Digitalization marches on, AI continues its rise—and a handful of tech companies are doing very well indeed. The supremacy of these big tech firms increasingly depends on their ability to stand up and manage digital ecosystems with a rich constellation of “complementors”: firms or individual developers collaborating and co-creating to generate something of value for the end customer.
Such ecosystems are set to become more important for more firms, as economies increasingly rely on digital technologies. That’s why a widening range of companies will need to develop ecosystem strategies based on a clear understanding of what makes an ecosystem successful—whether they are “orchestrators” building the ecosystem, or complementors contributing to and transacting within it.
Two recent developments should prompt reflection on the dilemma between openness and control in the design of ecosystems. On the one hand, Meta’s Mark Zuckerberg has publicly articulated his vision of, and arguments for, an open AI ecosystem, built around the Llama model family. Albeit in a different domain, this is a striking reversal from Meta’s years-long (and notoriously costly) play for a closed, tightly controlled metaverse. On the other hand, Apple has chosen to enter the metaverse, rebranded as “spatial computing,” with a similar model of strict orchestrator control, built around its VisionPro headset. The spatial computing ecosystem that Apple hopes to foster would give it total control of the hardware and operating system that would enable complementors to sell VR and AR experiences
What is one to make of these conflicting bets? Apple might be attempting to replicate its exceptional success with the iPhone, leveraging its mettle in creating a user-friendly interface to lock in consumers. However, we argue that an orchestrator’s belief that it must create and control everything itself is, more often than not, a cardinal sin of ecosystem design. Indeed, the lessons from Meta’s foray into the metaverse remain valuable, and applicable beyond web3. First: Total openness and absolute dominance pose a false dilemma for ecosystem orchestrators. And second: Complementors have power to shape the trajectory of a viable, scalable ecosystem.
Meta’s bitter lesson
A decade or so ago, it was Facebook that aspired to extend its social media dominance into a new, orchestrator-dominated virtual realm. The company began pouring resources into building an insular ecosystem where it could exercise total control. It developed its own proprietary devices to offer augmented (AR) and virtual reality (VR) technologies, bought headset maker Oculus for $2 billion in 2014, and, as recently as 2021, launched a social VR space called Horizon Worlds. The finishing touch, and a powerful statement of intent, was to rebrand itself as Meta.
The belief that the metaverse had arrived triggered a gold rush, as tech giants like Microsoft and Apple raced to construct their own immersion-focused ecosystems. But fast-forward a few years and reality has fallen far short of the hype. By 2023, two years after its launch, Meta’s Horizon Worlds had fewer than 200,000 monthly active users, and only 9% of its worlds had been visited by more than 50 people. Meanwhile, U.S. sales of VR headsets and AR glasses crashed, plummeting nearly 40% to $664 million in 2023. That understandably dismayed content creators who were paying fees of up to 47.5% to sell their virtual goods on the platform, shrinking the incentive for these complementors to invest in creating added-value content.
Humbled, Meta belatedly acknowledged that collaboration might ultimately be a better course; it opened up the operating system that powers its line of MetaQuest VR devices earlier this year. This shift enabled third-party hardware makers to build mixed reality (MR) experiences. Meta also allowed “pre-approved” partners to develop specific use cases: hardware manufacturer Asus would develop a gaming headset under its Republic of Gamers brand, while Lenovo developed MR devices for productivity, learning, and entertainment. In 2022, Meta struck a partnership with Microsoft to bring the immersive tools of Mesh for Microsoft Teams to MetaQuest devices.
Meta’s current approach represents a middle ground between open and closed ecosystems and has allowed it to entice complementors while still controlling who can use its operating system and (to some extent) what for.
Models of ecosystem-building
Meta’s change of heart shows that there’s more than one way for orchestrators to attract and benefit from complementors in their ecosystems. Meta first aspired to a “hardware lock,” where users are locked into proprietary devices, but third parties can create content for them. Microsoft, meanwhile, aimed even higher with its own metaverse play: a “total lock,” where both hardware and software are under the orchestrator’s control.
Apple’s strategy is somewhere in between: A hardware lock, with selective and limited opening for complementors. Other companies are exploring different paths: Blockchain game developer Axie pursued a “content lock” where it controlled content creation, but was device-agnostic; and Decentraland, a truly “open platform,” is not only device-agnostic but actively encourages content creation by multiple third parties.
Drawing on our recent research project on the metaverse, we looked at the approaches different orchestrators took, and found that, as a rule, firms that tried to control too much were unable to generate complementor support and ultimately failed. It makes sense then that we’re seeing a broader trend of orchestrators giving up on this “orchestrator takes all” approach and embracing partnerships that instead multiply use cases and drive user adoption.
Only time will tell whether Apple’s gamble on the sort of hardware-lock, high-fee, and tightly- controlled-content strategy that Meta abandoned will pay off. But for now, it is predictably struggling to entice participation. None of YouTube, Spotify, or Netflix currently offers a dedicated app for Apple’s Vision Pro headset. In fact, 46 of the most popular apps on Apple’s App Store weren’t available on Vision Pro’s launch day. These complementors are tight-lipped on why they’re staying away, but their reluctance is already having an impact, undermining the appeal of Apple’s high-quality user hardware and user experience, and reportedly prompting Apple to cut production targets for the Vision Pro.
Lessons for ecosystem participants
Even though Big Tech giants may be tempted to leverage their extraordinary power to impose their will onto a new ecosystem, this not only thwarts complementors—it can often endanger ecosystem success. An effective ecosystem strategy does not always benefit from hardball tactics, as Meta’s own evolution shows. Through its strategic pivot, Meta now hopes to address the two biggest drawbacks to extended reality: sluggish hardware development and clunky experiences that few users enjoy. To do that the company is now betting on the future of an interconnected, interoperable metaverse, and gunning for first-mover advantage when it comes to setting standards for transferable digital identities. And these lessons are now being carried over into Meta’s play in the generative AI space as well.
Opening up the ecosystem means making the social layer within the Horizon OS that powers its Horizon Worlds more appealing to users by allowing profiles to glide seamlessly between virtual spaces, giving them ownership of their virtual identities. Meta’s calculation is that if Horizon OS becomes the engine of users’ social experiences everywhere, the company can harvest customer data and strengthen its existing market position in digital advertising.
Ecosystem complementors, that is, the partners who are struggling to compete in a world of tech giants, still have power to influence the ecosystems in which they operate—most importantly, by refusing to do so. For complementors, strategic alignment can substitute for tech power, and clever positioning can make a difference.
In the new growth area of GenAI-enabled ecosystems, getting the ecosystem strategy right will be crucial for complementors and orchestrators alike. As we’ve argued before, the locus of innovation and investment will overtime move downstream, towards small, domain-specific models that numerous companies will be best positioned to develop “on top” of the foundation models created by leading GenAI startups and hyperscalers.
***
The metaverse itself is far from dead. In fact, the spatial computing race has only just begun. Which ecosystem-building strategy will prevail: Meta with its new semi-open strategy, or Apple and its “total lock” hardware approach? Or will it be another orchestrator that already dominates the metaverse, like the online game platform Roblox, with 60 million daily active users? One thing is for certain: Orchestrators will have to become much more sophisticated and nuanced in how they run their ecosystems—whether in the virtual world or the real one.
Read other Fortune columns by François Candelon.
François Candelon is a partner at private equity firm Seven2 and the former global director of the BCG Henderson Institute.
Michael G. Jacobides is the Sir Donald Gordon Professor of Entrepreneurship and Innovation at London Business School, academic advisor at the BCG Henderson Institute, and the lead advisor of Evolution Ltd.
Katie Round is a principal at BCG X, Boston Consulting Group’s tech build and design unit.
Some of the companies mentioned in this essay are past or current clients of the authors’ employers.
