On Streaks, and Admitting We Were a Little Wrong

Two of our blog posts make a clean argument against streaks. One of them calls them out by name, promising “no streaks to maintain, no red badges engineered to panic you.” The other places the streak alongside autoplay and the red badge as gambling-psychology mechanics borrowed by apps that were never really built for friendship. We meant both pieces. We still do.

We’re adding streaks to Colloquies.

This is the kind of contradiction worth sitting with for a paragraph before trying to explain it away. We told a story about ourselves to a few thousand readers, and we’re now shipping something that, at first glance, looks like the thing we said we wouldn’t. The rest of this post is our case for why it isn’t. We’ll show the seams. You can decide whether the result holds together.

What we knew when we said it

The case against streaks wasn’t theoretical. There’s a small library of writing now on how engagement mechanics work on the brain. Tristan Harris and the Center for Humane Technology have spent a decade documenting it. Stanford’s Anna Lembke has described the dopamine-deficit pattern the heaviest of these apps push the brain into. Duolingo’s green owl, Snapchat’s snapstreaks, Instagram’s red badges. These are the canonical examples, and they’re canonical because they work. They turn a feature into a small, persistent low-grade dread. The dread is the product.

Our position was simple. If the goal is to help close friends stay close, why borrow the mechanics that hollow out the rest of our digital lives?

We still believe most of that. The question that turned out to be harder is whether every nudge is the same kind of nudge.

What changed

We’ve been watching how people use Colloquies for the better part of a year, and one pattern is hard to ignore. The friends who already keep up with each other, who organize the standing dinner and the weekly walk and the long voice notes on the drive home, don’t actually need our app. They’ve already solved the problem the app is trying to help with. They might enjoy it as a small ritual on top of what they already have, but they would be fine without us.

The rest of us are the actual users. The rest of us reread “let’s catch up soon” texts from old friends and feel a small hot wave of guilt and never actually call. We care about the friendships and somehow let the weeks slip by anyway. We’re busy in a way our parents weren’t quite busy in: more screens, more inputs, more obligations spread across more time zones, more attention pulled in more directions before we’ve even had breakfast.

If you were already disciplined about reaching out, you wouldn’t be opening Colloquies on Sunday night. You’d be in the call. The honest reading of who needs this app is people who’d love a small, well-designed nudge toward the relationships they already say matter most.

That’s the same reason an owl on a phone teaches more Spanish than a textbook on a shelf. It’s the reason a meditation app gets more sittings than a cushion in the corner of a bedroom. It’s the reason a couch-to-5K plan finishes more 5Ks than a vague resolution. Engagement design is one of the most powerful tools software has built in the last twenty years, and powerful tools are mostly defined by what we aim them at. The same mechanic that pulls a teenager into a five-hour scroll can pull an adult back to a once-a-week conversation with the people they love. We get to choose what we use it for.

The behavioral science here isn’t subtle. Wendy Wood, the USC researcher who’s spent decades on how habits actually form, has shown that close to half of our daily behavior is habitual rather than deliberate. The small environmental cues (a phone on the nightstand, a running shoe by the door, a notification at the right hour) do most of the lifting. The mistake the engagement industry made wasn’t using cues. It was using them to extract attention rather than to support behavior people actually wanted to keep.

The objection we have to take seriously

There’s an obvious counterargument here, and it deserves its own paragraph rather than a parenthetical. Every engagement designer in the last fifteen years has told themselves a version of the story we’re now telling ourselves: it’s a gentle nudge, not manipulation, our users actually want this. That story is what every dark pattern hides behind. Anyone who’s ever felt the small panic of a 600-day Snapstreak on the verge of breaking can tell you the dread is the feature, not a side effect of it. The road from “ambient signal” to “loss-aversion engine” is short, and it’s paved with people who believed they were being more careful than the last app.

We can’t promise we’ll get this right. What we can do is show our work.

Where we drew the lines

The streak we built is not the streak you’ve used before. The differences aren’t cosmetic; they’re the entire feature.

It belongs to the circle, not the person

The streak lives with the circle, never with a single person. Each chapter (one full cycle of prompt, answers, and conversation) adds to the count when the circle has acknowledged it together. The mechanic that gives Snapstreaks their bite (I’m letting my friend down personally if this breaks) simply doesn’t exist here.

Any admin can turn it off

The whole feature is opt-out at the circle level. An admin can disable streaks at any time, and the flame disappears for everyone in the circle while the count freezes wherever it was. Real consent, where the friendship actually lives.

A quiet chapter preserves the streak

Each person in a circle can mark a chapter as quiet (a hard week, a vacation, a stretch where life is louder than usual) and skip it. The streak holds as if they had answered. Only silent absence breaks it.

The number lives on its own page

A small flame appears next to the circle once it has its first successful chapter. That’s the ambient signal. The actual count lives on the streak page, one tap away, alongside a plain explainer. We didn’t want a counter at the top of the screen turning a friendship into a leaderboard.

No streak-specific notifications

You will never get a buzzed “Your circle is about to lose its streak!” from us. The Strava-style “about to break” alarm is the heart of what makes streaks feel coercive, and we won’t ship it.

What it’s actually for

The point isn’t to be there every single week. Life isn’t built that way, and a friendship that demanded it would be a worse friendship. The friendships that actually carry a life run on a recurring cadence rather than constant presence, and quiet stretches are part of how the cadence breathes. The point is to notice the chapters you were there for each other. Ten chapters of one honest answer to one real question, with the same handful of people, is something most adult friendships don’t get any year of their lives. That’s worth a small, quiet flame.

If we got the design wrong, we’ll change it or remove it. The promise from the older posts still stands: nothing in Colloquies is allowed to be a guilt machine. That promise is older than this feature, and it’ll outlast the feature if we ever have to choose.

A streak isn’t a deadline. It’s a record of showing up. The flame doesn’t care how long it’s been; it only knows that the people you care about kept finding their way back to the same room.

Hero image by Clay Banks on Unsplash.