The matrix explained

A matrix organization design is a structural framework in which employees report to more than one manager—typically both a functional manager (e.g., head of engineering, marketing, or finance) and a project or product manager responsible for a specific initiative or client. This approach aims to improve flexibility, facilitate efficient resource allocation, and encourage collaboration across different parts of an organization.

Key Features:

Dual (or multiple) Reporting Relationships: Employees have two bosses: a functional manager and a project manager.
Shared Authority: Decision-making is distributed among different managers, leading to more balanced choices.
“Resource” Sharing: Personnel and resources are shared across projects and functions to maximize expertise.
“Enhanced” Communication: The structure encourages communication and coordination across traditional departmental boundaries.

Primary Sources:

Jay R. Galbraith, “Matrix Organization Designs: How to Combine Functional and Project Forms” (1971):
- Galbraith formalized the matrix structure concept, describing it as a response to increasingly complex projects and the need for flexibility.
- Quote: “A matrix organization is any organization that employs a multiple command system that includes not only the multiple command structure, but also related support mechanisms and an associated organizational culture and behavior pattern.”
Harold Kerzner, “Project Management: A Systems Approach to Planning, Scheduling, and Controlling” (1979):
- Kerzner discusses matrix organizations as a solution for companies handling multiple, complex projects.
- Quote: “The matrix organization attempts to combine the best of both the functional and projectized structures to facilitate the responsive use of people and resources.”
Lawrence & Lorsch, “Organization and Environment” (1967):
- Their research on differentiation and integration in organizations laid important groundwork for the matrix concept, observing companies balancing functional specialization with project-based integration.

References:

Galbraith, J. R. (1971). Matrix Organization Designs: How to Combine Functional and Project Forms. Business Horizons, 14(1), 29–40.
Kerzner, H. (1979). Project Management: A Systems Approach to Planning, Scheduling, and Controlling. Van Nostrand Reinhold.
Lawrence, P. R., & Lorsch, J. W. (1967). Organization and Environment: Managing Differentiation and Integration. Harvard Business School Press.

One Antidote: an archipelago of adaptiveness

My InfoQ article on “islands/archipelagos of agility” argues that when large organizations are bogged down in bureaucracy and legacy culture, creating a focused “island” that models organizational adaptiveness can be a powerful way to trigger broader change.

Core idea: islands of adaptiveness

The article proposes an “island of adaptiveness”: a substantial team-of-teams that works semi-independently on a significant product or problem, using genuinely adaptive ways of working and modern technical practices. This island operates in a “sea of less adaptive teams” and is designed to be permanent, gradually influencing the wider organization rather than being reabsorbed and diluted by old habits.

Why islands are needed

Many organizations face strong delivery pressure, superficial “transformation initiatives,” and measurement systems that reward local efficiency over real customer value, hindering meaningful adaptability. Crises such as disrupted markets, sustainability demands, or failed past changes can create urgency, but entrenched culture and structures often prevent adaptive behavior from taking root across the whole enterprise.

Bubbles, isolation, and islands

The article contrasts three patterns: culture bubbles, isolation patterns, and islands of adaptiveness. Culture bubbles and isolation teams can create temporary protected spaces, but they tend to be small, fragile, and still dependent on the “mainland,” whereas an island of adaptiveness has a larger scope and enduring mandate, so it can demonstrate systemic adaptiveness and radiate it outward.

Designing an island of adaptiveness

On an island of adaptiveness, people volunteer rather than being assigned, teams self-design, leadership is distributed, and work is organized around discovery, delivery, and value validation rather than functions or projects. The island deliberately eliminates siloed structures, coordination chaos, and “adaptiveness in a box” thinking, and instead focuses on customer-informed learning, flow, and human-centric interactions.

Protecting the archipelago

Over time, multiple islands form an “archipelago of adaptiveness,” taking on more work as the mainland slowly hands over value streams. To prevent cultural unraveling, work migrates to the archipelago rather than fully re-merging structures, while harmful mainland patterns such as misaligned metrics, NoNos (chronic naysayers), lip service, and competition-focused design are kept from polluting the adaptive culture.

Practical Starting Points

The article closes with practical guidance: co-create an inspiring purpose, educate curious volunteers and an executive sponsor in authentic adaptive approaches, and design measurement, funding, and organization structures to support systemic adaptiveness. The island then improves and expands carefully through intentional retrospectives and marginal gains (perhaps with nested PDSA loops), accepting that authentic, sustainable adaptiveness is hard but offers hope when traditional transformation efforts have failed.

Source:

https://www.infoq.com/articles/need-island-agility/

Plan Do Study Act

An evolution from the thought-leading works of W. Edwards Deming et al, inspired by works by Walter A. Shewhart et al

Adaptations:
1). Observe current condition as part of the Plan step
2). Enter or exit PSDA at any point,
3). Cultivate the work environment and climate, and
4). Sometimes replace “Act” with “Adjust” or “Adapt”

An evolution from the thought-leading works of W. Edwards Deming et al, inspired by works by Walter A. Shewhart et al

Adaptations:
1). Enter or exit PSDA at any point,
2). Cultivate the work environment and climate, and
3). Sometimes replace “Act” with “Adjust” or “Adapt”

The goalposts should move if evidence tells one the goalposts are in the wrong place; this is why goalposts and goals should be malleable

Through intentional pause and reflection, avoid execution bias (210) - the human tendency to persevere when one should pivot or stop

Better Attribution

Nested PDSA loops

What has changed?

For a start, there is a vacuum (”Agile” died, leadership died). And vacuums usually fill, not always with the best options.
AI has taken off, with some amazing results, a lot of over-trusted content, and a lack of “four eyes” monitoring/auditing of results.
Mob programming with AI in smaller “one pizza teams” is emerging as a success pattern, which is great for focus, exhilaration, and learning, but exhausting for introverts.
A trend of Product Developers using AI agents as teammates; Dave Farley, amongst others, is critical of the code quality produced and the side effects of AI-driven code changes, while others claim it’s under-hyped.
People are referred to even more as resources; humanity has exited the workplace; check out the latest Gallup annual report - we had the great resignation and now that great detachment.
Plan Do Study Act is back; some would argue it never went away
The matrix is where flow goes to die; no one in the matrix can make decisions on their own, so nothing much changes.

https://www.gallup.com/workplace/653711/great-detachment-why-employees-feel-stuck.aspx

Long story short, we’re maybe one step forward and two steps back.

Coping Strategies for Bubbles or Islands

In a bubble, the leadership team becomes more of a team:

There should be coopertition (cooperation and friendly competition) to prevent the bubble/island from being absorbed by the mainland.
The bubble/island requires at least 3 days of on-site work per week to foster focus and intentional collaboration.
It has a unifying, continuous, emergent strategy informed by evidence.
It has a team agreement that outlines how decisions are made.
It commits to avoiding the five dysfunctions of a team and to the leadership team being the team, not each matrix-driven functional silo; therefore, sometimes decisions are made that are good for the bubble/island, and a functional manager will suck them up for the bubble/island's good.
Agree with functional managers to allocate individual objectives to 50% organizational performance, 30% team-of-teams/product performance, 15% functional performance, and 5% individual performance. Why? Because perhaps over 95% of performance is constrained by systemic issues. OK, we could argue the numbers, and people do like to compete, but I hope you catch my drift.
It has a nested PDSA cycle for addressing system issues, owned by or devolved to the leadership team (delegation is too weak and temporary).
Have people “from the business” as well as digital folks in the teams so that the mainland does not infect the bubble/island.
Adaptiveness Guides with "skin in the game" (such as Nader or people of that ilk; "bubble gum" “agile coaches” or “product coaches” would not suffice) are in the frame as real change agents. They help managers, product leaders, and colleagues to become ambidextrous change agents.
Work and waste are visualized; you would be amazed at what conversations get triggered by posting obstacles and work on a physical wall, done in a way that is also editable digitally. We solved that using digital whiteboard forms and printing them on the wall, so that we have information radiators, not information refrigerators.
The leadership team agrees to act like the Avengers and close shields to deflect excessive silo pressure; there needs to be negative consequences for functional managers who break away into their own silo.
Form an archipelago of bubbles/islands and a network leadership team. The Avengers get stronger through evidence of results, delivering meaningful outcomes, and breaking dependencies on the mainland. Rinse and repeat.
Never subsume the archipelago under the mainland; friendly competition is preferable. If you’re to be naughty, encourage the mainland’s centres of excellence to grow so the archipelagos win even more.
Or start a new P&L, a totally separate, independent organization, not even a division. It’s difficult to disrupt from the inside.

Conclusion

The matrix must evolve. Networks result and thrive.

This article suggests only one antidote among several. Change agents are not clones. See Nader's unique take in his articles on this site.

Busting the Matrix