Engagement Is Not Wellbeing

Mental health apps are built and funded like consumer products. That means they measure what consumer products measure: daily active users, retention, session length, notification open rates, streak maintenance. These are the metrics in investor updates, pitch decks, and growth reviews.

The problem is that for mental health tools, optimizing for those metrics and optimizing for user wellbeing are often pulling in opposite directions.

The Engagement Playbook, Imported Wholesale

The same mechanics that made Duolingo's retention numbers impressive get imported, largely intact, into emotional wellness products. Streak systems that introduce loss aversion. Push notifications timed to break avoidance. Progress dashboards that reward consistent use with visual feedback. Content feeds that always have something new to engage with.

None of this marks bad actors. It's just how consumer product development works. You build, you instrument, you optimize for what you can measure. Most people building these products are genuinely trying to help users. But the optimization target pulls in a specific direction — toward ongoing engagement, not toward people being done.

The Subscription Conflict of Interest

When your revenue depends on users continuing to subscribe, you have a structural conflict against resolution.

A user who processes what they were carrying and no longer needs your app is, by your metrics, churn. A user who keeps returning is retention. These categories don't align with user outcomes — they're about revenue continuity.

This doesn't mean subscription mental wellness apps are deliberately keeping people unwell. Most aren't. But it creates pressure — subtle, unspoken, often invisible even to the people doing the work — toward maintenance over completion. Features that keep users engaged are easier to justify than features that help users finish and leave.

The reframe happens quietly. Emotional health becomes 'ongoing practice' rather than something that, for a specific thing you're carrying, might just need to end.

What the Research Actually Describes

The expressive writing research — James Pennebaker's foundational work at the University of Texas Austin, beginning in the late 1980s, and the broader literature it generated — typically tested brief, bounded interventions. Participants wrote for 15-20 minutes a day for 3-4 consecutive days. Subsequent studies looked at health outcomes, psychological wellbeing, and physical health markers after those interventions completed.

The paradigm isn't 'maintain a daily writing practice.' It's closer to: write intensely about something difficult for a few sessions, and measure what changed afterward. Pennebaker's own framing emphasizes the disclosure act itself — the externalization of internal experience — more than the ongoing accumulation of entries.

Apps built on this research often don't acknowledge that distinction. They take the mechanism (writing about emotional content) and attach it to an engagement model the underlying research doesn't support. It's not fraudulent, exactly. It's a loose interpretation with a business model incentive behind it.

The Ideal Outcome Looks Like Churn

Here's the inconvenient arithmetic: a tool that genuinely helps someone put down what they were carrying might get used intensely once or twice and never again.

By engagement metrics, that's a failed acquisition. It's indistinguishable from someone who downloaded the app, bounced at onboarding, and deleted it. Both show up as zero on retention curves.

But one of those is abandonment. The other is resolution. The metrics don't care which is which.

This creates a selection pressure in the market. Products designed around resolution don't produce retention numbers that attract investment or demonstrate growth. Products designed around ongoing engagement do. So the market naturally favors the latter — not because anyone decided that was the right outcome for users, but because that's what the incentive structure rewards.

Building Toward Completion

I built Unheavy as a one-time purchase partly for this reason: I don't want a financial model that creates an incentive against users finishing.

There's no subscription to maintain, no streak to protect, no engagement metric I'm optimizing around. When the destruction animation ends, the session is complete. The design goal isn't 'come back tomorrow.' It's that you put something down and walked away. That should look like churn by standard metrics — and that's fine.

I'm not arguing that subscription models are wrong for every emotional health tool. There are people whose wellbeing genuinely benefits from consistent daily practice, and tools built for them should exist. What I'm arguing is that the conflation of engagement with wellbeing is a specific, structural problem in how most of these products get built — and that the builders making these tradeoffs often don't name it directly.

The right question isn't 'how do we get users to come back?' It's 'what does success look like if they never have to?'