Friendships Don't Die From Fights. They Die From Drift.

You probably have a friend you genuinely love who you haven't seen in six months. Maybe longer. Nothing happened — no falling out, no bad text exchange. You both got busy, the group chat went quiet, and then you stopped noticing it had gone quiet. This is the most common way adult friendships end, and almost nobody talks about it honestly.

We have a vocabulary for breakups and a vocabulary for fights, but the slow fade — the one that accounts for most lost friendships after 30 — has no name and no rituals around it. It just happens, quietly, to people who would have sworn it never would.

What kills adult friendships?

Drift kills adult friendships far more often than conflict does. The relationship researcher Shasta Nelson, in her book Frientimacy, gives the problem a clean diagnosis. Every healthy friendship, she argues, rests on three pillars: positivity, consistency, and vulnerability. Positivity is the accumulation of good shared experiences. Vulnerability is the willingness to share real things. Consistency is the simple act of regular contact over time.

Her central, slightly uncomfortable claim is this: most friendships don't die from a failure of positivity or vulnerability. They die from a failure of consistency. The good times were good. The trust was real. You just stopped seeing each other often enough for any of it to matter.

Why is consistency the pillar we lose first?

Consistency fails first because it's the only one of the three that requires active maintenance against entropy. It's worth asking why this specific pillar is the one that gives out.

Positivity is easy to keep up when you do see someone — most friendships are built on people you actually enjoy. Vulnerability is harder, but it's also rarer and more episodic; you don't need a vulnerable conversation every week to maintain trust built over years.

Consistency is different. Left alone, it always trends toward zero. Your calendar fills with obligations, your bandwidth shrinks, and the people who are loudest in your life — work, immediate family, whoever messaged most recently — crowd out the people who are quietest. The quietest people are often the ones you love most, because love doesn't generate urgency. A close friend who hasn't reached out in two months isn't a fire alarm. They're a soft hum you stop hearing.

The diagnostic question Nelson offers cuts through the rationalisations we build around this: not "are we still friends?" — almost everyone says yes — but "when did we last actually connect?" The honest answer is usually further back than you think.

What does consistency actually require?

Consistency requires cadence appropriate to the depth of the bond — not intensity. Nelson is explicit about this: a friendship doesn't need weekly dinners or daily texts. Some friends need a quarterly catch-up to stay strong. Others — the inner-circle ones — need something closer to weekly contact, even if it's small.

The mistake most people make isn't aiming too low. It's having no aim at all. Without an explicit cadence, every friendship defaults to whichever cadence emerges from the chaos of your week. And the cadence that emerges from chaos is, reliably, too slow.

The second mistake is treating consistency as something you'll feel your way through. You won't. Six weeks is the same internal feeling as two weeks for most people — it doesn't register until someone reminds you. Memory is not a reliable instrument for tracking the passage of time across multiple relationships. This is why the people you live with stay close almost effortlessly and the people you don't drift almost inevitably: presence does the bookkeeping for you. Without presence, you need some other system.

What does the system have to do?

If you accept Nelson's diagnosis — and it's hard to argue with — then what you actually need is fairly specific:

• A way to make the cadence of each friendship visible, not felt.
• A way to notice drift early — while it's still a four-week gap and not a four-month one.
• A way to act on the noticing without it feeling like a chore or a guilt trip.

This is exactly the problem BondScore was built to solve, and it's where the abstract framework becomes a concrete product decision.

How does BondScore implement consistency?

BondScore tracks two things for each of the five or so people you most want to stay close to — what we call your Inner Circle: how recently you've spent meaningful time together, and whether that cadence matches what you said you wanted at the start.

When you add someone to your Inner Circle, you set a weekly time goal. It's modest by design — most people set somewhere between 30 minutes and a couple of hours a week. That goal becomes the baseline cadence for the bond. Then BondScore quietly does the bookkeeping. If you spend time together, the bond's score goes up. If you don't, it gently decays — not as punishment, but as a visible signal that the gap is widening.

The decay isn't aggressive. A good week of contact buys you weeks of grace. The point isn't to make you feel bad on a Tuesday because you haven't seen someone since the previous Tuesday. The point is to make sure that when six weeks have passed without contact, you actually notice — instead of waking up one Sunday and realising it's been four months.

Once a week, on Sunday, the app surfaces a planning view. It shows you who in your Inner Circle is drifting, suggests specific activities the two of you might actually enjoy based on what you've told it about them, and lets you put something in the calendar in about three minutes. The Sunday ritual is small on purpose — anything longer wouldn't survive contact with a real week.

For the people you care about but who don't need weekly attention — the Outer Circle — the system is gentler still. No score, no decay, just a quiet timestamp of when you last connected and a soft nudge when the gap has grown longer than feels right. Nelson's point about cadence-appropriate-to-depth is, structurally, the difference between how the two circles work.

What the framework doesn't cover yet

In the interest of honesty: Nelson's three pillars are positivity, consistency, and vulnerability. BondScore is good at consistency, decent at positivity (every logged activity captures how it felt), and currently does nothing structured about vulnerability. That's a real gap. We think vulnerability is harder to model without becoming intrusive, and we'd rather build it slowly than badly. It's on the roadmap, not in the product.

But for the specific problem most adult friendships actually have — the one Nelson named — the consistency engine is the part that matters most. It's the pillar that fails first and fails quietly. Get that right, and the other two have something to stand on.

How to start: a small experiment

If you do nothing else after reading this, try the diagnostic question on the five people you'd most regret losing touch with. Not "are we still friends?" but "when did we last actually connect?" — meaning a real conversation, in person or on video, not a meme exchange.

Write the dates down. Look at the list.

If the list surprises you, that's the gap Nelson was writing about. And it's the gap BondScore was built to close.