What the data actually shows
Sociological research on how personal networks change over time — including studies by Mollenhorst, Völker, and Flap that re-interviewed the same people years apart — finds that networks turn over substantially. A frequently cited finding from this line of work is that people lose touch with a large share of their close network over a period of several years, on the order of replacing roughly half of it over about seven years. The specific fraction varies by study and population, but the broad pattern is that networks are far more fluid than they feel from the inside.
Crucially, this turnover is mostly passive rather than the product of conflict. The same research finds that ties tend to fade because the shared contexts that sustained them — a workplace, a neighbourhood, a stage of life — change or disappear, not because of arguments or falling-outs. People drift apart largely because the structures that kept them in regular contact dissolve.
Dunbar's 'layers' model gives this a shape. It describes social networks as nested circles — a small inner core of a few close relationships, surrounded by progressively larger and less intense layers out to a few hundred acquaintances. Maintaining the inner layers takes regular contact and emotional investment, which is limited, so people continuously cycle relationships inward and outward across these rings as time and circumstances allow.
Why this feels different from how it actually is
Drifting apart feels like a loss or a failure partly because it is so gradual that there is no event to make sense of. A friendship that simply faded leaves no clear cause, so the mind tends to supply one — usually self-blame or the sense that the bond 'should' have lasted. The absence of a falling-out can paradoxically make the fading harder to interpret than an outright argument would.
It also feels different because we carry an implicit ideal of permanent friendship — the notion that real closeness should endure regardless of circumstance. Measured against that ideal, ordinary turnover looks like something going wrong. But the network research suggests permanence is the exception; continuous replacement is the norm, even among close ties.
Finally, the passive nature of drift hides its real cause. Because no one decided to end the friendship, it can feel like neglect or indifference. In the data, though, the dominant driver is structural — the loss of a shared context and the simple scarcity of time — rather than any change in how much people valued each other.
What the research says to do about it
Because drift is driven mainly by lost shared contexts, the research-aligned response is to recreate shared context deliberately for the relationships you want to keep. Ties that no longer have a built-in reason for regular contact need a manufactured one — a standing commitment, a recurring activity — to survive, since proximity and frequency are what sustained them in the first place.
It also helps to accept the layered, shifting structure rather than resist it. Dunbar's model implies that the close inner circle is necessarily small and that maintaining it has a real cost in time and attention. Concentrating limited relational energy on a few inner-layer ties, rather than trying to hold every relationship at full intensity, fits how the system actually works.
And reframing drift as circumstance rather than rejection tends to ease the self-blame and, often, to keep doors open. Relationships that faded passively can frequently be reactivated, precisely because they ended through lost contact rather than conflict — there is no breach to repair, only a context to rebuild.
What the research says does not help
Reading every faded friendship as a personal failure is both inaccurate and unhelpful. The network research shows substantial turnover happening to nearly everyone, mostly through changing circumstances rather than fault. Treating an ordinary, near-universal pattern as evidence of a character flaw misdiagnoses it and adds avoidable distress.
Trying to keep a large number of relationships at full closeness tends not to work, because the time and emotional investment that close ties require is genuinely limited. Dunbar's model suggests the inner layers are small for structural reasons; spreading the same energy across far more people usually means none of them stay close.
Assuming a drifted relationship is permanently over is often wrong. Because most drift is passive and not the result of conflict, many faded ties can be picked back up when a shared context is restored. Writing them off entirely forecloses a reconnection that the absence of any real rupture would otherwise allow.
Real numbers in context
Personal networks turn over substantially over time. Sociological research that re-interviewed the same people years apart (Mollenhorst, Völker, and Flap) found a large share of close ties replaced over a span of years — a widely cited estimate is losing touch with roughly half of one's close network over about seven years. The exact fraction varies by study, but the takeaway is that networks are far more fluid than they feel.
Most of that turnover is passive, not the result of conflict. The same research attributes fading ties largely to the loss of shared contexts — a job, a neighbourhood, a life stage — rather than to arguments or falling-outs. People drift apart mainly because the structures that kept them in contact change.
Dunbar's layers model frames friendship as concentric circles — a small close core surrounded by wider, less intense rings out to a few hundred acquaintances — that people move in and out of across a life. Drifting between layers is the ordinary motion of the system, constrained by the limited time and attention that close relationships require.