Swarming is coordinated action by many dispersed units that converge on a target or objective without centralised command. No single node directs the whole; the units share a goal, sense the environment, and act in concert through local communication and simple rules. The result is resilience (no single point of failure), speed (decisions happen at the edge), and adaptability (the swarm can shift as the situation changes).
The model comes from nature (bees, ants) and from military and insurgent doctrine: distributed forces that strike from many directions, regroup, and strike again. In business it applies to distributed teams, ecosystem competition, and go-to-market where many small actors coordinate without a central controller. The strategic question is when to favour swarm structure (when speed, resilience, and local information matter) over hierarchy (when coordination cost or consistency of message demands central control).
Use the model when you face or deploy many coordinated but not centrally directed units — partners, teams, or nodes in a network — and when the environment is fast-changing and local information is decisive.
Section 2
How to See It
Swarming reveals itself when many actors achieve coordinated effect without a single boss. Look for: shared objective and simple rules; local decision authority; and communication that is horizontal (node to node) as well as or instead of vertical (leader to node). The diagnostic: could the system still function if the "center" were removed? If yes, you are seeing swarm logic.
Business
You're seeing Swarming when a company's go-to-market relies on many partners, affiliates, or channel nodes that each pursue the same goal (e.g. lead gen, adoption) with local autonomy. There is no single command post; coordination comes from shared incentives, data, and lightweight rules. The swarm can adapt to local conditions faster than a central team could direct.
Technology
You're seeing Swarming when a platform or protocol is designed so that many independent nodes (developers, validators, users) contribute to a common outcome — security, liquidity, or content — without a central controller. The system behaves like a swarm: resilient, adaptive, and hard to kill by attacking one node.
Investing
You're seeing Swarming when a category has many small players that collectively shift the market — e.g. indie developers, micro-influencers, or long-tail providers — and the investable question is whether their collective behaviour will converge on a new standard or winner. The swarm is the market structure.
Military
You're seeing Swarming when a force operates in small, dispersed units that can concentrate on a target when needed, then disperse again. Each unit has enough information and authority to act; they coordinate through shared situational awareness and simple rules (e.g. "attack when you have local superiority"). No single commander directs every move.
Section 3
How to Use It
Decision filter
"When coordination could be centralised or distributed, ask: do we need a swarm? If speed, resilience, and local information matter more than tight consistency, design for swarm — shared goal, simple rules, local authority, horizontal communication. If we need one message, one plan, or tight control, prefer hierarchy."
As a founder
Use swarm structure where it beats hierarchy: product teams that can ship without waiting for central approval, partner ecosystems that sell and support without your direct control, or community-driven growth where many nodes spread the message. The discipline is giving the swarm a clear objective and simple rules, then getting out of the way. The mistake is calling something a swarm while still requiring central approval for every move. The second mistake is using a swarm when you need one coordinated story (e.g. crisis comms, regulatory response); then centralise.
As an investor
Assess whether the company's strategy depends on swarm dynamics — many small units or partners that collectively achieve scale. If yes, evaluate whether the design supports real decentralisation (authority, information, incentives at the edge) or is hierarchy in disguise. Swarm can be a strength when the environment is uncertain and local action matters; it can be a weakness when consistency and control are required.
As a decision-maker
When you lead a swarm, your job is to set the objective, establish the rules, and ensure information and authority flow to the edge. Do not micromanage; do not become the bottleneck. When you face a swarm (e.g. a distributed competitor or an ecosystem), recognise that there is no single head to cut off; you must out-compete the collective behaviour or change the rules that shape it.
Common misapplication: Calling a structure a swarm when decisions are still centralised. If every node needs approval from the center, it is not a swarm; it is a hub-and-spoke with delay. Real swarms have authority at the edge.
Second misapplication: Using swarm when the objective requires one voice or one plan. Swarms are good at adaptation and coverage; they are bad at delivering a single, consistent message or a tightly coordinated sequence. Match the structure to the task.
Hastings built Netflix around distributed decision-making: small teams with clear scope and authority, "context not control," and information flowing to the edge. The company behaves like a swarm of product and regional units that share the same goal (engagement, retention) but make local choices. The lesson: give teams the context and authority to act; avoid becoming the bottleneck.
Grove pushed decision authority down at Intel: "only the paranoid survive" meant every unit had to sense and respond to threats. The company combined strategic direction from the top with swarm-like responsiveness at the business-unit level. The lesson: you can have a center that sets direction while still enabling swarm behaviour in execution.
Section 6
Visual Explanation
Swarming — Many nodes, shared goal, simple rules. No single controller; coordination emerges from local action and communication.
Section 7
Connected Models
Swarming sits with decentralisation, OODA loop, and network effects. The models below either explain the structure (decentralisation, team of teams), the advantage (OODA, asymmetric warfare), or the dynamics (network effects, distributed systems).
Reinforces
Decentralization
Decentralisation is the distribution of authority and information away from a single center. Swarming is decentralisation applied to action: many nodes act with local authority. The reinforcement: both favour resilience and speed at the edge over central control.
Reinforces
OODA Loop
Swarms can cycle through Observe-Orient-Decide-Act faster because each node does it locally; there is no wait for central approval. The reinforcement: swarm structure is one way to compress the OODA loop at the organisational level.
Reinforces
Asymmetric Warfare
Weaker sides often use swarm-like structure: dispersed, resilient, hard to target. Asymmetric warfare and swarming overlap when the underdog avoids concentration and uses many small units to harass or outlast a stronger, centralised opponent.
Leads-to
Network Effects
Networks can behave like swarms: many nodes (users, developers) contribute to a common outcome (value, content) without central direction. The connection: platform and protocol design that enables swarm behaviour often amplifies network effects.
Reinforces
Section 8
One Key Quote
"The key distinction is between hierarchies and networks... In the network form, there is no single central leader... The network has a flat structure with multiple nodes and links."
— John Arquilla & David Ronfeldt, Networks and Netwars (2001)
Swarming is the network form in action: no single leader, many nodes, many links. Coordination comes from shared purpose and from the structure of the network, not from a chain of command.
Section 9
Analyst's Take
Faster Than Normal — Editorial View
Real swarms have authority at the edge. If every decision still goes through a central approval, you have delay and bottleneck, not swarm. The test: can a node act on its own within clear bounds? If not, push authority and information down until the answer is yes.
Simple rules and a shared goal are the glue. Without a clear objective and a few rules (e.g. "do not conflict with other nodes on the same customer," "escalate only when X"), the swarm becomes chaos. Invest in clarity of purpose and in the minimal rules that prevent conflict and enable convergence.
Use swarm where the environment is uncertain and local. When the best information is at the edge and the situation changes fast, swarm wins. When you need one message, one plan, or one negotiation position, hierarchy wins. Do not default to one or the other; match the structure to the problem.
Facing a swarm: do not expect a single knockout. The opponent has no single head. You must either out-compete the collective (better product, better economics) or change the rules that shape their behaviour (e.g. platform policy, standards). Decapitation does not work against a swarm.
Section 10
Test Yourself
Is this mental model at work here?
Scenario 1
A company has 20 regional teams that each set local pricing and partnerships; HQ sets overall targets and guardrails but does not approve each deal.
Scenario 2
An open-source project has hundreds of contributors; there is a maintainer team but most features and fixes come from the community without central assignment.
Scenario 3
A sales org requires every deal above $50k to be approved by VP Sales.
Scenario 4
A competitor is a federation of independent affiliates that all sell the same brand; each affiliate sets its own tactics and local partnerships.
Section 11
Top Resources
Swarming and network forms appear in military doctrine, organisational design, and systems thinking. These resources connect the concept to practice.
The RAND volume that framed netwar and swarm in modern conflict. Distinguishes network forms from hierarchies and explains when and how swarms outperform centralised forces.
McChrystal on restructuring the Joint Special Operations Command into small, empowered teams that shared context and could act without central approval. Organisational swarming in practice.
Starfish (decentralised) vs spider (centralised) organisations. The starfish can regenerate; cutting off a leg does not kill it. The book popularised swarm-like resilience in business.
Hastings' "context not control" and freedom-and-responsibility culture are a blueprint for swarm-like teams: give context, set goals, let people act. The deck is a practical reference for distributed authority.
05
Swarm Intelligence — Biology and CS literature
Reference
The biological and computational basis for swarm behaviour: ants, bees, and algorithms that use simple rules to produce coordinated behaviour. The same principles apply to human and organisational swarms.
Team of Teams
Stan McChrystal's "team of teams" — small, empowered teams that share context and align through trust and communication — is organisational swarming. The reinforcement: break the organisation into teams that can act like swarm nodes.
Tension
Distributed Systems
Distributed systems in tech (e.g. consensus, replication) implement swarm-like resilience: no single point of failure. The tension: distribution adds complexity and consistency challenges. Use when resilience and scale outweigh the cost of coordination.
