👊💰 We need a revolution in social media business models

Deepti Doshi on the business incentives underlying social media

Cory Doctorow has popularized the concept of “enshittification”: tech companies begin by creating real value for users, then gradually degrade as they increasingly optimize for profit and finally shift toward extracting value from users to benefit investors.

We’ve watched this happen to nearly every social media platform, and few people have had a better front row view of this than Deepti Doshi, a former Director at Meta and one of New_ Public’s Co-Founders.

These incentives are one of the most fundamental and least innovated upon factors shaping the internet. They’re a little like the moon — silently omnipresent, but creating the ocean waves that direct everything, sometimes resulting in typhoons and catastrophic erosion.

Below, Deepti lays out the way they actually work and how we can begin to shift them. Take it away Deepti!

–Josh Kramer, Head of Editorial, New_ Public

The case for building tech that serves people not investors

Underneath all the polarization, harassment, misinformation, and declining trust, social media runs on a business model that prioritizes investors over people and communities.

To really understand why so many of today’s platforms produce such similar harms, and to avoid repeating past mistakes in the next, AI-fueled, generation of social apps, we have to start here — surveillance advertising. It’s an elegant system financially, and can be a destructive one socially.

Here’s a thorough reminder of how this works. In short: Each platform is designed to maximize the time users spend on the platform. The more people join, and the longer people keep engaging, the more ads the platform can serve.

Meta, with over three billion users worldwide, generates more than $160 billion in annual revenue, and counting. As long as they keep growing, shareholders and investors are happy. Even at that scale, the pressure for growing engagement never lets up.

My experience from being inside has me convinced that everything follows from this.

Conceptualized using Google’s Nano Banana, designed by Josh Kramer and Esther Abosch for New_ Public

How incentives shape design … and outcomes

We’ve all seen these incentives play out in the dominant design choices of the past decade. Algorithms elevate emotionally-charged or provocative posts, even when they are polarizing, harmful, or misleading for billions of people.

Features like infinite scroll and autoplay remove natural stopping points and keep people moving through the feed without friction, for far longer than people intend. And notifications like “someone liked your post” aren’t about connection; they’re engineered to pull people back in after they’ve left. Each of these choices reflects the same simple underlying logic: more engagement means more revenue.

As the Director of Community Partnerships at Meta, I worked closely with the team responsible for Facebook Groups. From local neighborhood spaces to large-scale convenings of millions, Groups were one of the few places on the platform where people were still building genuine communities and really connecting with one another. But I experienced firsthand how our business model of serving advertisers with data and attention created its own gravitational pull beyond the leadership decisions of any one person or group.

Specifically, my team was responsible for the company’s relationship with group admins like Lola Omolola of the massive women’s group Female IN (FIN), who were the backbone of Facebook Groups. They dedicated enormous time and emotional labor to maintaining trust and keeping conversations and their communities civil. They were on the front lines, seeing where conversations were getting ugly, and hacking together the tools they needed for themselves.

For example, FIN members use a grumpy cat meme to show that a comment isn’t following group norms. My team tried to get these ideas embedded into the product at scale, but we would lose out in the product roadmapping process to features like recommending new groups in users’ news feeds, which could more directly boost metrics like time spent on the app.

Many of my colleagues and I cared deeply about the people and communities we served, and did manage to create innovative ways to support Group Admins, including establishing the Facebook Community Leadership Program and deploying $10M to them, but that wasn’t good enough. I came to Facebook attracted by the opportunity to build a better community-building platform for two billion people, but my values had no path to influence the product to become healthier at its core. I simply couldn’t compete against the culture that this business model created.

How these incentives show up in our lives

These everyday design choices don’t operate in a vacuum; they sit inside broader social dynamics. They aren’t the sole cause of falling trust, polarization, or disconnection, but they routinely amplify and accelerate those problems.

You can also see this in local platforms like Nextdoor, where fear-based posts about crime receive disproportionate visibility because they generate high engagement. As I shared in Jonathan Haidt’s newsletter, the consequences are not only about what we feel when we turn on our phones or open our computers — these apps are rewiring our trust in one another, and in our families and neighborhoods.

The Algorithm Next Door

When a system is optimized for attention, these distortions become the norm, not the exception. In un-ending, always-on feeds, people can feel surrounded by information yet strangely unanchored from their actual communities.

Another consequence is the rise of what friends of New_ Public Sam Pressler and Soren Duggan describe in Connective Tissue as “synthetic connection.” Platforms present an endless stream of posts from other people, which can feel like community but lacks the core elements that make real communities meaningful: reciprocity, accountability, shared norms, and productive friction.

Real community involves negotiation. It involves showing up for one another, navigating disagreement, and giving time and care even when it is not convenient. Platforms strip away these dimensions because friction interrupts engagement. The result is an experience that looks like connection but does not build the trust or belonging that real relationships require.

Over time, synthetic connection can begin to replace the real thing. Participation becomes consumption. Community becomes a feed. People spend more time online yet feel less rooted in the relationships around them. These outcomes flow directly from the incentives that guide the system and quietly shape how people show up, how they relate to one another, and how collective life functions.

It’s only going to get worse

The same incentives that shaped social media are now actively shaping the next generation of technology. As AI tools and chatbots become woven into everyday life and increasingly mediate how people communicate, search for information, and make sense of the world, the stakes grow larger.

Unless we address the incentives underneath, we risk replaying the same dynamics at an even greater scale. The companies building these systems are under intense pressure to find profitable business models, and the fastest path is the same one that shaped social media: attention capture and data extraction through surveillance advertising.

Just like social media, AI tools aren’t necessarily harmful on their own; they can genuinely help people communicate, navigate conflict, access information, or find what they need in their communities. But when they’re developed inside that business model, the technology is bent toward those incentives.

AI companions are a clear example. For some people, they may offer moments of comfort or reduce loneliness. But when these tools are built to serve retention-driven goals, they are pushed toward deeper and more persuasive forms of synthetic intimacy. Instead of helping people build relationships with other humans and civic institutions, the system is optimized to keep people engaged in the app itself.

None of this is inevitable. If the incentive is to maximize engagement, AI will find efficient ways to capture attention. If, instead, the incentive is to strengthen community well-being, AI can help support community stewards, identify shared norms, mediate conflict, increase accessibility, and lower the barriers to civic participation.

AI could help people spend less time in digital loops and more time in real relationships. We have agency to shape these new tools and technologies into something prosocial and public-spirited.

What better incentive structures could look like

To understand why these dynamics are so hard to change, we also have to look at how the tech ecosystem is funded. Traditional venture capital depends on a few companies generating massive, fast returns, pushing founders to focus on extreme growth, which gets embedded into their culture. No one has articulated this better than Catherine Bracy in her book, World Eaters: through one case study after another, she describes how this ethos undermines workers, communities, and even the health of the systems being built.

Because this is the dominant funding model in Silicon Valley, social platforms are built to optimize for growth in revenue above all else, and surveillance advertising has become the easiest business model to deliver this promise. There’s been vanishingly little innovation over several decades of social media, and that’s what we are trying to unlock.

If we want digital spaces that support trust, belonging, and civic health, and we want them to be self-sustaining, we need business models that align incentives with community well-being. How do we build incentives oriented towards the public good, instead of meeting the growth aspirations of VC investors?

We have a few examples that show what this could look like. Public media organizations and projects like Wikipedia are sustained through broad-based public support, demonstrating how widely relied-upon resources can be funded and stewarded by the communities that depend on them.

Local social platforms like Front Porch Forum, which centers trust and neighbor-to-neighbor connection, provide a proof point on how local advertising — where the small businesses advertisers are equally invested in the health of the community as the members — can be a healthy and profitable business model too. Not everything has to be a “rocket ship” or “hockey stick.” As Ben Waldman, CEO of Canadian social startup Gander, told us recently: “We’re happy to be a small business that does a good job.”

How New_ Public is approaching this

At New_ Public, we start from the idea that technology works better when it’s accountable to the people who rely on it. Instead of building for investors, we’re building for the public. Everything follows from this.

With Roundabout, our local community app currently in private beta testing, we’re building a platform where community stewards, not engagement-driven algorithms, shape each local space. Stewards bring the context, culture, and lived knowledge of their neighbors.

They help set expectations, address conflict, and cultivate a comfortable atmosphere — the kind of participation that consistently produces stronger community outcomes (as reflected in recent research from New_ Public Co-Founder Talia Stroud) than spaces designed around monetization. Not only do our teams prioritize their feedback in our roadmaps, Roundabout is establishing a model in which stewards can be compensated for this vital work.

👋 Introducing Roundabout: built for neighbors, with neighbors

Philanthropy has provided the patient capital and runway for us to get the design right without VC pressure, but it also presents its own challenges. So to build Roundabout sustainably, we are exploring revenue models that align with community benefit, including some of the examples above, such as member support, and even local ads and business sponsorships that reflect the community. We’re committed to using user data only to deliver a better experience, and never profiling or targeting members or selling their data.

Yes, we may grow more slowly. We may not generate the same returns as traditional social media. But we will not sacrifice healthy community connection on anything we build.

Everything comes back to one question: Who is the platform accountable to? If the answer is investors, the platform will eventually warp toward extraction and surveillance. If the answer is communities, the result can be a product designed around the needs of real people, not VC investors. Something that can grow healthier over time, the way Wikipedia has over decades, sustained by volunteers, grounded in shared norms, and accountable to its users. The incentives decide which path we get.

I know this is all easier to write about than it is to do. But I worry that if we don’t apply the same rigor to innovating new business models as we do on the design and product innovation for healthier communities, we won’t develop sustainable pathways for healthy digital public spaces to scale. The ways we connect with each other depend on us getting this right.

– Deepti Doshi, Co-Director, New_ Public

---

Thanks Deepti!

Making a list and checking it twice,

–Josh