Skip to main content
Category: Strategies
Type: Interfirm orchestration and cooperative strategy
Origin: J. Carlos Jarillo, IESE, 1988; Powell, Lorenzoni & Baden-Fuller, Dyer & Singh, Gulati, Nohria & Zaheer, 1990–2000
Also known as: Strategic networks; network orchestration; relational strategy
Quick Answer — Network Strategy is competing by designing and governing a web of independent partners, rather than owning every step of the chain. J. Carlos Jarillo named the strategic network in 1988; later work showed that advantage can live in the relationship itself. The practical lesson is that a vendor list is not a network: without a center that creates value for partners, you have outsourcing, not strategy.

​
What is Network Strategy?

Network Strategy is a way of competing in which a firm designs, governs, and learns through a web of independent partners instead of trying to own the whole value chain.
The cooperative relationships of a firm can be the source of its competitive strength.
Think of a conductor who does not play every instrument. The score, the cues, and the trust that Saturday’s players will arrive are the work. Network Thinking is the habit of seeing nodes and links. Network Strategy is the business bet that those links can be designed. It is not the same as platform strategy. A platform matches two or more customer sides so that network effects compound. A network firm may never run a marketplace. It still wins if it can recombine partners faster, and more reliably, than a rival can rebuild the chain in-house.

​
Network Strategy in 3 Depths

  • Beginner: You do not have to own every piece. You need a living map of who can do what with whom, and a reason they will still pick up the phone.
  • Practitioner: Name the center, the nodes, what you will never own, and the rule that keeps partners skilled. Then run one real order through that map before you draw an org chart.
  • Advanced: Advantage sits in relation-specific assets, shared routines, complementary capabilities, and governance. A thick contact list is not a moat. A two-sided platform is one special case, not the whole family.

​
Origin

The academic name is younger than the practice. Japanese keiretsu, Italian industrial districts, and Hong Kong trading houses had already been competing through partners. Strategy research needed a language that was neither “the firm as fortress” nor “buy it on the spot market.” J. Carlos Jarillo, at IESE in Barcelona, supplied that language in 1988. In “On Strategic Networks,” Strategic Management Journal 9(1): 31–41, he treated the strategic network as a mode of organization. Cooperation and competition were not opposites. A firm’s chosen partners could be the source of its strength. Two years later, Walter W. Powell argued that networks are not a mushy hybrid of market and hierarchy. In “Neither Market nor Hierarchy: Network Forms of Organization,” Research in Organizational Behavior 12 (1990): 295–336, he described them as a third form, held by reciprocity, reputation, and complementary strengths rather than by price or by employment. The “who holds the web together?” question came next. Gianni Lorenzoni and Charles Baden-Fuller, in California Management Review 37(3) (1995): 146–163, studied firms such as Benetton and Apple. Successful webs, they wrote, are guided by a strategic center—more than a broker of phone numbers. The center creates value for itself and for partners, sets an agenda, and changes structure as strategy changes. The economic claim sharpened in 1998. Jeffrey H. Dyer and Harbir Singh, in “The Relational View,” Academy of Management Review 23(4): 660–679, argued that a firm’s critical resources may span its legal boundary. They named four sources of interorganizational advantage: relation-specific assets, knowledge-sharing routines, complementary resources, and effective governance. In March 2000, Ranjay Gulati, Nitin Nohria, and Akbar Zaheer opened a Strategic Management Journal special issue (21(3): 203–215) with a sharper warning: strategy research had treated firms as autonomous entities. The networks they sit in shape conduct and performance. The same issue carried Dyer and Kentaro Nobeoka’s Toyota knowledge-sharing study. The later practitioner label is network orchestration, stated at book length by Victor K. Fung, William K. Fung, and Yoram Wind in Competing in a Flat World (2008).
The networks of relationships in which firms are embedded profoundly influence their conduct and performance.

​
Key Points

Network Strategy earns its keep when the scarce skill is recombination: matching the right partners to this order, this week, without owning every plant. Comparative advantage says specialize. This strategy asks who will hold the specialties together.
1

Treat the relationship as the unit of advantage

Industry maps and in-house resources still matter. They do not exhaust the question. Ask what you can do with this partner that neither of you can do alone, and what would vanish if the tie broke. A one-off purchase is a market. A repeating, specialized tie is the start of a network. If the advantage walks out with a single contract, you do not yet have one.
2

Be a strategic center, not a switchboard

Lorenzoni and Baden-Fuller’s test is blunt: does the center create value for partners, or only skim introductions? A center sets the problem, helps partners raise their own skill, and keeps information moving. A switchboard collects quotes. The first is a strategy. The second is a directory. If partners would be equally well off without you, they will be, soon.
3

Invest in the four relational sources

Dyer and Singh’s list is a checklist, not a slogan. Relation-specific assets are tools, data, or habits that only pay off with these partners. Knowledge-sharing routines are the meetings and inspections that move know-how without waiting for a lawsuit. Complementary capabilities are the pieces that click. Governance is how you settle a fight before it reaches court. Skip the investments and you have a vendor list with extra adjectives.
4

Leave partners independent enough to stay skilled

If you take 100 percent of a factory’s book, you have hired it without the employment contract. In a 2007 interview, Victor Fung described occupying about 30 to 70 percent of a partner’s capacity, never all of it, so the partner still serves others and keeps learning. The 1998 version of the same idea was a flexible hold: reserve undyed yarn and mill time, then name the color five weeks before delivery. Control the configuration. Do not pretend you own the people.

​
Applications

Use Network Strategy when the work is assembling a chain that does not yet exist, not when a single owned process must be run to a safety spec. Do not use it as a slogan for “we outsource now.”

Source a product without owning the plants

Write the steps as nodes: yarn, dye, cut, sew, trim, ship. For the next order, assign each node to a named partner and a backup. Time the color decision as late as the network will bear. This is how a small brand can run a seasonal line without a mill.

Build a professional web, not a lone-wolf shop

A clinic, a design studio, or a research lab rarely needs every specialty on payroll. Name the center (who owns the client), the repeating partners, and the quality rule. Then run one joint case before you print “ecosystem” on the website. Win-win talk without a governance rule is a toast, not a network.

Assemble skills early in a career

You will not be illustrator, copywriter, printer, and salesperson in the same year. Map three people you can already help, and one craft you will keep in-house. Trade work on a clock, not as favors that never get repaid. The web is the asset; the software you learn this month is a node.

Share what a town cannot staff alone

A rural clinic network, a volunteer fire pact, or a school that shares a specialist teacher is a public version of the same design. Write who is the center this season, how hours are repaid, and what happens when two towns need the specialist on the same Tuesday. Reciprocity is the contract. A poster about “community” is not.

​
Case Study

The numbered window onto Network Strategy as a going concern is Li & Fung in Joan Magretta’s interview with chairman Victor Fung, Harvard Business Review, September–October 1998 (vol. 76, no. 5). The firm was founded in Guangzhou in 1906 as a Chinese-owned export house. By 1998 it was Hong Kong’s largest export trader, still run by the founder’s grandsons, Victor and William Fung. It did not own factories and did not manage factory workers. It inspected production, designed the chain, and recombined partners for each order. The 1997 snapshot is concrete. Revenue was about $1.7 billion. The company had 35 offices in 20 countries. It worked with about 7,500 suppliers in more than 26 countries and about 350 customers. Fung’s own arithmetic: if a factory averaged 200 workers, more than a million people were engaged on behalf of those customers. A typical garment was not “made in one country.” Yarn might be bought in Korea, woven and dyed in Taiwan, sewn in Thailand, and finished with zippers from a Japanese maker. One order of 10,000 garments, split across five Thai factories, could be on European shelves five weeks later, matched as if they had come from a single plant. The strategic point was time, not cheap labor. Shortening the buying cycle from three months to five weeks bought eight weeks of watching the market before committing color and quantity. Trust in the supply network let Li & Fung reserve undyed yarn and mill capacity, then name colors five weeks before delivery. That is orchestration: the configuration is the product. Boundary note: these figures are from a 1998 executive interview, not an independent census of factories. The case shows a mechanism. It does not prove that Li & Fung’s later years stayed easy, and it is not a warrant that every industry should hollow out its plants.

​
Boundaries and Failure Modes

Network Strategy fails when the work is a tightly coupled process that cannot be inspected at the seams. A drug’s sterile fill, a nuclear plant’s control software, or a recipe whose know-how is the firm, may need hierarchy. Partners who can walk with the spec leave you teaching tomorrow’s rival. It also fails when the center stops creating value. If you only auction volume, partners will bid the relationship down to a spot price. Relation-specific tools then become hostages. The hold-up problem is not a footnote. It is how a “network” turns into a fight over who owns the jig. The common misuse is to rename a vendor list. Outsourcing without a strategic center, without shared routines, and without a rule that keeps partners independently skilled is cost-cutting with new stationery. Another misuse is to treat platform strategy as a synonym. Matching riders and drivers is one architecture. Orchestrating five factories for one pair of shorts is another. Both can ride network effects. They are not the same bet.

​
Common Misconceptions

The English name collides with telecom planning, with social-media marketing, and with platform strategy, so it is easy to copy the word and miss the design.
Network effects describe how value rises as more users join. Platform strategy builds a multi-sided market to harvest that rise. Network Strategy is broader and older in the strategy literature: it is competing through a governed web of firms. A Li & Fung chain has no consumer-side marketplace. It still stands or falls on how the web is designed.
A long vendor list is a market with extra names. Powell’s point is that networks run on reciprocity and complementary strengths, not on the next bid. Dyer and Singh’s point is that you have to invest in the tie. If partners are interchangeable at noon, you do not have a strategic network at 12:01.
Flexibility is a result of unused capacity and trusted holds, not of the word “network.” If every partner is fully booked, or if you have outsourced the skill needed to inspect them, the web is brittle. Vertical integration is sometimes the cheaper way to keep a secret, a safety standard, or a learning loop inside the firm.
These pages sit next to the same problem: where advantage lives when the work no longer fits inside one legal boundary.

Platform Strategy

A multi-sided market that harvests network effects; one architecture inside a larger family of network bets.

Network Effects

Value that rises with the number of participants; a mechanism a network may use, not a synonym for the strategy.

Network Thinking

The cognitive habit of reading nodes and links; Network Strategy is the business use of that map.

Moat Building

Defensibility. A network moat is the hard-to-copy web, not the logo on the purchase order.

Comparative Advantage

Why specialize. Network Strategy is how those specialties are held together without one owner.

Win-Win

Mutual gains in a deal. A network needs that plus governance that survives the second order.

​
One-Line Takeaway

Draw the web, name the center, and leave partners skilled enough to leave—then give them a reason not to.