A market can be someone’s product. The operator brings buyers and sellers together, gives them a way to find one another, records what they have done and decides the conditions under which they may return. The sellers may own their tools, choose their hours and supply the service themselves. Yet an organization stands behind the place where they meet their customers, and its decisions reach their working lives. The firm’s boundary has become difficult to see because several boundaries no longer coincide. Ownership of the equipment, direction of the work, possession of the customer relationship and the right to exclude can lie in different hands. The question is what has moved outside the firm, and what the firm continues to govern from within.
The older account begins with a comparison of costs, and the arrival of cheap communication does not repeal it. A producer can buy a component more easily when specifications travel quickly, suppliers can be compared and performance can be checked. The same improvements can help a manager direct more plants, coordinate their inventories and see a failure sooner. Coase recognized both directions in 1937: an invention can alter the cost of using the market and the cost of organizing inside a firm. Their relative movement matters. A technology that carries information farther can make an independent supplier more accessible, or make a distant establishment easier to administer. The device alone does not tell us which organization will prevail.
The later Dissolution made this point with an invented component whose arithmetic is worth retaining. Making it uses eighty units in production and twelve in internal coordination, for ninety-two altogether. A specialist can make the equivalent component for seventy-five, but external coordination adds twenty-two, bringing the cost of buying to ninety-seven. Lower that external cost to ten and buying costs eighty-five and becomes cheaper than organizing the work inside. Now let the same advance lower internal coordination to two. Making costs eighty-two, and the earlier conclusion reverses. Quality, quantity and every other consequence are held constant. A cheaper connection creates an opportunity to redraw the line; the remaining costs determine where drawing it saves resources.
Outsourcing also leaves an organization on the other side of the invoice. The cleaner may cease to work for the building’s owner and begin to work for a specialist who sets the schedule, supplies the materials and supervises several buildings. One firm has contracted while another has acquired the work. A franchise divides the arrangement differently: a local owner can supply capital and effort under standards associated with a common name. These arrangements separate ownership and coordination without eliminating either. The recovered Fissure followed their consequences for the people doing the work. Its central question survives the change in form: which decisions and risks cross the boundary, and which remain attached to the enterprise whose customers the workers serve?
A platform makes a further distinction by enabling the provider to deal directly with the customer. Hagiu and Wright distinguish this mode from the integrated firm and from a reseller that purchases an output and sells it onward. In their account, the provider retains control over important features of the transaction, while the relevant groups affiliate with the platform in order to interact. The definition concerns the allocation of decisions. It does not require the intermediary to relinquish every condition of access, and a business can combine modes across its activities. A seller who sets a service’s price can still depend on another organization for admission to the marketplace. A buyer who chooses among providers can still be choosing within a field someone else has arranged.
The advantage of leaving a decision with the provider can be precisely the information that a central manager cannot see. A specialist may know which service a particular customer needs and bear the return from adapting it well. In Hagiu and Wright’s professional-services model, this local information and the incentive to exert unobservable effort favor the platform. Integration has a different advantage when one provider’s decision affects the others. Separate advertising choices may draw customers to the whole group or merely draw them away from neighboring providers. An integrated organization can coordinate those effects. The comparison turns on the relative importance of information, effort and spillovers, together with the contracts available to support them. These are reasons for different allocations of control, rather than stages on an inevitable journey toward independence.
Payment can repair part of the difference without erasing it. A firm can reward an employee for the business she brings in; a platform can charge or subsidize transactions to alter the provider’s incentives. In the model, those instruments change both the effort supplied and the incentives to make other costly decisions. Giving the employee more of the return can weaken the integrated firm’s incentive to undertake its own activity. Altering the platform’s charge can improve one spillover while distorting another choice. The resulting comparison is conditional, even within the model. It gives the firm reasons to buy, make or enable different parts of its business, and it gives the observer a better question than whether an enterprise happens to call itself a platform.
For a participant, the practical boundary can sit at the entrance. Imagine a marketplace in which providers choose their methods and prices, while the operator controls listings, reputation records and access to customers. A provider may have ample discretion over a job already obtained and little power to contest the loss of future work. Whether that dependence is severe turns partly on alternatives: can customers find her elsewhere, can her record travel with her, and can she use several intermediaries at once? A credible opportunity to leave can constrain the gatekeeper. Dependence on one difficult-to-replace route can strengthen it. The arrangement may coordinate without daily orders while still locating consequential authority in the rules of participation.
The fee makes the division visible, but its movement alone does not measure a gain in efficiency. In a separate invented transaction, the customer pays one hundred, the service uses sixty in resources, and the platform uses eight in operating resources. A fee of twelve leaves the provider twenty-eight and the operator four. Raise the fee to eighteen while holding the service, costs and participation unchanged: the provider retains twenty-two and the operator ten. Their combined return is thirty-two in both cases. The larger platform margin is a transfer within this account. If instead the platform lowers its real operating cost from eight to five, the combined return rises to thirty-five. Prices can change the organization’s incentives and therefore its future output, but that further effect has to be shown; a new distribution of the same receipts cannot establish it.
The same discipline applies to risk. Moving the obligation to replace equipment or absorb a period without customers can lower the lead organization’s recorded expense without eliminating the underlying exposure. Someone still supplies the equipment and lives through the interruption. The new arrangement may handle that risk better, or put it on someone less able to bear it. It may create useful specialization while weakening a particular worker’s claim on the gains. The economic boundary therefore needs an account of obligations alongside its account of decisions. Who promises the service, who stands behind a failure, and who can demand an explanation when a rule changes? Those questions remain even when the transaction is initiated by a screen.
The firm can give up the task and keep the gate.
The platform does not settle whether the firm is melting into the market. It shows how a firm can organize a market, retain some decisions and distribute others among participants whose independence is real in one respect and limited in another. Cheap communication makes new arrangements possible; it can also extend the reach of the organization that chooses their terms. The useful boundary is consequently drawn decision by decision, obligation by obligation, with the gains and burdens followed across it. Once the firm’s authority has been traced beyond its employees and assets, the volume’s final question becomes unavoidable: for whose benefit should that authority be exercised, and to whom should its holder answer? The argument remains open.