K-factor answers a specific question: does your invite mechanic replace itself? At K equal to 1, every user brings in exactly one more user on average, and the chain neither grows nor dies on its own. Above 1, each generation is larger than the last — genuine, self-sustaining compounding growth. Below 1, each generation is smaller, and the mechanic decays toward zero without other acquisition feeding it.
The number gets treated as more dramatic than it usually is. A K-factor of 0.4 isn't a failure — it means invites extend your reach by 40% on top of whatever else is bringing users in, a real and valuable multiplier that just isn't sufficient alone. The K ≥ 1 threshold matters because it's the only point where the mechanic could theoretically sustain a business with zero other acquisition, not because anything below it is worthless.
K-factor also isn't stable over time the way a single reported number implies. It typically declines as a network approaches saturation in its addressable audience — the same invite mechanic that produced K=1.2 in an underserved market often can't sustain that once most plausible users are already in the network. Treating an early K-factor as a permanent growth engine is a common and expensive misread.
Formula
K = i × c, where i = average invites sent per existing user, c = conversion rate of those invites into new signups
Measure both inputs over the same cohort and time window — mixing an all-time invite average with a recent conversion rate produces a K-factor that doesn't describe any real cohort's actual behavior. Project compounding growth as Users(n) = Users(0) × (1 + K)^n only as a best-case ceiling, since K typically declines as the network saturates.
Why Viral Coefficient matters
K-factor is the only number that tells you whether an invite mechanic can theoretically sustain growth on its own — everything below K=1 is a real multiplier on other acquisition, not a standalone growth engine, and conflating the two leads teams to underinvest in paid or organic channels while waiting for virality to compound on its own.
The same K, two different verdicts
A consumer app reports K=0.6 in its first quarter and a K=0.6 again eighteen months later. In quarter one, that's a strong signal — the invite mechanic is meaningfully extending reach in an unsaturated market and likely to be joined by paid and organic growth. Eighteen months later, in a market where the addressable audience is mostly already reached, a flat K=0.6 usually means the mechanic hasn't improved and the team has been coasting on an early result rather than treating K-factor as something to actively grow.
Common mistakes
Treating any K below 1 as a failed mechanic
A K-factor of 0.3 to 0.7 is a genuine, valuable multiplier on other acquisition — the K=1 threshold marks self-sustaining growth, not the line between 'working' and 'not working.'
Reporting K-factor without a time window
K-factor changes meaningfully depending on the window it's measured over — a lifetime average and a trailing-30-day figure can tell completely different stories about whether the mechanic is currently healthy.
Assuming K-factor holds as the network scales
K-factor typically declines as a network approaches saturation in its addressable audience — an early, small-market K-factor is a starting point, not a permanent growth rate to model against.