Every industry that connects producers to consumers through intermediaries faces the same architectural choice. Should the channel be open — a neutral conduit through which any qualifying product can flow, chosen on merit — or closed — a proprietary pipeline distributing only the house's own offerings, or those of partners who paid for placement? The question appears in wealth management (can your advisor recommend any fund, or only the bank's?), in technology platforms (open protocols versus walled gardens), in pharmaceuticals (open formularies versus exclusive deals), in app stores, in media distribution. The vocabulary differs; the structure is identical. And the choice is never merely technical. It determines who competes with whom, whose interests the intermediary serves, how innovation happens, and where the profits pool. Understanding the logic of each architecture — including the genuine virtues of closure, which advocates of openness too easily dismiss — is essential to seeing why industries oscillate between the two and why "open" so often degrades into "nominally open."
The core trade-off: alignment versus integration
The case for open architecture is at bottom the case examined throughout the economics of agency: intermediaries who distribute only their own products face a structural conflict between what is best for the client and what is best for the house. A closed-distribution advisor at a bank that manufactures mutual funds is not really an advisor; they are a salesperson wearing an advisor's title, and empirical audit studies of financial advice confirm that recommendations track compensation grids with depressing fidelity. Open architecture severs manufacturing from distribution, letting the intermediary select from the whole market. Competition then disciplines products directly: the fund must survive comparison, not merely capture a channel. Clients get better prices and better fit; new entrants get a route to market without building their own distribution; the market as a whole gets faster selection pressure on quality. The historical shift of wealth management from proprietary sales forces toward open platforms — accelerated by fiduciary regulation and by clients' growing ability to compare — reflects exactly this logic. So does the generativity argument in technology: Jonathan Zittrain's analysis of the "generative" internet showed that open, permissionless architectures produce innovation their designers never imagined, because the edge — not the owner — decides what gets built and distributed.
But closure has real economics behind it, not merely greed. Integrated distribution can guarantee quality end to end: Apple's closed App Store is also a curated one, and the malware rate on tightly controlled platforms is genuinely lower. Closure solves coordination problems — an integrated pipeline can optimize the whole product experience where an open bazaar cannot. It funds the channel itself: distribution is costly, and proprietary margins pay for the branches, the sales force, the platform development that pure conduits struggle to finance. And it creates accountability of a certain kind: when one firm controls the whole stack, there is one throat to choke when things fail, whereas open ecosystems diffuse responsibility across a supply chain no one fully governs. The honest framing is thus not virtue versus vice but a trade-off: openness aligns the intermediary with the client and maximizes variety and competition; closure enables curation, coherence, and investment in the channel — while creating a standing temptation to abuse the gatekeeping position.
Why "open" decays: the political economy of platforms
The deeper insight from two decades of platform economics is that openness is not a stable equilibrium but a strategy phase. Platforms characteristically open to grow and close to harvest. In the growth phase, openness solves the chicken-and-egg problem: welcoming all producers attracts consumers, whose presence attracts more producers — the indirect network effects analyzed in the foundational two-sided-market literature of Rochet, Tirole, and Parker/Van Alstyne. But once the network effects lock users in, the platform's incentives rotate. It begins enveloping its complementors — copying successful third-party products into the core (the fate of countless apps whose features became operating-system functions), demoting rivals in search and shelf placement, raising the take-rate on the channel, and converting neutral infrastructure into a tollbooth. Amazon's marketplace illustrates the full cycle: radically open to third-party sellers, and simultaneously a competitor to them, with documented cases of the platform's private-label arm mining seller data to clone winning products. The architecture remains formally open; the distribution has effectively closed around whoever pays or belongs.
Finance exhibits the same decay in its own dialect. "Open architecture" platforms that charge funds for shelf space — revenue sharing, pay-to-play placement on the recommended list — have reconstructed closed distribution beneath an open façade: the client sees a marketplace, but the menu was auctioned, not curated on merit. This is guided architecture, and it is arguably more dangerous than honest closure because it launders the conflict: the proprietary salesperson's bias is at least legible, while the "open" platform's bias hides inside an apparently neutral choice screen. The general law is Goodhart-like: once "openness" became a marketing asset, the label detached from the substance. What matters is never the label but three concrete questions — Who can get on the shelf, and at what price? How is the default and the ranking determined? Does the intermediary earn differently depending on what the client chooses? An architecture is only as open as its answers.
The systemic stakes: innovation, resilience, and power
Zoom out from any single firm and the architectural choice shapes entire innovation systems. Open architectures externalize experimentation: thousands of independent producers bear the cost of trying things, and the system harvests the few successes — the evolutionary logic that made the open internet, open-source software, and open fund platforms such prolific generators of variety. Closed systems internalize experimentation, which yields coherence but caps the search at the imagination and budget of one firm. Open systems are also more resilient to any single point of failure but more vulnerable to quality collapse and fraud at the edges (the open web's spam, the open marketplace's counterfeits); closed systems invert the profile — high average quality, catastrophic exposure to the gatekeeper's errors, stagnation, or extraction.
There is finally a question of power, which the purely economic framing understates. Distribution chokepoints are governance: whoever controls the channel taxes and regulates everyone who must pass through it, functioning as a private legislature over an industry. The escalating antitrust confrontations over app-store terms, self-preferencing, and platform interoperability mandates (the EU's Digital Markets Act being the most explicit) are attempts to impose open-architecture obligations on infrastructures that became too essential to leave closed. The historical rhyme is instructive: railroads, telephone networks, and electricity grids each began as proprietary closed systems and were each eventually converted — by regulation — into common carriers obliged to serve all comers on nondiscriminatory terms. Digital and financial distribution are now retracing that arc, and the recurring lesson is that essential distribution eventually forfeits the right to be closed.
Choosing, and designing, honestly
For a firm, the choice is contingent: closure is defensible where quality assurance is genuinely hard, integration genuinely improves the product, and the firm resists milking the gate; openness is superior where client trust is the product — which is why fiduciary advice and closed distribution are ultimately incompatible, whatever the compliance department says. For a system designer or regulator, the task is subtler than mandating openness: it is policing the integrity of openness — transparent listing criteria, nondiscriminatory ranking, separation of the platform's merchant role from its umpire role, and disclosure not of boilerplate conflicts but of the actual economics of placement. And for the client or user, the protective habit is the structural question already familiar from the study of incentives: how does the channel earn, and does it earn differently depending on what I choose? A truly open architecture is one where that answer is "no" — and everything else, however labeled, is a distribution strategy wearing openness as a costume.
The oscillation between open and closed will not end, because each architecture accumulates the pathologies that make the other attractive: closed systems breed conflict and stagnation until openness looks like liberation; open systems breed chaos and covert tollbooths until curation looks like salvation. The mature position is to stop treating either as an ideology and start reading every channel — bank, platform, app store, formulary — through the same three questions about shelf, ranking, and payment. Architecture, in distribution as in buildings, is destiny; but only inspection reveals which architecture you are actually standing in.
References
- Parker, G. G., Van Alstyne, M. W., & Choudary, S. P. (2016). Platform Revolution: How Networked Markets Are Transforming the Economy and How to Make Them Work for You. W. W. Norton.
- Zittrain, J. (2008). The Future of the Internet — And How to Stop It. Yale University Press.
- Rochet, J.-C., & Tirole, J. (2003). "Platform Competition in Two-Sided Markets." Journal of the European Economic Association, 1(4), 990–1029.
- Eisenmann, T., Parker, G., & Van Alstyne, M. (2011). "Platform Envelopment." Strategic Management Journal, 32(12), 1270–1285.