies and changes, and produce evidence-backed observations for leadership.
The objective is not to collect more information for its own sake. It is to become progressively better at understanding reality.
2 · Think
Create the improvement plan
Use your observations to develop an integrated plan for improving the organisation.
The plan should consider all five modalities and all five components of the Theory:
- Strengthening Direction.
- Aligning the supporting modalities with Direction.
- Improving alignment among the supporting modalities.
- Improving the quality and speed of the OTA loops within each modality.
- Using AI to operate the management system more effectively, while maintaining human accountability and key decision authority.
Not everything needs to change at once. Identify the highest-impact opportunities, their dependencies and the best sequence for acting on them.
Use AI as an active thinking partner.
Ask it to develop alternative improvement plans, identify dependencies, assess trade-offs and propose priorities.
Use AI to run Pre-Mortems: Why might the plan fail? What are we overlooking? What assumptions could prove wrong?
Use Scenario Mapping to explore different futures and how the plan would perform in each.
Use Red Teams and the Adversarial Engine to challenge the proposed Direction, the improvement priorities and the underlying assumptions.
Do not accept AI’s first answer. Challenge its reasoning, ask for counterarguments, test its assumptions and iterate. Combine AI-generated insights with human knowledge, judgement and experience.
Use the Toolkit throughout this process.
The resulting plan should specify what will change, why it matters, who is accountable, what resources and decisions are required, which KPIs will indicate progress and how outcomes will be evaluated.
The objective is not an impressive presentation. It is a coherent plan that people can start implementing.
3 · Act
Implement, measure and learn
Start implementing the plan.
Assign responsibilities, including an accountable owner for the OTA loop within each modality and clear responsibility for coordination between modalities.
Begin with the most valuable improvements. Move quickly when decisions are reversible. Experiment, measure outcomes and adjust. Apply greater deliberation to consequential, difficult-to-reverse decisions.
Use AI throughout implementation.
AI can help coordinate initiatives, prepare decisions, monitor progress, identify emerging problems and recommend corrective actions. Agentic architectures can increasingly execute authorised workflows and automate parts of the monitoring and learning process.
Use Reality Instrumentation and, where appropriate, the AI-powered Reality Lattice to continue gathering information as the plan progresses.
Monitor whether the expected improvements are occurring, whether new misalignments are appearing and whether the original assumptions remain valid.
Use After-Action Reviews and other Toolkit practices to compare what actually happened with what was expected.
Keep challenging AI’s recommendations. Verify important conclusions, examine unexpected results and adjust both the plan and the AI systems themselves as you learn.
Humans retain accountability, make the key decisions and determine which actions AI is authorised to perform.
Act ↺ Observe
Then rotate the loop again.
The next round of observation should begin with everything learned during implementation.
The ambition is not to complete an organisational improvement project. It is to build an organisation whose ability to Observe, Think and Act keeps improving.
Observe. Think. Act. Repeat — better and faster, with AI helping throughout.
The Theory tells us what to improve. The Toolkit provides practical ways to do it.
Explore the ToolkitConcepts that travel with the tools
Not tools themselves
- Calibration
- WWHTBT
- Psychological safety
- Paranoid optimism
- Calibrated trust
- Triangulation
- Reversibility
- Layered rollout
Score your own organisation
The way an OTA-200 rater scores a real case
“The two case types ask fundamentally different questions. Failure cases apply root-cause analysis: the scoring asks which phase or phases caused the outcome to go wrong, and weight follows causation backward from the failure. … Success cases apply differentiation analysis: the scoring asks which phase or phases distinguished this organisation from a reasonably-resourced peer that would have done less well, and weight follows the contribution that made the difference.”OTA Scoring Methodology v4, §1
What is the core reason your organisation could have an extraordinary future success?
Allocation 1 of 2
Divide 100 points across the phases
Which phase of the loop would be most differentiating? Drag the point, click anywhere in the triangle, or use the +/− controls.
- Observe.
- The organisation's perception of its environment: competitors, customers, technology shifts, regulatory change, and its own internal state.
- Think.
- The organisation's internal reasoning about what its observation implied and what it should do in response.
- Act.
- The organisation's external actions taken on the basis of that reasoning: announcements, investments, reorganisations, product launches, disposals, hires, cuts.
Allocation 2 of 2
Pick three modalities, then divide 100 points
Identify the main differentiator — what set this company apart from peers who never became great successes.
Select 3 more (0 of 3 chosen).
“Each case is scored along five modalities. Modality percentages sum to 100 per cent per case. No case is scored along more than three modalities at once; the other two are recorded as zero. Weights use 5-percentage-point increments with a 10 per cent floor on any non-zero modality.”OTA Scoring Methodology v4, §3
Select three modalities to continue.