Scenario 05 — Cross-functional Alignment and Decision Follow-through

Part 01 · Scenario boundaries

What this page covers, and what it does not

Pull the familiar cross-functional conversation up to the governance layer before any framework.

Scene description

A clear portfolio priority does not make a study move by itself. Resources placed around a critical constraint do not automatically resolve a cross-functional issue. A capable team can still stop when Clinical Development, Regulatory Affairs, Biostatistics, Data Management, Quality, Clinical Supply and Clinical Operations reach different conclusions.

Scenario 02 addresses portfolio priorities and delivery trade-offs. Scenario 03 addresses constraints and operational support. Scenario 04 addresses team capability and delegation. Scenario 05 builds on those foundations and asks a more specific question:

When a critical decision requires several functions and no single function can resolve the issue alone, how does the Clinical Operations leader move the organization from fragmented judgment to a clear decision, and from that decision to an operational result?

Cross-functional work in clinical development involves more than exchanging information. Four activities must connect:

【1】align on the project situation;

【2】frame the decision;

【3】complete the governance decision;

【4】drive implementation and verification.

If one activity is missing, everyone may attend the meeting while the study remains in the same place.

Scope

This page focuses on cross-functional governance in clinical development. The discussion centers on how a Clinical Operations leader organizes facts, professional input, decision rights and implementation feedback around important clinical-development and clinical-operations issues. "Cross-functional" may include Clinical Development/Medical, Biostatistics, Data Management, Regulatory Affairs, Pharmacovigilance, Clinical Supply, Quality, Finance and Clinical Operations. Organization names and governance structures vary by company; this page does not assume one universal model.

This page covers

Why do different functions build different interpretations of the same study?
How should the team separate project facts, professional judgments, key assumptions and functional positions?
How can "functional disagreement" become a decision that governance can actually resolve?
Which issues should follow an established process, receive expert analysis, be explored through bounded probes, or be stabilized immediately?
Who provides input, who recommends, whose agreement is required, who decides, and who performs?
How does a governance decision enter the study plan, resource allocation, risk management and subsequent review?
How can a Clinical Operations leader move an issue forward without holding final decision authority?

Out of scope

Generic communication tips or workplace relationship advice
Meeting management whose only objective is to "run a better meeting"
Enterprise-wide organization redesign
External-partner governance centered on vendor performance and contracts
A complete method for CAPA, root-cause analysis or quality effectiveness checks
One company's fixed committee or delegation structure
Using rhetoric to disguise facts, override professional dissent or manufacture superficial agreement

Expected outputs

Cross-functional decision brief
Shared fact base and key-assumption list
Cynefin context assessment
Option set and integrated impact assessment
RAPID decision-rights map
Formal decision record
Decision implementation and review log
Governance escalation path

Part 02 · Problem decomposition

The core question

How does a Clinical Operations leader turn partial functional judgments into a project decision with a credible fact base, clear decision rights, an executable path and evidence that can test the original judgment?

Question map — six sub-questions

1. Are people actually discussing the same problem?

What project facts have been confirmed?
Are all functions using the same data cut and status information?
Which statements are facts, which are professional judgments, and which remain assumptions?
How does each function explain the cause of the problem?
Which statements have been repeated without ever being tested?
What information is missing, and would the missing information change the decision?

2. What decision does the project actually need?

Does the team need information sharing, professional assessment, or a formal decision?
Can the decision be written in one complete sentence?
Which project objective does the decision serve?
Which participant-protection, scientific, quality and regulatory guardrails must be preserved?
Is the real trade-off about evidence, time, resources, risk, or the development window?
What is the latest responsible decision date?
What will continued delay cost the project?

3. What kind of problem is this?

Is cause and effect clear enough for an existing process to address the issue?
Does the issue require analysis from one or more expert functions?
Do participant, investigator, site or system interactions make a reliable answer impossible in advance?
Would a bounded, low-risk probe provide better evidence?
Is there an urgent risk that must be stabilized before analysis?
Does one large issue contain several different types of subproblem?

4. Which professional inputs and options are required?

Which functions hold facts or expertise that could change the decision?
Which inputs are essential, and which inputs simply reflect a wish to be heard?
Has the team developed options that can genuinely be compared?
How would each option affect scientific integrity, participant protection, statistics, regulatory strategy, operations, data, supply, timeline and budget?
Which risks can be accepted, and which risks require explicit safeguards?
Is there a staged or reversible option that preserves future choices?

5. Who develops the recommendation, and who decides?

Who integrates the fact base and functional input into a formal recommendation?
Which functions provide Input?
Whose formal Agree is required?
Who holds Decide?
Who will Perform once the decision is made?
Is the Clinical Operations leader the recommender, process owner, execution owner or escalation initiator for this decision?
Where should the issue escalate if the current governance level cannot decide?

6. How will the decision enter the study and be tested?

Has the decision been recorded formally?
Have the integrated study plan, baseline and key milestones been updated?
Have country/site strategy, budget, supply forecast, service-provider direction or system plan changed where required?
Who owns each implementation action, and by when?
Which measures will show whether the decision worked?
What new evidence would invalidate the original assumptions?
Under what conditions should governance reopen the decision?

Part 03 · Decision framework

From project narrative to decision follow-through

Clinical studies do not produce a completely neutral presentation of facts. Each function selects information, interprets cause and effect, and recommends action through the lens of its remit, professional model and accountable risk.

Clinical Development may read enrollment variance as evidence of a genuinely scarce target population. Clinical Operations may see a mismatch between the protocol and real care pathways. Biostatistics may be concerned that change could affect endpoints or analysis. Regulatory Affairs may focus on the consequences for agency interaction. Each interpretation may contain valid evidence, but no single interpretation represents the whole study.

The concept of narrative is useful here, not because the project needs a more persuasive story, but because groups rarely act on isolated facts alone: facts must be arranged into a line of reasoning with sequence, causal interpretation, direction and implications for action before a group can decide and coordinate. The Clinical Operations leader should not protect one function's version. The role is to maintain a project narrative that can be corrected by facts, resolved through governance and tested through execution.

The framework has four steps:

Four-step path from project narrative to decision follow-through — **Figure 1.** From project narrative to decision follow-through — four steps: ① align on the situation → ② frame the decision → ③ complete the governance decision → ④ drive implementation and verification.

1. Align on the project situation

Build a shared fact base

The first task is not to force agreement. The first task is to make competing views visible, distinguishable and testable. The Clinical Operations leader separates information into four levels:

【1】confirmed facts, traceable to data, records, interviews or approved documents;

【2】professional judgments, through which functions interpret those facts;

【3】key assumptions, not yet proven but already influencing action;

【4】information gaps, including whether further information is worth obtaining.

The team also reconstructs the timeline: what was the original plan; when did the variance begin; which events changed the study trajectory; what actions have already been taken; did the results match the prediction behind those actions.

Outputs: shared fact base; study-event timeline; functional narratives and key assumptions; information gaps and evidence plan; impact on study objectives.

2. Frame the decision

Convert disagreement into a governance decision

Many cross-functional issues remain unresolved because nobody reframes the disagreement as an organizational trade-off. A decision-ready issue should include: a clear decision statement, the project objective the decision serves, the scientific / participant-protection / quality / regulatory guardrails, realistic options, the trade-offs each option protects and gives up, the latest responsible decision date, the cost of continued delay, and what further evidence is required (and whether waiting for it is justified).

Cynefin supports context assessment at this stage. The Clinical Operations leader should not label the entire issue with one domain. Clear, complicated, complex and chaotic elements should be separated and handled in different ways.

Outputs: clear decision statement; key trade-offs and guardrails; options; integrated impact assessment; context assessment and response; decision deadline and governance route.

3. Complete the governance decision

Bring input into clear decision accountability

Cross-functional work requires shared understanding, not unanimous enthusiasm for the final option. The organization needs alignment at four levels: acceptance of the fact base, a shared understanding of the question being decided, acceptance of who provides input, who recommends and who decides, and a clear execution commitment once the decision has been made.

RAPID distinguishes five decision roles: Recommend integrates facts and inputs into a formal recommendation; Agree provides required formal agreement within an accountable remit, or states why agreement cannot be given; Perform executes the decision; Input supplies professional facts, judgment and execution constraints; Decide makes the final decision and commits the organization to action. The acronym does not describe workflow sequence — in practice, Input informs Recommend, required Agree roles are addressed, Decide reaches the conclusion, and Perform carries the decision into execution. A critical decision should have one clear D wherever possible; Agree should remain limited to roles with genuine formal accountability for a relevant boundary.

Outputs: RAPID decision-rights map; formal recommendation and integrated impact assessment; governance decision; rationale, conditions and dissenting views; execution responsibility and escalation items.

4. Drive implementation and verification

Change how the study runs and learn from it

"Agreed" in meeting minutes does not mean the decision has taken effect. A decision becomes part of the project only after entering the plan, resources, documents, systems and operating cadence.

The Clinical Operations leader drives changes into the relevant workstreams: update the integrated study plan and baseline; adjust milestones, country/site mix and enrollment forecast; initiate protocol-amendment and impact-assessment processes where required; update budget, contracts and clinical-supply assumptions; update data, system, document and training plans; clarify execution requirements for CROs, service providers and sites; update the risk register, issue log and governance escalation route.

Every significant decision also needs feedback: which measures show whether the chosen direction is working; when the team will review the evidence; which results support continued execution; which results call for adjustment; what evidence would invalidate the decision basis; what conditions would send the issue back to governance.

Outputs: decision implementation plan; updated study plan and accountabilities; outcome measures and review points; triggers for reopening the decision; decision and implementation record.

Self-correcting project narrative loop — **Figure 2.** The feedback loop — execution evidence updates the project narrative. The objective is not to protect one story indefinitely; a strong project narrative can be revised when evidence changes.

Part 03 · Context assessment

Cynefin: assess the context before choosing the response

Cynefin does not supply the correct answer. The framework helps leaders recognize the situation they face and choose an appropriate way to analyze and act.

Five contexts and the right response

Clear. Cause and effect are understood; the organization has an established rule, process or standard practice. Response: sense the facts → categorize against the rule → respond through established practice. Clinical Operations examples include an agreed status report not submitted on time, an assigned action that remains overdue, a defined approval step missed, or an approved plan not followed. These issues call for execution and correction, not a new expert forum every time.

Complicated. Cause and effect can be analyzed, but the answer is not self-evident; several valid options may exist, and expert judgment is required. Response: gather evidence → conduct expert analysis → compare options → choose. Clinical-development examples include the effect of a protocol amendment on endpoints and sample size, the regulatory and timeline effect of adding a country, the effect of eligibility changes on the target population and interpretability, and the effect of a longer enrollment period on clinical supply and budget. The Clinical Operations leader does not perform every expert analysis; the role is to keep professional input focused on the same decision and integrate the analyses into comparable options.

Complex. Multiple actors interact, cause and effect cannot be known reliably in advance, and patterns may become clear only after action. Response: run bounded, safe-to-learn probes → observe feedback → amplify useful patterns or stop weak ones. Clinical Operations examples include why eligible participants are not entering the study, why investigators support the protocol but do not consistently refer participants, whether a different site-support approach will improve screening, and why the same protocol produces different operational outcomes across regions. A single meeting cannot discover one definitive answer; a better response may be to review the participant pathway at representative sites, test a limited operational intervention, define the measures in advance and use the evidence to decide whether to scale.

Chaotic. The situation is unstable, risk may be spreading, and there is no time for a complete analysis before action. Response: act to stabilize → sense the emerging situation → move into structured analysis. Clinical-development examples include an urgent event with potential ongoing participant-safety impact, a major compliance risk that continues to expand, a sudden loss of a critical system or clinical supply, and widespread loss of control over an important data or operational process. The immediate objective is to contain risk and protect participants and critical study assets; full cross-functional agreement comes after stabilization.

Confused. Most cross-functional issues begin here; the team has not yet determined which context applies. The Clinical Operations leader should split the issue into clear process failures, expert questions, elements that require learning through a probe, and elements that require immediate stabilization. Cynefin is useful because the framework reduces the risk of applying the wrong management response.

Part 03 · Decision rights

RAPID: separate participation from decision authority

The most common structural weakness in a cross-functional forum is not low attendance; the weakness is unclear decision accountability. RAPID assigns distinct roles for a specific decision.

Five decision roles

Recommend. R integrates the fact base, professional input and options, then submits a formal recommendation to D. A decision-ready recommendation explains the recommended option, why that option is preferred, which alternatives were considered, the impact and risk of each option, the assumptions behind the recommendation, and required resources and safeguards. For some operational decisions the Clinical Operations leader or project lead may hold R; for a decision centered on core scientific judgment, Clinical Development or another formally authorized role may hold R.

Agree. A applies only to roles whose formal accountability requires agreement. Depending on the decision, Clinical Development, Regulatory Affairs, Quality, Finance or another authorized function may need to confirm a matter within its remit. Not every decision requires every function to hold A. Too many A roles blur accountability; Agree protects a formal boundary and is not a general invitation to express a view.

Perform. P carries the decision into execution. Perform roles may include Clinical Operations, Data Management, Regulatory Affairs, Clinical Supply, CROs, other service providers or site-support teams. Perform should contribute to feasibility before the final decision; otherwise governance may select an option that is logically sound but operationally unworkable.

Input. I supplies professional evidence, experience and execution constraints that affect decision quality. Input should be heard and assessed, but not every comment must be adopted, and Input does not automatically carry veto authority. Recommend should show which inputs were adopted, which were not, and why.

Decide. D reaches the final conclusion and commits the organization to action. A critical decision needs a clear D. Depending on governance and the nature of the decision, D may sit with the Clinical Development lead, a program lead, senior management or a formal governance committee. The Clinical Operations leader may not hold D, but should know where D sits, what D needs in order to decide, which governance forum should receive the issue, and the latest responsible decision date.

Three principles for using RAPID

【1】Apply RAPID to a defined decision, not permanently to an entire function.

【2】Assign one clear D wherever possible.

【3】Clear roles do not replace a strong fact base, real options or an integrated impact assessment.

RAPID cannot replace professional judgment. RAPID prevents professional judgment from remaining trapped in discussion because decision rights are unclear.

Part 04 · Worked example

Enrollment remains below the baseline forecast: should the program initiate a protocol-amendment assessment and reshape country, site and recruitment strategy?

The example demonstrates the framework. No specific trial is represented, and no single answer is assumed.

Starting situation

A global multicenter phase III oncology clinical study has entered enrollment. Enrollment continues to fall below the baseline forecast. Some activated sites have not screened participants for an extended period. Other sites identify potential participants but show a high rate of screen failure. The study team has discussed the issue repeatedly without achieving sustained improvement.

Different functions have developed different explanations.

Several partial narratives

Clinical Development / Medical. The target population is inherently limited. Current eligibility criteria protect participant safety and population consistency. A poorly considered change could weaken scientific integrity and interpretability.

Clinical Operations. Some eligibility criteria may not fit real care pathways. Friction exists in participant identification, referral and screening. Adding more sites with the same profile may not improve enrollment.

Biostatistics. Any change to the study population, endpoint or sample size requires an assessment of statistical assumptions, analysis and interpretability.

Regulatory Affairs. A protocol amendment, country change or development-timeline change may affect agency interaction, submission content and approval timing.

Clinical Supply. Changes to the enrollment curve and study duration affect manufacturing, release, inventory, distribution and expiry management.

Management. The study is consuming budget and development time. Management needs a credible recovery path rather than continued waiting.

These narratives do not necessarily contradict one another. The real problem is that several explanations have not yet become one decision-ready issue.

Step one: align on the project situation

The Clinical Operations leader does not begin with "Should we amend the protocol?" The first task is to establish a cross-functional fact base. The fact base should include activation, screening and enrollment performance by country and site; participant funnel and major reasons for screen failure; differences in participant pathways across countries and sites; investigator feedback and the strength of supporting evidence; site-support and recruitment actions already attempted; observed results after those actions; current milestone, budget and supply implications; and any matter affecting participant protection or scientific integrity.

Key assumptions are listed separately: is the target population genuinely too small; are eligibility criteria the main barrier across different sites; does evidence support the belief that more sites will recover the gap; have recruitment-support measures already been tested properly; does the belief that a protocol amendment will improve enrollment overlook participant pathways and investigator behavior.

The team does not need immediate agreement on causation. The team does need to agree on confirmed facts, causal explanations that remain assumptions, information gaps worth addressing, and project objectives already affected by the enrollment variance.

Step two: separate the problem through Cynefin

Clear elements: overdue site actions; late status updates; approved participant-identification processes that were not implemented. These items receive immediate correction through existing mechanisms.

Complicated elements: the scientific, statistical and regulatory effect of a potential protocol amendment; the timeline, budget and supply effect of changing the country/site mix; whether the study baseline should be revised. The relevant experts conduct structured analysis.

Complex elements: why participants do not enter the study through real care pathways; why investigators do not identify or refer participants consistently; which site-support interventions can improve screening conversion. The team does not wait for a single final explanation. Representative sites are selected for bounded probes, such as participant-pathway reviews or different site-support approaches, with measures defined in advance.

Unless an urgent participant-safety, major compliance or supply-continuity issue is present, the enrollment issue should not be treated as wholly chaotic.

Step three: frame the decision

The team replaces:

Without weakening participant protection, scientific integrity or regulatory acceptability, should the program initiate a formal protocol-amendment assessment and, in parallel, adjust the country/site mix, recruitment strategy and study baseline?

The team develops options that governance can compare.

Option A: maintain the protocol and concentrate on execution. Increase support for high-potential sites; close or deprioritize persistently low-contributing sites; improve participant identification, referral and screening; observe results over a defined period.

Option B: change the country/site mix and operating strategy. Add countries or sites that better match the participant pathway; reallocate resources across the site portfolio; assess budget, supply and timeline implications.

Option C: initiate a protocol amendment. Assess potential changes across participant protection, science, statistics, regulatory strategy, data, supply and operations; define required approvals and regulatory steps; assess implementation time and opportunity cost.

Option D: use a staged dual-track approach. Correct clear execution failures immediately; run bounded probes for complex issues; complete integrated assessments for protocol amendment and country/site changes in parallel; make the next decision at a predefined governance point using the new evidence.

The options do not need to remain limited to these four. The important shift is from arguing over one function's preferred solution to comparing complete project paths.

Step four: complete governance through RAPID

The program assigns RAPID for this specific decision under the organization's governance model. The table below is illustrative.

Role	Possible responsibility in this example
Recommend	Integrate cross-functional input and produce a formal recommendation and option set. Depending on the governance structure, the project lead, Clinical Development lead or an authorized Clinical Operations leader may hold R.
Agree	Limited to roles with formal accountability for the relevant scientific, participant-protection, regulatory, quality or budget boundary.
Perform	Clinical Operations, Clinical Development, Biostatistics, Regulatory Affairs, Data Management, Clinical Supply and relevant CROs / service providers execute the decision.
Input	Provide evidence on site performance, participant pathways, statistics, data, regulatory impact, supply, safety, quality and budget.
Decide	The Clinical Development lead, program governance committee or another formally authorized governance body makes the final decision.

The Clinical Operations leader may add value by bringing site and participant-pathway evidence into the decision, coordinating the integrated impact assessment, keeping all professional inputs focused on the same decision statement, making the decision deadline and cost of delay visible, and ensuring that the selected option has an executable path.

Step five: carry the decision into the study

Assume governance selects a staged dual-track approach. The Clinical Operations leader drives immediate correction of clear execution failures, participant-pathway and site-support probes at representative sites, an integrated protocol-amendment assessment by Clinical Development, Biostatistics, Regulatory Affairs, Data Management and Clinical Supply, an updated country/site strategy and resource recommendation, a clear evidence package for the next governance decision, accountabilities, dependencies and timing in the integrated study plan, and predefined continue, adjust and escalate triggers.

Meeting completion is not the measure of success. Better questions are: has new site and participant-pathway evidence reduced a critical uncertainty; have the probes produced a signal that can be replicated; can governance now compare the integrated effects of amendment and country/site change; can the program make the next decision at the agreed governance point; and does the original project narrative need revision.

The example captures the central point of Scenario 05: the Clinical Operations leader does not replace each function's professional judgment. The role is to turn distributed evidence and judgment into a project action that can be decided, implemented and tested.

Part 05 · Key actions

Actions for the Clinical Operations leader

Build a shared fact base and align data definitions and status.
Identify each function's causal interpretation, key assumptions and risk concerns.
Rewrite functional conflict as a clear decision statement.
Use Cynefin to separate contexts instead of sending every issue through more meetings, more chasing or more expert analysis.
Develop options and coordinate an integrated cross-functional impact assessment.
Define RAPID roles, the governance forum and the escalation path.
Set a deadline for important decisions and make the cost of delay visible.
Translate governance decisions into plans, resources, documents, systems and service-provider actions.
Set outcome measures, review points and conditions for reopening the decision.
Preserve the rationale and later evidence so future teams can understand why the decision was made.

Part 05 · Decision boundaries

Boundaries the Clinical Operations leader should not cross

Cross-functional coordination does not authorize Clinical Operations to replace professional functions.

Functional boundaries

Clinical Development / Medical. Owns professional judgments on clinical science, benefit-risk, the study population and medical content of the protocol.

Biostatistics. Owns professional judgments on statistical assumptions, sample size, endpoints, analysis and interpretability.

Regulatory Affairs. Owns professional judgments on regulatory strategy, agency interaction, submission requirements and the regulatory impact of protocol changes.

Data Management. Owns professional judgments and execution concerning data collection, data flow, the database and related system changes.

Pharmacovigilance / Safety. Owns professional judgments related to safety signals, safety assessment and risk management.

Quality. Provides independent quality and compliance judgment and oversight under the organization's quality management system and the specific issue. Quality should not be presented as a fixed approver for every project decision.

Clinical Supply. Owns judgments on manufacturing, release, inventory, distribution, expiry and continuity of supply.

Clinical Operations / COD. Brings operational evidence, site and participant pathways, timelines, resources and execution feasibility into the decision. Clinical Operations coordinates cross-functional dependencies and drives governance decisions into the study with visible feedback.

Related capability domains

Project narrative and issue framing · Cross-functional governance · Decision-rights design · Integrated impact assessment · Protocol-feasibility judgment · Decision escalation · Governance-forum design · Stakeholder management · Judgment under complex conditions · Implementation follow-through and feedback.

Part 05 · Tools

Tools that support this scenario

Cross-functional decision brief

Convert an issue discussed repeatedly without resolution into a complete record of facts, judgments, assumptions, decision statement, guardrails, options, deadline and governance level.

Cross-functional integrated impact assessment

Evaluate a protocol amendment, country/site strategy change or study rebaseline across the functions it actually affects.

Cynefin context assessment

Split a mixed issue into clear, complicated, complex and chaotic subproblems and match the response to each one.

RAPID decision-rights map

Assign Input, Recommend, Agree, Decide and Perform for a specific decision so the meeting does not stop at discussion.

Decision and implementation follow-through log

Capture decision, rationale, roles, dependencies, review points and reopening conditions so later teams can understand the call.

Part 05 · Conceptual foundations

How narrative, Cynefin and RAPID belong together

The three concepts answer different questions. Narrative helps the organization understand the project situation. Cynefin matches the response to the context. RAPID completes the governance decision. Execution feedback then updates the original narrative.

Conceptual source

Narrative is not fiction. Narrative arranges facts into a structure with time, cause, direction and implications for action. In clinical development, functions develop different project narratives through their professional models and accountabilities. The Clinical Operations leader should not eliminate those differences; the work is to bring partial narratives into a project line that evidence can test, project objectives can guide, and a group can use to act. Three functions of narrative are especially relevant: prediction (explain how the project reached the current state), meaning (clarify the objective the organization now needs to protect), and public coordination (allow functions to act around a shared decision).

Cynefin shows why clear, complicated, complex and chaotic contexts require different responses. In Scenario 05, Cynefin helps prevent using SOPs and chasing to address complex human behavior, delaying a clear execution correction through endless analysis, waiting for complete agreement during an urgent risk, and forcing a mixed issue into one category.

RAPID addresses decision accountability, not professional judgment. RAPID separates "everyone is involved" into who provides essential input, who integrates the input into a recommendation, whose formal agreement is required, who closes the discussion with a decision, and who turns the decision into an operational result. RAPID cannot replace professional judgment. RAPID prevents professional judgment from remaining trapped in discussion because decision rights are unclear.

Robert J. Shiller

Economist and author of Narrative Economics. Studies how narratives spread and influence collective expectations and behavior.

Dave Snowden

Originator of the Cynefin framework. Long focus on sensemaking and decision-making in complex environments.

Paul Rogers & Marcia Blenko

Important exponents of RAPID decision accountability. Authors of Who Has the D? How Clear Decision Roles Enhance Organizational Performance.

References

Dennett, D. C. (1991). Consciousness Explained. Little, Brown and Company.
Shiller, R. J. (2019). Narrative Economics: How Stories Go Viral and Drive Major Economic Events. Princeton University Press.
Kurtz, C. F., & Snowden, D. J. (2003). The New Dynamics of Strategy: Sense-making in a Complex and Complicated World. IBM Systems Journal, 42(3), 462–483.
Snowden, D. J., & Boone, M. E. (2007). A Leader's Framework for Decision Making. Harvard Business Review, 85(11), 68–76.
Rogers, P., & Blenko, M. (2006). Who Has the D? How Clear Decision Roles Enhance Organizational Performance. Harvard Business Review, 84(1), 52–61.
万维钢：《现代思维工具 100 讲》，"叙事"相关课程内容。 This page draws only on the course's high-level ideas about prediction, meaning and public coordination and does not reproduce the course text.

Cross-functional alignment and decision follow-through