Viral coefficient (K-factor) calculator

The viral coefficient, or K-factor, is the number of new users each existing user brings in. The idea comes from epidemiology, where the reproduction number describes how many people one infected person infects. For a product it is the product of two measurable rates:

If users send 3 invites on average and 20% of invites convert, K = 3 × 0.2 = 0.6: every 10 users bring 6 more, those 6 bring 3.6, and so on. The calculator models this generation by generation:

Why K < 1 still matters: amplification 1/(1−K)

K below 1 is often dismissed as “not viral”. That misses the point. Summing the generations gives a geometric series: 1 + K + K² + … = 1/(1−K). Every user you acquire through any channel turns into 1/(1−K) users in total. At K = 0.6 that is 2.5×: a campaign that brings 1,000 users ends up bringing 2,500. At K = 0.9 it is 10×. Amplification grows sharply as K approaches 1, which is why moving K from 0.8 to 0.9 doubles the effect of all your acquisition.

With 1,000 starting users and K = 0.6 the base converges to 1,000 / 0.4 = 2,500. With a steady 200 users a cycle from other channels, new users per cycle converge to 200 / 0.4 = 500.

K tells you how much a loop grows; viral cycle time tells you how fast. It is the time between a user joining and the users they invite joining. Two products with the same K can grow at very different speeds: with K = 1.2, a 2-day cycle gives 15 cycles a month, a 30-day cycle gives one.

David Skok’s essay on viral marketing shows that shortening the cycle often beats raising K. Cycle time shrinks when inviting is part of the first session rather than a later prompt, when the invitee gets value before signing up, and when the invite arrives at the moment of need — a shared document, a meeting link, a request for feedback.

How to measure the inputs

• Invites per user. Take a cohort that has been active for at least one cycle and divide all invites they sent by the size of the cohort, including users who sent none.
• Invite conversion. Invites that led to a new active user, not just a sign-up, divided by invites sent by that cohort. Deduplicate people invited several times.
• Cycle time. Median days from the inviter’s sign-up to the invitee’s sign-up.

Measure by cohort, not across the whole base: older users invite less, and averaging them in hides how the loop performs for new users. For the steps between invite and activation, map the loop in the funnel conversion calculator, and test changes to the invite flow with the A/B test calculator.

• Make sharing part of the job. Products where collaboration is the value — documents, design files, scheduling — generate invites without asking. Look for the moment in your product when a user naturally needs someone else.
• Lift invite conversion. A clear preview of what the invitee gets, sign-up in one step, and a first screen that continues the invite context instead of a generic onboarding.
• Activate invited users. An invited user who never reaches value will not invite anyone. Invitee onboarding is part of the loop: it sets both conversion and the next generation’s invites.
• Ask at the moment of success. Prompts after a user achieves something convert better than prompts on day one.

Common mistakes

• Counting sign-ups instead of active users. Inflates K with users who never come back and never invite.
• Averaging only over inviters. Invites per user must include users who invite nobody.
• Ignoring saturation. K measured in a small community falls as invites reach people who already use the product or are not interested.
• Confusing K with referral programs. Paid referral bonuses raise K but also cost money per user; account for that in CAC payback.

1. Skok D. Lessons Learned – Viral Marketing. forEntrepreneurs.com — viral coefficient, viral cycle time and why cycle time matters.
2. Anderson R. M., May R. M. Infectious Diseases of Humans: Dynamics and Control. Oxford University Press, 1991 — the reproduction number that the K-factor borrows from epidemiology.