The Coordination Stack: The Five Questions Behind Coordinated Work
Ask a leadership team why a project slipped and "coordination" will usually come up within the first few minutes. Ask them what coordination actually consists of, and the answers get vague: better communication, more alignment, fewer silos.
The Coordination Stack™ is one attempt to make the word concrete. It treats coordination as a set of specific questions that either have answers or don't. This article explains what the framework is, how each of its five layers works, how practitioners apply it, and where it sits relative to other well-known work management frameworks.
What is the Coordination Stack?
The Coordination Stack™ is a system-level work management framework that defines five questions every piece of shared work must answer before people, teams, or AI agents can coordinate it reliably. It was developed by Brandon Hatton and is formalized and stewarded by the Work Management Institute (WMI).
The canonical definition reads:
The Coordination Stack™ is a foundational work management framework that defines the five essential coordination questions—why, what, who, when, and how—that must be answered for people, teams, or AI agents to reliably coordinate work across its lifecycle.
The framework's central claim is simple: coordination fails when one or more of those questions goes unanswered. It does not prescribe a tool, a methodology, or a meeting cadence. It describes the conditions any of those things must satisfy to work.
The five layers
Each layer of the stack corresponds to one coordination question. Together they move from intent to execution.
In order, the five layers are Why, What, Who, When, and How.
Why — Purpose & Intent
The Why layer asks why the work exists and what outcome it is meant to achieve. It establishes intent through desired outcomes, success criteria, strategic intent, and clear problem statements. It is also the layer most often assumed rather than stated, because the person requesting the work knows the reason and forgets that others don't.
When Why is unclear, work drifts and value erodes. A team can hit every deadline and still deliver something no one needed.
What — Work Definition & Scope
The What layer asks what exactly needs to be done and what "done" means. It translates intent into concrete scope through tasks and deliverables, scope boundaries, acceptance criteria, and definitions of done. Its most important element is a shared picture of the finished result, since two people can agree on a task name and still imagine very different outcomes.
When What is unclear, rework, duplication, and misalignment increase.
Who — Ownership & Responsibility
The Who layer asks who owns the work, who contributes, and who decides. It makes accountability explicit through named owners and assignees, decision makers, contributors and reviewers, and the accountability model a team uses. A strong Who layer separates the person accountable for the outcome from the people who help produce it.
When Who is unclear, coordination slows and responsibility diffuses. Work waits in the gaps between people, each assuming someone else has it.
When — Timing & Dependencies
The When layer asks by when the work needs to happen and in what order. It governs flow through deadlines and milestones, dependencies, sequencing and prioritization, and the cadence a team works on. In cross-functional work, sequencing often matters more than any single deadline, because a late handoff upstream moves every date downstream.
When When is unclear, teams experience delays, thrash, and missed commitments.
How — Execution & Alignment
The How layer asks how the work is carried out and how people stay aligned as it progresses. It defines the execution environment: communication practices, collaboration tools, norms for updates and decisions, and the workflows work moves through day to day. It is usually the most visible layer, and often the first place organizations invest when coordination feels strained.
When How is unclear, even well-defined work becomes chaotic.
Reading the stack as a system
The layers are described as reinforcing, which means a gap in one tends to show up as a problem in another. The framework's own guidance is that a breakdown at a higher layer often traces back to an unresolved lower layer.
This is the most practical idea in the framework. Problems are usually noticed at the How and When layers, because that is where work visibly stalls: a missed deadline, a confusing handoff, a status meeting that runs long. Teams then try to fix those layers directly, with a new tool, a tighter schedule, or more check-ins. If the real gap is an undefined scope or an unstated purpose, those fixes treat the symptom and the problem returns on the next project.
How practitioners use it
The Coordination Stack is described as a diagnostic and design framework, not a methodology. In practice it is applied in four ways:
Diagnosing breakdowns. Walk a stalled initiative through the five questions and note which ones have no clear, shared answer.
Designing work systems. Use the layers as a checklist when setting up a new project template, intake form, or recurring workflow, so each question has a defined home.
Evaluating tools and processes. Ask which layers a given tool or process actually supports. Most task tools are strong on What, Who, and When; few hold Why well.
Supporting human, AI, and hybrid work. Apply the same five questions whether the participant receiving the work is a person or an AI agent.
A worked example
A marketing team is three weeks late launching a product page. The visible symptom is a timing problem: the page missed its date. Walking the stack tells a different story.
How: Updates were scattered across chat threads and email. A symptom, but not the root.
When: The design review happened after copy was finalized, forcing a rewrite. A sequencing gap.
Who: No one was named as the final approver, so the page circulated through four reviewers.
What: Reviewers disagreed on whether the page needed pricing and a comparison table. Scope was never written down.
Why: The page's purpose, lead generation or sales enablement, was never stated. That ambiguity is what produced the scope disagreement.
The team could have bought a better chat tool or tightened deadlines. Neither would have fixed the unstated purpose, which is where the delay started.
How it relates to other work management frameworks
Most familiar frameworks go deep on one or two of the stack's questions. The Coordination Stack works at a broader level: rather than replacing them, it shows which coordination question each one answers and which ones still need a home.
OKRs are built for setting and measuring objectives, so they live mostly in the Why and What layers. RACI assigns roles on tasks and decisions, which makes it a Who tool. Kanban visualizes work in progress and limits how much is underway at once, supporting the When and How layers. Critical path and Gantt scheduling focus on sequencing and dependencies, the core of the When layer.
Personal systems like Getting Things Done (GTD) work at the individual level, mostly within What and How. A team that uses several of these together can map each one onto the stack and quickly see where coverage overlaps and where a layer has no owner.
The resemblance to 5W1H (who, what, when, where, why, how) is worth naming directly. The Coordination Stack draws on the same instinct to ask basic questions, but differs in three ways: it is scoped specifically to coordinating shared work, it orders the questions as layers running from intent to execution, and it has no separate "where" question.
Within WMI's own body of knowledge, the stack is positioned alongside two companion frameworks. The C4 Flywheel™ describes how work builds momentum through Clarity, Coordination, and Completion. The Work Value Pyramid™ evaluates what kind of value the work produces. In that grouping, the Coordination Stack answers the prior question: whether the work can be coordinated at all.
Coordination with AI agents
The framework's definition explicitly includes AI agents alongside people and teams. That inclusion matters more each year, because agents expose coordination gaps that humans quietly paper over.
A human colleague handed a vague request will often fill in the missing Why from hallway context, or ask a clarifying question. An agent typically acts on what it was given. If the Why is missing, it optimizes for the literal instruction. If the Who is missing, no one reviews its output. If the When is missing, its work lands out of sequence.
Used this way, the five questions work as a minimum handoff standard for delegating work to an agent: state the purpose, define done, name the human owner and approver, set the timing and dependencies, and specify where the output goes and how it is reported.
What the Coordination Stack is not
Part of what makes the Coordination Stack easy to adopt is how focused it is. It is designed to work alongside the methods and tools a team already has.
It is not a methodology. It does not tell a team how to run sprints, hold standups, or structure a project plan. Teams still need a delivery approach.
It is not a maturity model. It identifies which questions are unanswered, not how advanced an organization's coordination is overall. Maturity models address that separately.
It is not a tool category. Tools, processes, and governance support the layers, but they are not the layers themselves. A tool that stores a due date has not answered When if the dependencies behind that date are unknown.
It does not measure quality of answers. The stack checks whether each question has a clear, shared answer. Whether the answer is a good one, a wise strategy or a realistic deadline, still depends on judgment.
That focus is what lets the stack pair cleanly with whatever delivery methodology, maturity model, or toolset an organization already uses.
Frequently asked questions
What are the five layers of the Coordination Stack?
Why (purpose and intent), What (work definition and scope), Who (ownership and responsibility), When (timing and dependencies), and How (execution and alignment).
Who created the Coordination Stack?
It was developed by Brandon Hatton and is formalized and stewarded by the Work Management Institute.
Is the Coordination Stack a replacement for RACI or OKRs?
No. RACI and OKRs go deep on individual layers (Who, and Why/What respectively). The Coordination Stack is a broader check that shows which layer each of those tools is covering and which ones are left open.
Can it be used by a single team, or only across an organization?
Both. The same five questions apply to one task, a team's recurring workflow, or a cross-functional program. The scale changes how detailed the answers need to be, not the questions.
Does it apply to AI agents?
Yes. Its canonical definition names people, teams, and AI agents as participants. Agents tend to make gaps in the stack more visible because they act on what they are given rather than filling in missing context.
