#209 1995 · Li & Fung (Victor Fung) · Supply chain and trading
Li & Fung capped itself at 70% of any factory's business, on purpose
the problem
Relying too heavily on one factory is a risk that cuts both ways
background
A company that sources goods through independent factories rather than owning them faces a dilemma in both directions. Take too little of a factory's capacity and you're a minor customer with no leverage, first to be bumped when a bigger order comes in. Take too much and the factory becomes dependent on you alone, loses the outside relationships that keep it sharp and innovative, and its problems — a bad quarter, a cash crunch, a quality slip — become your problems too, because you have nowhere else to go on short notice.
Victor Fung's answer wasn't to diversify suppliers in the usual sense — spread orders thin across many factories to reduce risk. It was to manage a precise range with each one: enough of their business to matter, never so much that either side stopped needing anyone else. The company was, in effect, deliberately declining available volume to preserve the very independence that made its partners worth working with.
what everyone would do
The standard sourcing instinct is to maximize your share of a good factory's capacity — the more of its output you control, the more pricing power and priority you get, which is why buyers routinely compete to become a factory's largest or even exclusive customer.
what they saw
Victor Fung saw that this instinct creates a mirror-image version of the very risk it's meant to solve: past a certain share, a factory's fate becomes your fate, because it loses the other customers whose orders kept it financially diversified and technically sharp — so buying more of a factory's business doesn't reduce dependency risk, it just moves the dependency from 'we rely on them' to 'they rely entirely on us,' with nowhere else for either side to turn if something goes wrong.
the move
Rather than maximize the share of any factory's output it controlled — the obvious way to lock in capacity and pricing power — Li & Fung, under chairman Victor Fung, deliberately kept its own purchases within a band: never less than 30% of a given factory's business, enough to matter to them, and never more than 70%, enough to leave room for other customers. Above 70%, the factory loses exposure to other clients' new ideas and techniques and Li & Fung risks inheriting the factory's problems as its own; below 30%, it isn't a big enough customer to command attention or negotiate.
why it works
Capping purchases below 70% keeps the factory exposed to other customers' orders, feedback and techniques, which keeps it competitively sharp and gives it a revenue base that survives a slow quarter from Li & Fung alone. Setting a floor at 30% ensures Li & Fung still represents enough of the factory's business to get priority and real attention rather than being squeezed out when a bigger order arrives. Managing deliberately within that band, instead of maximizing toward either extreme, means either side can absorb a shock — a bad month, a lost customer, a quality dispute — without the other's business collapsing alongside it, which is what let Li & Fung scale the same discipline across 15,000 factories without ever needing to own one.
the payoff
The discipline let Li & Fung orchestrate a network that grew to a reported 15,000 factories across more than 60 countries, sourcing several billion dollars of goods a year without owning a single factory itself — a model management scholars later named "network orchestration" and Harvard Business School turned into a teaching case.
where it breaks
The discipline requires enough scale and enough alternative factories that walking away from excess volume is actually an option — a smaller buyer without alternative sourcing has no spare capacity to redirect and may have to take whatever share a factory offers. It also requires a market with multiple comparably capable suppliers to diversify across; where only one factory can do the specialized work at all, there is no band to manage, since the buyer may need 100% of that single relationship or nothing. And sustaining the band requires real, ongoing tracking of what share of each factory's capacity you represent, an investment in relationship management a purely transactional buyer has no reason to make.
what came after
The 30/70 discipline became a textbook example of what business scholars call network orchestration — coordinating a vast web of independent producers without owning them — and Li & Fung's model, built on walking away from business on purpose, is still taught as a counterpoint to the instinct to capture as much supplier capacity as possible.
references
- [1]What Does It Take to Compete in a Flat World?Knowledge at Wharton, 2007knowledge.wharton.upenn.edu
- [2]Li & Fung shows its mettle in flat worldSouth China Morning Post, 2007scmp.com