As Jim Barksdale famously said: “There’s only two ways I know of to make money: bundling and unbundling — most people spend half their time adding and other people spend half their time subtracting, so that’s what works out.”

Yesterday, Facebook put a bullet in Parse, its “unbundled” platform-as-a-service for mobile developers, throwing its substantial weight behind the fully bundled world of Facebook apps. Given the company’s most recent earnings announcement, its strategy of pulling content from all over the Web into the Facebook app vortex is monetizing well. No surprise, given Facebook’s all-seeing access to all user data and activities.

You have to marvel at how fast things have changed: After Facebook acquired Parse for $85 million in 2013, Parse was the centerpiece of Zuckerberg’s Facebook keynote just last year. Since then, however, Facebook has doubled down on Messenger, which it believes to be the platform of the future as it chases WeChat’s success in China. This might seem strange, since most people still consider Facebook a social network, but here’s the big picture:

In recent months, Facebook has greatly shifted focus and resources from its original, browsing Web-based social network to its mobile Messenger app, and it is busy rolling out enterprise-friendly services, such as customer communication bots, to increase its appeal to businesses. This pivot for a $300 billion cap company is right up there with Bill Gates’ similar move in 1995, when he realized the Internet Tidal Wave was coming for Microsoft. There’s a very good reason for this: Messaging and notifications are rapidly eating up the Web experience, becoming the dominant way we interact with and consume Internet content and services. As we browse the Web less, we browse our social media notifications and messages more. Browsing the Web on our phones is convenient, but it’s also sub-optimal. Which in turn drives us back into our social media, where Web links are contextualized — mostly by friends we trust and people/brands we’ve opted to follow.

Facebook is savvy to see this and shut down Parse — which after all, only enables companies to build products that compete with Facebook and its apps — while retrofitting Messenger into a communication tool for enterprise. So a key question in the coming years is whether enterprise will embrace Messenger, even if that means handing over their customer data and customized user experience to Facebook and its walled garden. (On this topic, social game developers may have some hard-earned wisdom to share.) Developers will get distribution for their content and a potential reach of over 1 billion users, but in exchange, they lose control over the customer experience and their data. Companies will soon need to decide if they’re willing to give up so much to play in Facebook’s walled garden.

However, it would be a mistake to assume that one messaging app will engulf our Internet activity, because we’re entering a phase of the Internet where the mass explosion of messaging is in non-messaging apps: Dating, on-demand services, retail, gaming, and even enterprise apps utterly depend on messaging for the core experience and to retain users and drive engagement. (Or as designer Luke Wroblewski put it as a corollary to Zawinski’s Law, “Every mobile app attempts to expand until it includes chat. Those applications which do not are replaced by ones which can.”) For that reason, I think the future of Internet engagement may not belong to any one platform or any single walled garden but to thousands of mobile mega-niches of messaging and conversation.

If I’m right about that, Facebook just did developers who’ve come to rely on Parse an enormous favor by killing it — no longer dependent on a company dedicated to channeling the entire Internet experience into its own apps, they are now free to consider platforms that enable them to build experiences that benefit their users — as opposed to Facebook’s dominance.

Ron Palmeri is founder and CEO of Layer.