When a Session Ends, and Why We Sweated the Small Stuff

The Problem We Didn't Expect

When we first built Intentr, we thought the session model would be straightforward. You set an intention, a timer starts, you consume content in bounded time. Done. But real people don't work in clean blocks. Someone opens the app to check a creator's latest post. Sets a 10-minute intention. Reads three pieces. Phone rings. They step away. Come back 45 minutes later. Scroll for two more minutes. Close the app.

Is that one session interrupted? Two sessions? Did they fail their intention, or did they succeed at the part they actually did? We realised we'd built a strict container but hadn't thought carefully about the edges. The attention ledger was tracking time spent, sure, but we were forcing users into a binary: stay in the session or abandon it. No middle ground. No grace for human life.

What Our Data Told Us (and Didn't)

We looked at the first thousand sessions from our early testers. About 8 per cent of people left the app and came back within the same calendar day, usually within 30 minutes. Some came back within five minutes. Others vanished for hours. On the surface, it looked random. But when we read the feedback, patterns emerged. People weren't flaking out of bad intentions. They were getting interrupted by work, kids, messages. Or they'd finished their stated goal and just... left, thinking the session had ended naturally.

One tester wrote: 'I read the two articles I meant to read, so I left. But then I worried I'd somehow messed up the session tracker.' That comment stung. We'd made intentionality feel like homework. The mechanics were getting in the way of the actual experience. We needed a rule that felt less like a gotcha and more like, well, common sense.

The Rule We Built

We settled on this: a session ends early if you leave the app and don't return within ten minutes. If you come back within ten minutes, it's still the same session. If you return after ten minutes, it's a new one.

Ten minutes felt right for a few reasons. It's long enough to handle a real interruption (a call, a message, someone at the door) without resetting the whole session. It's short enough that if you're genuinely done, you move on. And crucially, it puts the user in control. You're not being judged for stepping away. The app just remembers you were here. If you meant to keep going, the boundary is forgiving.

We also changed the language. Instead of 'session abandoned', the ledger now shows 'session paused' if you left and came back. Small word change, but it shifted the mental model. You're not failing. You're living. The app is just keeping track.

Why Ten Minutes (Not Five, Not Thirty)

The specificity matters. If we'd gone with five minutes, we'd fracture legitimate pauses. A creator's notification pops up, you check it, come back. Session reset. That feels petty. If we'd gone with thirty, we'd be too forgiving, and the ledger becomes noise. Is this person actually spending 47 minutes reading, or 15 minutes reading plus a 30-minute phone call? The ledger is supposed to show you where your attention actually went, not hide it.

Ten minutes sat in the middle. It's enough time for life to interrupt, but short enough that intentionality still has teeth. When you come back after ten minutes, you made a choice to re-engage. That matters. Your ledger will reflect that you paused, came back, and spent more time. That's not failure. That's the real story of how you spent your attention.

How Creators Fit Into This

We also had to think about creators. Creators on Intentr get 85 per cent of subscription revenue from engaged viewers. But what does 'engaged' mean if sessions keep fragmenting? We decided: if someone pauses and returns, and you're tracking it in the ledger, that counts as one continuous engagement window. The creator sees that someone came back. They see real attention, not just session counts. That's way more valuable for understanding who's genuinely interested in their work.

One creator told us after we shipped this: 'I can actually see when someone read one post, came back an hour later, and read another. That's a real person, not a bot. I care about that pattern way more than perfect session metrics.' That vindicated the whole approach. We were building for humans, not for numbers.

The Thing We Didn't Anticipate

A couple of months after launch, someone asked us: what if someone uses the app, leaves, and comes back after nine minutes, then leaves again and comes back after another eight minutes? Is that one session or three? Technically, by our rule, it's one. But is it one intentional session, or someone endlessly drifting in and out?

We haven't changed the rule. We don't think we should. Part of intentionality is being honest about how you actually consume. If you're a person who checks in and out every few minutes, that's real information. The ledger captures it. You can see it. That matters more than us enforcing some abstract ideal of what a 'proper' session should look like.