COMPETITIVE STRATEGY

Your Fiercest Rival Is Secretly Your Distribution Network

Google and Reddit, Threads and Mastodon, Amazon and the merchants it could crush all look like competitors, but each pair survives because one platform's weakness is a hole only the other can fill.

Based on the research ofCarballa-Smichowski, Duch-Brown, Martens, and Gomez-Losada, "Inter-platform ecosystems," Strategic Management Journal, 2026

Two "rivals," one shared externality, four fixes PLATFORM A e.g. a search engine PLATFORM B e.g. a community forum A's gap: no first-person answers B's gap: no monetized reach four ways to close the gap Network fusion host rival's content Community-driven users cross-post Meta-platform aggregate rival inside Concatenation refer out, hand off Cross-platform search costs fall "Rivals" become a working inter-platform ecosystem
The platforms a company spends the most energy "competing" with are often the ones quietly carrying its customers to the door.

The Externality Two "Rivals" Impose on Each Other

Carballa-Smichowski, Duch-Brown, Martens, and Gomez-Losada spent their paper looking at web traffic across 241 European platforms and found something that standard platform strategy does not expect: many platforms that look like substitutes for the same user attention are, functionally, complements. They call the pattern an inter-platform ecosystem, and their explanation for why it forms is sharper than "network effects are nice." Each platform, simply by existing and growing, imposes an externality on the others around it. A search engine that cannot produce fresh, first-person answers imposes a quality externality on its own users. A forum that cannot get in front of anyone outside its own app imposes a distribution externality on its own content. Neither platform can fix its own externality alone, because the fix requires something the other platform already has.

The paper identifies four structural answers to that matching problem. Network fusion means hosting a rival's content directly inside your own product. User-community-driven interactions mean letting ordinary users do the connecting themselves, through cross-posting and linking, without the platforms ever signing anything. A meta-platform strategy means aggregating another platform's output inside your own interface, so your users never have to leave. Platform concatenation means referring a user onward to a different platform for one specialized, complementary step, then taking them back. None of these require the two platforms to like each other, or even to admit the arrangement exists. They require only that the externality be expensive enough, and the fix cheap enough, that someone builds the bridge.

Google and Reddit Did Not Merge, But They Might As Well Have

Google Search and Reddit compete for exactly the same five seconds of a user's attention: the moment someone types a question into a box, deciding whether to trust an algorithm's synthesis or a stranger's forum post. By any conventional competitive map, they are rivals for the same query. By the paper's logic, they are an inter-platform ecosystem, and the traffic data shows it plainly.

In February 2024, Reuters confirmed, citing people familiar with the matter, that Reddit had struck a data-licensing deal with Google worth about $60 million a year, giving Google structured, real-time access to Reddit's content for search and AI products, alongside a separate arrangement letting Reddit build AI features on Google's Vertex AI. Three months later, Search Engine Land reported that Google had begun testing special visual treatment for Reddit results in its search pages, surfacing upvote and comment counts directly in the snippet, a change that gives Reddit threads a more prominent, information-rich presence on Google's own results page. That is the meta-platform strategy in the paper's own terms: one platform aggregating another's content, with its community signals intact, inside its own interface.

The traffic effect of this arrangement, deal and algorithm change together, was not subtle. Data from Ahrefs, reported by the marketing trade outlet B&T, shows Reddit's global organic search traffic rose from roughly 160 million visits in August 2023 to about 480 million in February 2024, in the six months after Google's Helpful Content Update started favoring first-person, forum-style content. Industry SEO analysts quoted in that same reporting openly debated whether the commercial deal was shaping Google's algorithm, a question Google has publicly denied. Whatever the causal mix, the practical result was that Reddit's traffic nearly quadrupled while Google's own answer quality, by its own stated goal, improved.

Google and Reddit compete for the same query, yet each one's product is now being assembled, in part, with pieces the other one supplies.

Neither company needed to stop competing to make this work. Google still wants to keep users inside its own results page as long as possible; Reddit still wants its own app to be the destination, not Google. But Google's content externality, that it cannot manufacture authentic first-person opinion at scale, is fixed by routing queries toward Reddit. Reddit's distribution externality, that it cannot reach anyone who does not already open the Reddit app, is fixed by Google sending them in. A licensing contract and a snippet redesign did more to resolve that mutual deficiency than any amount of head-to-head competition could.

Threads Opened a Door Into a Network It Doesn't Own

Meta's Threads was built, explicitly, to compete with X for the real-time, public conversation audience. The fediverse, the loose federation of Mastodon, Pixelfed, PeerTube, and similar apps built on the open ActivityPub protocol, is not merely a competitor to Threads; it is a project whose founding premise is resistance to exactly the kind of centralized, advertising-funded platform Meta runs. These two things should have nothing to offer each other.

In December 2023, Meta began testing the opposite. Engadget reported that a new feature let some Threads content become visible to, and followable from, Mastodon and other ActivityPub-based apps for the first time. Mark Zuckerberg announced the test directly on Threads: "Making Threads interoperable will give people more choice over how they interact and it will help content reach more people. I'm pretty optimistic about this." Instagram head Adam Mosseri added that the company was starting with the ability to follow Threads users from ActivityPub clients, with following in the other direction to come later. Meta had already let users verify a Threads profile on Mastodon before this; the fediverse test extended that into actual cross-network content flow.

This is the paper's network fusion strategy, opening your product to host and be hosted by a notionally rival network, layered with user-community-driven interactions, since the actual linking, following, and replying across the two networks is something individual users and server administrators choose to do, not something the platforms broker centrally. The externality each side was carrying is visible in how each side talked about it. Threads needed legitimacy and reach among the portion of the social web that refuses closed platforms on principle, plus a credible answer to regulatory pressure for interoperability. The fediverse needed exactly what decentralization structurally lacks: gravitational pull. Mastodon's creator, Eugen Rochko, captured the trade explicitly in a blog post at the time, writing that "the fact that large platforms are adopting ActivityPub is not only validation of the movement towards decentralized social media, but a path forward for people locked into these platforms to switch to better providers." He welcomed the bridge while remaining, structurally, the rival project on the other side of it.

Amazon Lends Its Checkout to the Sites It Could Just As Easily Crush

Independent direct-to-consumer brands exist, in part, to escape Amazon's marketplace fees and algorithmic ranking. Their own websites are supposed to be the alternative to Amazon, not an extension of it. Yet Amazon's Buy with Prime program runs Prime's fast shipping, one-click checkout, and easy returns on those same independent merchants' own sites, outside Amazon.com entirely.

Amazon's own announcement, released in January 2023, reported that offering Buy with Prime increased shopper conversion by an average of 25 percent, measured as the difference in completed orders when the option was available versus when it was not, on the same site over the same period. Individual merchants cited in that release reported even larger gains: the skincare brand Trophy Skin reported a conversion increase of more than 30 percent, the smart-home brand Wyze reported 25 percent and rolled Buy with Prime out across its entire catalog, and the hydration brand Hydralyte reported a 14 percent increase.

This is platform concatenation in its cleanest form: a shopper is handed off, for one specific and complementary step, to a different platform's infrastructure, and then handed back to finish the purchase on the merchant's own site. The externality driving it is cart abandonment. A small, unfamiliar storefront loses sales at checkout because shoppers will not hand a stranger their card details for an unfamiliar guarantee on shipping or returns. Amazon's logistics and trust infrastructure patches that hole directly. In exchange, Amazon gets Prime engagement and purchase data generated entirely off its own marketplace, from traffic it never had to acquire. A brand that exists to compete with Amazon's marketplace becomes, through one embedded button, a distribution partner for Amazon Prime.

Reading Your Own Ecosystem the Right Way

None of these three pairs required a truce. Google still wants to keep searchers on its own results page; Reddit still wants to be the destination, not a source Google cites. Threads still wants to be the bigger network; Mastodon's whole premise is resisting networks like Threads. Amazon's marketplace still wants the sale on Amazon.com; independent brands still exist to avoid exactly that. Competition between each pair is real and ongoing. It coexists with a second relationship running underneath it, where one platform's structural weakness is a gap only the other is positioned to fill, and both sides are better off building the bridge than pretending the gap does not exist.

The operating question this paper hands to anyone running a platform is not "who competes with us for attention." It is: whose users are worse off because of a gap in what we offer, and is there a platform positioned to close that gap in a way we can capture value from, rather than simply lose to. With 241 platforms already showing this pattern in Europe alone, inter-platform ecosystems are not a handful of clever exceptions. They are the predictable structural response wherever growth on one platform quietly creates a problem that only a different platform, often one wearing the label of rival, is actually built to solve.

Sources

  • Carballa-Smichowski, Duch-Brown, Martens, and Gomez-Losada, "Inter-platform ecosystems," Strategic Management Journal, 2026 doi.org
  • Anna Tong, Echo Wang, and Martin Coulter, "Exclusive: Reddit in AI content licensing deal with Google," Reuters, February 22, 2024 reuters.com
  • Barry Schwartz, "Google is testing special snippet treatment for Reddit search results," Search Engine Land, May 22, 2024 searchengineland.com
  • Arvind Hickman, "'Quality Of Google Search Crashed', Concerns As Reddit's Traffic Nearly Quadruples In Six Months," B&T, March 11, 2024 bandt.com.au
  • Karissa Bell, "Meta tests making Threads posts available on Mastodon and other fediverse apps," Engadget, December 13, 2023 engadget.com
  • "Amazon's Buy with Prime Increases Shopper Conversion by an Average of 25%," Amazon (US Press Center), January 10, 2023 press.aboutamazon.com
  • Matthias Bastian, "Reddit reportedly signs $60 million annual training data deal with Google," The Decoder, February 22, 2024 the-decoder.com
← More on the blog