Viral Loop

Users bringing in new users as part of using the product, measured by the viral coefficient.

Also known as: viral loop, virality, K-factor, viral coefficient

A viral loop makes each user bring more users through normal use: invitations to collaborate, shared content carrying branding, network effects requiring counterparts. The viral coefficient (K) measures it — average new users each user generates. Above 1.0 sustained, growth compounds exponentially; below, it decays regardless of marketing spend.

K = invites sent per user × conversion rate of invites
K = 1.2 → 100 users become ~120, then ~144... (compounding)
K = 0.8 → 100 users become ~80, then ~64... (decaying despite spend)

True virality is designed into core workflows (collaboration invites, public sharing, multiplayer necessity), not bolted on as referral popups. Users invite because the product works better with others aboard — the loop serves them first, growth second.

The classic mistakes:

Incentivized virality mistaken for real. Cash-for-invites buys mercenary signups that churn; K looks healthy while retention dies. Measure invited-user retention separately — mercenaries reveal themselves in weeks.
K without cycle time. K=1.5 with a year-long cycle grows slower than K=1.1 weekly. Report K alongside cycle time; speed of the loop matters as much as its ratio.
Ignoring saturation. Viral loops exhaust susceptible populations — early explosive K decays as markets saturate. Plan the second loop before the first plateaus.
Dark patterns for invites. Address-book scraping, misleading share buttons, spammy defaults — growth that burns trust converts poorly and invites platform bans. Earn invites with value, never extract them.
Confusing sharing with virality. Social shares that nobody acts on are vanity. Track full loop completion (share → signup → active), not top-of-funnel shares.

Validate ruthlessly: cohorted invited-user retention versus organic. Real virality retains; manufactured virality churns. See growth loops for the broader compounding family and network effects for the defensibility version.