K
KnowMBAAdvisory
Change ManagementAdvanced7 min read

Dual Operating System

John Kotter's Dual Operating System argues that established organizations need to run two structures simultaneously: (1) the traditional hierarchy — the org chart, the budgets, the management routines that keep the existing business running reliably — and (2) a network — a parallel, voluntary, cross-functional 'change army' that pursues strategic transformation at the speed the market demands. The core insight: hierarchies are excellent at execution and terrible at large-scale change. Networks are excellent at speed and adaptability but terrible at running daily operations at scale. Both are required. Most companies attempt to do transformation 'through the line' (using the hierarchy alone), which moves at the speed of budget cycles and reporting layers — far too slow for the digital era. The dual system uses the network for the change work and the hierarchy for the run-the-business work, with deliberate connective tissue between them.

Also known asKotter Dual SystemHierarchy Plus NetworkNetwork-Hierarchy Hybrid

The Trap

The first trap is creating a network that's actually just another hierarchy in disguise — a transformation steering committee staffed by the same VPs who run the business, meeting once a month, generating slide decks. That's not a network; it's a slow committee. A real network is voluntary, has 50+ members across levels and functions, and operates with much faster decision cycles than the hierarchy. The second trap is building a strong network with no connective tissue back to the hierarchy — the network produces brilliant ideas that the line organization rejects because they 'weren't involved' or 'don't have budget.' The third trap: leaving the hierarchy untouched. The network can innovate all it wants, but if the hierarchy continues to reward old behaviors and block resources, nothing ships.

What to Do

Stand up the network with explicit principles: voluntary participation across levels (not just senior people), 8 'accelerators' (Kotter's specific roles: urgency, guiding coalition, vision, communicate, enable, generate wins, sustain, institutionalize), regular cadence (weekly or bi-weekly, not monthly), and visible executive sponsorship. Critically, design the connective tissue: a small steering team where senior hierarchy leaders meet with network leaders, decisions about resourcing get made jointly, and network outputs get formally adopted into hierarchy plans. Measure the network on speed of decision and number of cross-functional barriers removed, not on number of meetings held.

Formula

Transformation Velocity ≈ (Network Size × Network Cycle Speed) ÷ Hierarchy Friction — increasing network reach matters, but reducing hierarchy friction matters more

In Practice

ING Bank's 2015 'agile transformation' is one of the clearest dual operating system implementations at scale. Rather than fully reorganizing the bank, ING kept its core hierarchy intact for compliance, risk, and operations — and stood up a parallel agile network of ~350 'squads' organized into 13 'tribes' working on customer-facing innovation. The tribes operated with their own leadership, cadence, and decision rights, but were connected to the hierarchy through a senior leadership team that funded and prioritized them. The result: innovation cycle time dropped from quarters to weeks for digital products, while the regulated banking core continued to operate within its required hierarchical controls. ING's model became a widely-studied template for how large regulated companies can pursue transformation without abandoning the structures their core business requires.

Pro Tips

  • 01

    The network must be voluntary. The moment you mandate participation, you've recreated the hierarchy. Voluntary participation filters for the people who actually want to drive change — and their energy is the engine of the whole system. Mandated participants attend meetings; volunteers ship work.

  • 02

    Senior executive sponsorship is non-negotiable. Without an active CEO or COO sponsor with weekly engagement, the network becomes a side project that the hierarchy will eventually starve of resources. Kotter's research is clear: the difference between dual systems that work and ones that fade is sustained executive air cover.

  • 03

    The most-overlooked failure mode is middle-management championship. Senior executives sponsor the network, frontline volunteers staff it — but the people who control day-to-day resources, time, and political support are middle managers. If they see the network as a threat or distraction, they'll quietly block participation by overloading their teams. Bring middle managers into the network design, not just the rollout.

Myth vs Reality

Myth

The network replaces the hierarchy

Reality

Kotter is explicit: the dual system is 'and,' not 'or.' Hierarchies are extraordinary inventions for running large reliable operations. The network supplements the hierarchy for transformation work; it does not replace it. Companies that try to flatten everything into a network usually lose their ability to operate reliably at scale.

Myth

Agile or holacracy is the same thing as a dual operating system

Reality

Agile and holacracy are organizational design philosophies typically applied to a whole unit or company. The dual operating system is a deliberate two-structure design where one part stays hierarchical. Many 'agile transformations' fail because they attempt full conversion of the hierarchy rather than running a parallel network.

Try it

Run the numbers.

Pressure-test the concept against your own knowledge — answer the challenge or try the live scenario.

🧪

Knowledge Check

A 25,000-person bank stands up a 'Digital Transformation Office' staffed by 12 senior VPs from the existing hierarchy who meet monthly. After 18 months, only 3 minor digital initiatives have shipped. Why is this not a true dual operating system?

Industry benchmarks

Is your number good?

Calibrate against real-world tiers. Use these ranges as targets — not absolutes.

Transformation Network Size (% of Total Workforce)

Enterprise transformation programs (Kotter benchmark)

Robust Network

5-15% of workforce voluntarily engaged

Functional Network

2-5% engaged

Underpowered (Common)

< 1% (steering committee only)

Symbolic

12-person committee, monthly cadence

Source: John Kotter, 'Accelerate' (HBR Press, 2014)

Real-world cases

Companies that lived this.

Verified narratives with the numbers that prove (or break) the concept.

🏦

ING Bank

2015-2018

success

ING faced the classic large-bank transformation dilemma: how to compete with digital-native fintechs while maintaining the regulatory and operational rigor banking requires. Rather than reorganizing the entire bank, ING stood up a parallel agile structure of 350+ squads grouped into 13 tribes, focused on customer-facing innovation. The traditional hierarchy continued to run risk, compliance, and core operations. The two systems were connected via a senior leadership team that funded and prioritized tribes and ensured that what the squads built was operationalized through the hierarchy. Innovation cycle time dropped from quarters to weeks for digital products without compromising the regulated core.

Squads created

350+

Tribes (functional clusters)

13

Digital product cycle time

Quarters → weeks

Hierarchy maintained for

Risk, compliance, operations

ING demonstrated that you can run a fast-moving network on top of a regulated hierarchy without breaking either. The key was deliberate connective tissue — funding flows, decision rights, and operational handoffs were explicitly designed, not left to chance.

Source ↗
🎵

Spotify (cautionary)

2012-2020 (Squad Model evolution)

mixed

Spotify's 'squads, tribes, chapters, guilds' model became globally famous and was widely copied as the gold standard of network organization. But internally, Spotify itself moved away from the literal model over time, with engineers later writing publicly that the model worked when Spotify was 200 engineers but became increasingly dysfunctional at 1,000+. The root issue: the network worked, but Spotify under-invested in the connective tissue back to traditional hierarchical structures (career paths, cross-squad coordination, technical standards). Many companies that copied 'the Spotify model' missed that Spotify had also rebuilt its hierarchy — they only copied the network, leaving them with a disorganized tribe structure and no functioning hierarchy at all.

Companies that copied the model

Hundreds globally

Engineers at peak of model success

~200

Engineers when problems emerged

1,000+

Lesson

Network without hierarchy fails too

A dual operating system is dual. Companies that adopted 'the Spotify model' as a hierarchy replacement (not a hierarchy supplement) created chaos. The lesson isn't that networks fail — it's that they require the hierarchy to remain functional. You need both halves of the dual system, not just the cool half.

Source ↗

Related concepts

Keep connecting.

The concepts that orbit this one — each one sharpens the others.

Beyond the concept

Turn Dual Operating System into a live operating decision.

Use this concept as the framing layer, then move into a diagnostic if it maps directly to a current bottleneck.

Typical response time: 24h · No retainer required

Turn Dual Operating System into a live operating decision.

Use Dual Operating System as the framing layer, then move into diagnostics or advisory if this maps directly to a current business bottleneck.